Original Research

Nurse practitioners (NPs) and physician assistants (PAs) often help provide dermatologic care but lack the same mandatory specialized postgraduate training required of board-certified dermatologists (BCDs), which includes at least 3 years of dermatology-focused education in an accredited residency program in addition to an intern year of general medicine, pediatrics, or surgery. Dermatology residency is followed by a certification examination administered by the American Board of Dermatology (ABD) or the American Osteopathic Board of Dermatology, leading to board certification. Some physicians choose to do a fellowship, which typically involves an additional 1 to 2 years of postresidency subspeciality training.

Optional postgraduate dermatology training programs for advanced practice providers (APPs) have been offered by some academic institutions and private practice groups since at least 2003, including Lahey Hospital and Medical Center (Burlington, Massachusetts) as well as the University of Rochester Medical Center (Rochester, New York). Despite a lack of accreditation or standardization, the programs can be beneficial for NPs and PAs to expand their dermatologic knowledge and skills and help bridge the care gap within the specialty. Didactics often are conducted in parallel with the educational activities of the parent institution’s traditional dermatology residency program (eg, lectures, grand rounds). While these programs often are managed by practicing dermatology NPs and PAs, dermatologists also may be involved in their education with didactic instruction, curriculum development, and clinical preceptorship. 

In this cross-sectional study, we identified and evaluated 10 postgraduate dermatology training programs for APPs across the United States. With the growing number of NPs and PAs in the dermatology workforce—both in academic and private practice—it is important for BCDs to be aware of the differences in the dermatology training received in order to ensure safe and effective care is provided through supervisory or collaborative roles (depending on state independent practice laws for APPs and to be aware of the implications these programs may have on the field of dermatology.

Methods

To identify postgraduate dermatology training programs for APPs in the United States, we conducted a cross-sectional study using data obtained via a Google search of various combinations of the following terms: nurse practitioner, NP, physician assistant, PA, advance practice provider, APP, dermatology, postgraduate training, residency, and fellowship. We excluded postgraduate dermatology training programs for APPs that required tuition and did not provide a stipend, as well as programs that lacked the formal structure and credibility needed to qualify as legitimate postgraduate training. Many of the excluded programs operate in a manner that raises ethical concerns, offering pay-to-play opportunities under the guise of education. Information collected on each program included the program name, location, parent institution, program length, class size, curriculum, and any associated salary and benefits.

Results

Ten academic and private practice organizations across the United States that offer postgraduate dermatologic training programs for APPs were identified (eTable). Four (40%) programs were advertised as fellowships. Six (60%) of the programs were offered at academic medical centers, and 4 (40%) were offered by private practices. Most programs were located east of the Mississippi River, and many institutions offered instruction at 1 or more locations within the same state (eFigure). The Advanced Dermatology and Cosmetic Surgery private practice group offered training opportunities in multiple states.

Six programs required APPs to become board-certified NPs or PAs prior to enrolling. Most programs enrolled both NPs and PAs, while some only enrolled NPs (eTable). Only 1 (10%) program required NPs to be board certified as a family NP, while another (10%) recommended that applicants have experience in urgent care, emergency medicine, or trauma medicine. Lahey Hospital & Medical Center required experience as an NP in a general setting for 1 to 2 years prior to applying. No program required prior experience in the field of dermatology.

Program length varied from 6 to 24 months, and cohort size typically was limited to 1 to 2 providers (eTable). Although the exact numbers could not be ascertained, most curricula focused on medical dermatology, including clinical and didactic components, but many offered electives such as cosmetic and procedural dermatology. Two institutions (20%) required independent research. Work typically was limited to 40 hours per week, and most paid a full-time employee salary and provided benefits such as health insurance, retirement, and paid leave (eTable). Kansas Medical Clinic (Topeka, Kansas) required at least 3 years of employment in an underserved community following program completion. The Oasis Dermatology private practice group in Texas required a 1-year teaching commitment after program completion. The Advanced Dermatology and Cosmetic Surgery group offered a full-time position upon program completion.

Comment

There is a large difference in the total number of training and credentialing hours when comparing graduate school training and postgraduate credentialing of medical and osteopathic physicians compared with APPs. A new graduate physician has at least twice as many clinical hours as a PA and 10 times as many clinical hours as an NP prior to starting residency. Physicians also typically complete at least 6 times the number of hours of certification examinations compared to NPs and PAs.¹

Nurse practitioner students typically complete the 500 hours of prelicensure clinical training required for NP school in 2 to 4 years.^2,3 The amount of time required for completion is dependent on the degree and experience of the student upon program entry (eg, bachelor of science in nursing vs master of science in nursing as a terminal degree). Physician assistant students are required to complete 2000 prelicensure clinical hours, and most PA programs are 3 years in duration.⁴ Many NP and PA programs require some degree of clinical experience prior to beginning graduate education.⁵

When comparing prelicensure examinations, questions assessing dermatologic knowledge comprise approximately 6% to 10% of the total questions on the United States Medical Licensing Examination Steps 1 and 2.⁶ The Comprehensive Osteopathic Medical Licensing Examination of the United States Level 1 and Level 2-Cognitive Evaluation both have at least 5% of questions dedicated to dermatology.⁷ Approximately 5% of the questions on the Physician Assistant National Certifying Examination are dedicated to dermatology.⁸ The dermatology content on either of the NP certification examinations is unclear.^2,3 In the states of California, Indiana, and New York, national certification through the American Association of Nurse Practitioners or American Nurses Credentialing Center is not required for NPs to practice in their respective states.⁹

Regarding dermatologic board certification, a new graduate NP may obtain certification from the Dermatology Nurse Practitioner Certification Board with 3000 hours of general dermatology practice that may occur during normal working hours.¹⁰ These hours do not have to occur in one of the previously identified postgraduate APP training programs. The National Board of Dermatology Physician Assistants was founded in 2018 and has since dissolved. The National Board of Dermatology Physician Assistants was not accredited and required at least 3 years of training in dermatology with the same dermatologist in addition to completing a 125-question multiple-choice examination.¹¹ Of note, this examination was opposed by both the ABD and the Society for Dermatology Physician Associates.¹² A PA also may become a Diplomate Fellow with the Society of Dermatology Physician Associates after completion of 64.5 hours of online continuing education modules.⁴ Some PAs may choose to obtain a Certificate of Added Qualifications, which is a voluntary credential that helps document specialty experience and expertise in dermatology or other specialties.

In contrast, a dermatology resident physician requires nearly 11,000 to 13,000 hours of clinical training hours, which last 3 to 4 years following medical school.¹³ This training involves direct patient care under supervision in various settings, including hospitals, outpatient clinics, and surgical procedures. In addition to this clinical experience, dermatology residents must pass a 3-step certification examination process administered by the ABD.¹³ This process includes approximately 20 hours of examinations designed to assess both knowledge and practical skills. For those who wish to further specialize, additional fellowship training in areas such as pediatric dermatology, dermatopathology, or Mohs surgery may follow residency; such fellowships involve an extra 2500 to 3500 hours of training and culminate in another certification examination, further refining a resident’s expertise in a specific dermatologic field. Osteopathic physicians may opt out of the ABD 3-step pathway and obtain board certification through the American Osteopathic Board of Dermatology.¹⁴

Many of the programs we evaluated integrate APP trainees into resident education, allowing participation in equivalent didactic curricula, clinical rotations, and departmental academic activities. The salary and benefits associated with these programs are somewhat like those of resident physicians.^15,16 While most tuition-based programs were excluded from our study due to their lack of credibility and alignment with our study criteria, we identified 2 specific programs that stood out as credible despite requiring students to pay tuition. These programs demonstrated a structured and rigorous curriculum with a clear focus on comprehensive dermatologic training, meeting our standards for inclusion. These programs offer dermatologic training for graduates of NP and PA programs at a cost to the student.^15,16 The program at the Florida Atlantic University, Boca Raton, is largely online,¹⁵ and the program at the University of Miami, Florida, offers no direct clinical contact.¹⁶ These programs illustrate the variety of postgraduate dermatology curricula available nationally in comparison to resident salaries; however, they were not included in our formal analysis because they do not provide structured, in-person clinical training consistent with our inclusion criteria. Neither of these programs would enable participants to qualify for credentialing with the Dermatology Nurse Practitioner Certification Board after completion. While this study identified postgraduate training programs for APPs in dermatology advertised online, it is possible some were omitted or not advertised online.

While many of the postgraduate programs we evaluated provide unique educational opportunities for APPs, it is unknown if graduating providers are equipped to handle the care of patients with complex dermatologic needs. Regardless, the increased utilization of APPs by BCDs has been well documented over the past 2 decades.^17-20 It has been suggested that a higher ratio of APPs to dermatologists can decrease the time it takes for a patient to be seen in a clinic.^21-23 However, investigators have expressed concerns that APPs lack standardized surgical training and clinical hour requirements in the field of dermatology.²⁴ Despite these concerns, Medicare claims data show that APPs are performing advanced surgical and cosmetic procedures at increasing rates.^17,18 Other authors have questioned the cost-effectiveness of APPs, as multiple studies have shown that the number of biopsies needed to diagnose 1 case of skin cancer is higher for midlevel providers than for dermatologists.^25-27

Conclusion

With the anticipated expansion of private equity in dermatology and the growth of our Medicare-eligible population, we are likely to see increased utilization of APPs to address the shortage of BCDs.^28,29 Understanding the prelicensure and postlicensure clinical training requirements, examination hours, and extent of dermatology-focused education among APPs and BCDs can help dermatologists collaborate more effectively and ensure safe, high-quality patient care. Standardizing, improving, and providing high-quality education and promoting lifelong learning in the field of dermatology should be celebrated, and dermatologists are the skin experts best equipped to lead dermatologic education forward.

Nurse practitioners (NPs) and physician assistants (PAs) often help provide dermatologic care but lack the same mandatory specialized postgraduate training required of board-certified dermatologists (BCDs), which includes at least 3 years of dermatology-focused education in an accredited residency program in addition to an intern year of general medicine, pediatrics, or surgery. Dermatology residency is followed by a certification examination administered by the American Board of Dermatology (ABD) or the American Osteopathic Board of Dermatology, leading to board certification. Some physicians choose to do a fellowship, which typically involves an additional 1 to 2 years of postresidency subspeciality training.

Optional postgraduate dermatology training programs for advanced practice providers (APPs) have been offered by some academic institutions and private practice groups since at least 2003, including Lahey Hospital and Medical Center (Burlington, Massachusetts) as well as the University of Rochester Medical Center (Rochester, New York). Despite a lack of accreditation or standardization, the programs can be beneficial for NPs and PAs to expand their dermatologic knowledge and skills and help bridge the care gap within the specialty. Didactics often are conducted in parallel with the educational activities of the parent institution’s traditional dermatology residency program (eg, lectures, grand rounds). While these programs often are managed by practicing dermatology NPs and PAs, dermatologists also may be involved in their education with didactic instruction, curriculum development, and clinical preceptorship. 

In this cross-sectional study, we identified and evaluated 10 postgraduate dermatology training programs for APPs across the United States. With the growing number of NPs and PAs in the dermatology workforce—both in academic and private practice—it is important for BCDs to be aware of the differences in the dermatology training received in order to ensure safe and effective care is provided through supervisory or collaborative roles (depending on state independent practice laws for APPs and to be aware of the implications these programs may have on the field of dermatology.

Methods

To identify postgraduate dermatology training programs for APPs in the United States, we conducted a cross-sectional study using data obtained via a Google search of various combinations of the following terms: nurse practitioner, NP, physician assistant, PA, advance practice provider, APP, dermatology, postgraduate training, residency, and fellowship. We excluded postgraduate dermatology training programs for APPs that required tuition and did not provide a stipend, as well as programs that lacked the formal structure and credibility needed to qualify as legitimate postgraduate training. Many of the excluded programs operate in a manner that raises ethical concerns, offering pay-to-play opportunities under the guise of education. Information collected on each program included the program name, location, parent institution, program length, class size, curriculum, and any associated salary and benefits.

Results

Ten academic and private practice organizations across the United States that offer postgraduate dermatologic training programs for APPs were identified (eTable). Four (40%) programs were advertised as fellowships. Six (60%) of the programs were offered at academic medical centers, and 4 (40%) were offered by private practices. Most programs were located east of the Mississippi River, and many institutions offered instruction at 1 or more locations within the same state (eFigure). The Advanced Dermatology and Cosmetic Surgery private practice group offered training opportunities in multiple states.

Six programs required APPs to become board-certified NPs or PAs prior to enrolling. Most programs enrolled both NPs and PAs, while some only enrolled NPs (eTable). Only 1 (10%) program required NPs to be board certified as a family NP, while another (10%) recommended that applicants have experience in urgent care, emergency medicine, or trauma medicine. Lahey Hospital & Medical Center required experience as an NP in a general setting for 1 to 2 years prior to applying. No program required prior experience in the field of dermatology.

Program length varied from 6 to 24 months, and cohort size typically was limited to 1 to 2 providers (eTable). Although the exact numbers could not be ascertained, most curricula focused on medical dermatology, including clinical and didactic components, but many offered electives such as cosmetic and procedural dermatology. Two institutions (20%) required independent research. Work typically was limited to 40 hours per week, and most paid a full-time employee salary and provided benefits such as health insurance, retirement, and paid leave (eTable). Kansas Medical Clinic (Topeka, Kansas) required at least 3 years of employment in an underserved community following program completion. The Oasis Dermatology private practice group in Texas required a 1-year teaching commitment after program completion. The Advanced Dermatology and Cosmetic Surgery group offered a full-time position upon program completion.

Comment

There is a large difference in the total number of training and credentialing hours when comparing graduate school training and postgraduate credentialing of medical and osteopathic physicians compared with APPs. A new graduate physician has at least twice as many clinical hours as a PA and 10 times as many clinical hours as an NP prior to starting residency. Physicians also typically complete at least 6 times the number of hours of certification examinations compared to NPs and PAs.¹

Nurse practitioner students typically complete the 500 hours of prelicensure clinical training required for NP school in 2 to 4 years.^2,3 The amount of time required for completion is dependent on the degree and experience of the student upon program entry (eg, bachelor of science in nursing vs master of science in nursing as a terminal degree). Physician assistant students are required to complete 2000 prelicensure clinical hours, and most PA programs are 3 years in duration.⁴ Many NP and PA programs require some degree of clinical experience prior to beginning graduate education.⁵

When comparing prelicensure examinations, questions assessing dermatologic knowledge comprise approximately 6% to 10% of the total questions on the United States Medical Licensing Examination Steps 1 and 2.⁶ The Comprehensive Osteopathic Medical Licensing Examination of the United States Level 1 and Level 2-Cognitive Evaluation both have at least 5% of questions dedicated to dermatology.⁷ Approximately 5% of the questions on the Physician Assistant National Certifying Examination are dedicated to dermatology.⁸ The dermatology content on either of the NP certification examinations is unclear.^2,3 In the states of California, Indiana, and New York, national certification through the American Association of Nurse Practitioners or American Nurses Credentialing Center is not required for NPs to practice in their respective states.⁹

Regarding dermatologic board certification, a new graduate NP may obtain certification from the Dermatology Nurse Practitioner Certification Board with 3000 hours of general dermatology practice that may occur during normal working hours.¹⁰ These hours do not have to occur in one of the previously identified postgraduate APP training programs. The National Board of Dermatology Physician Assistants was founded in 2018 and has since dissolved. The National Board of Dermatology Physician Assistants was not accredited and required at least 3 years of training in dermatology with the same dermatologist in addition to completing a 125-question multiple-choice examination.¹¹ Of note, this examination was opposed by both the ABD and the Society for Dermatology Physician Associates.¹² A PA also may become a Diplomate Fellow with the Society of Dermatology Physician Associates after completion of 64.5 hours of online continuing education modules.⁴ Some PAs may choose to obtain a Certificate of Added Qualifications, which is a voluntary credential that helps document specialty experience and expertise in dermatology or other specialties.

In contrast, a dermatology resident physician requires nearly 11,000 to 13,000 hours of clinical training hours, which last 3 to 4 years following medical school.¹³ This training involves direct patient care under supervision in various settings, including hospitals, outpatient clinics, and surgical procedures. In addition to this clinical experience, dermatology residents must pass a 3-step certification examination process administered by the ABD.¹³ This process includes approximately 20 hours of examinations designed to assess both knowledge and practical skills. For those who wish to further specialize, additional fellowship training in areas such as pediatric dermatology, dermatopathology, or Mohs surgery may follow residency; such fellowships involve an extra 2500 to 3500 hours of training and culminate in another certification examination, further refining a resident’s expertise in a specific dermatologic field. Osteopathic physicians may opt out of the ABD 3-step pathway and obtain board certification through the American Osteopathic Board of Dermatology.¹⁴

Many of the programs we evaluated integrate APP trainees into resident education, allowing participation in equivalent didactic curricula, clinical rotations, and departmental academic activities. The salary and benefits associated with these programs are somewhat like those of resident physicians.^15,16 While most tuition-based programs were excluded from our study due to their lack of credibility and alignment with our study criteria, we identified 2 specific programs that stood out as credible despite requiring students to pay tuition. These programs demonstrated a structured and rigorous curriculum with a clear focus on comprehensive dermatologic training, meeting our standards for inclusion. These programs offer dermatologic training for graduates of NP and PA programs at a cost to the student.^15,16 The program at the Florida Atlantic University, Boca Raton, is largely online,¹⁵ and the program at the University of Miami, Florida, offers no direct clinical contact.¹⁶ These programs illustrate the variety of postgraduate dermatology curricula available nationally in comparison to resident salaries; however, they were not included in our formal analysis because they do not provide structured, in-person clinical training consistent with our inclusion criteria. Neither of these programs would enable participants to qualify for credentialing with the Dermatology Nurse Practitioner Certification Board after completion. While this study identified postgraduate training programs for APPs in dermatology advertised online, it is possible some were omitted or not advertised online.

While many of the postgraduate programs we evaluated provide unique educational opportunities for APPs, it is unknown if graduating providers are equipped to handle the care of patients with complex dermatologic needs. Regardless, the increased utilization of APPs by BCDs has been well documented over the past 2 decades.^17-20 It has been suggested that a higher ratio of APPs to dermatologists can decrease the time it takes for a patient to be seen in a clinic.^21-23 However, investigators have expressed concerns that APPs lack standardized surgical training and clinical hour requirements in the field of dermatology.²⁴ Despite these concerns, Medicare claims data show that APPs are performing advanced surgical and cosmetic procedures at increasing rates.^17,18 Other authors have questioned the cost-effectiveness of APPs, as multiple studies have shown that the number of biopsies needed to diagnose 1 case of skin cancer is higher for midlevel providers than for dermatologists.^25-27

Conclusion

With the anticipated expansion of private equity in dermatology and the growth of our Medicare-eligible population, we are likely to see increased utilization of APPs to address the shortage of BCDs.^28,29 Understanding the prelicensure and postlicensure clinical training requirements, examination hours, and extent of dermatology-focused education among APPs and BCDs can help dermatologists collaborate more effectively and ensure safe, high-quality patient care. Standardizing, improving, and providing high-quality education and promoting lifelong learning in the field of dermatology should be celebrated, and dermatologists are the skin experts best equipped to lead dermatologic education forward.

Nurse practitioners (NPs) and physician assistants (PAs) often help provide dermatologic care but lack the same mandatory specialized postgraduate training required of board-certified dermatologists (BCDs), which includes at least 3 years of dermatology-focused education in an accredited residency program in addition to an intern year of general medicine, pediatrics, or surgery. Dermatology residency is followed by a certification examination administered by the American Board of Dermatology (ABD) or the American Osteopathic Board of Dermatology, leading to board certification. Some physicians choose to do a fellowship, which typically involves an additional 1 to 2 years of postresidency subspeciality training.

Optional postgraduate dermatology training programs for advanced practice providers (APPs) have been offered by some academic institutions and private practice groups since at least 2003, including Lahey Hospital and Medical Center (Burlington, Massachusetts) as well as the University of Rochester Medical Center (Rochester, New York). Despite a lack of accreditation or standardization, the programs can be beneficial for NPs and PAs to expand their dermatologic knowledge and skills and help bridge the care gap within the specialty. Didactics often are conducted in parallel with the educational activities of the parent institution’s traditional dermatology residency program (eg, lectures, grand rounds). While these programs often are managed by practicing dermatology NPs and PAs, dermatologists also may be involved in their education with didactic instruction, curriculum development, and clinical preceptorship. 

In this cross-sectional study, we identified and evaluated 10 postgraduate dermatology training programs for APPs across the United States. With the growing number of NPs and PAs in the dermatology workforce—both in academic and private practice—it is important for BCDs to be aware of the differences in the dermatology training received in order to ensure safe and effective care is provided through supervisory or collaborative roles (depending on state independent practice laws for APPs and to be aware of the implications these programs may have on the field of dermatology.

Methods

To identify postgraduate dermatology training programs for APPs in the United States, we conducted a cross-sectional study using data obtained via a Google search of various combinations of the following terms: nurse practitioner, NP, physician assistant, PA, advance practice provider, APP, dermatology, postgraduate training, residency, and fellowship. We excluded postgraduate dermatology training programs for APPs that required tuition and did not provide a stipend, as well as programs that lacked the formal structure and credibility needed to qualify as legitimate postgraduate training. Many of the excluded programs operate in a manner that raises ethical concerns, offering pay-to-play opportunities under the guise of education. Information collected on each program included the program name, location, parent institution, program length, class size, curriculum, and any associated salary and benefits.

Results

Ten academic and private practice organizations across the United States that offer postgraduate dermatologic training programs for APPs were identified (eTable). Four (40%) programs were advertised as fellowships. Six (60%) of the programs were offered at academic medical centers, and 4 (40%) were offered by private practices. Most programs were located east of the Mississippi River, and many institutions offered instruction at 1 or more locations within the same state (eFigure). The Advanced Dermatology and Cosmetic Surgery private practice group offered training opportunities in multiple states.

Six programs required APPs to become board-certified NPs or PAs prior to enrolling. Most programs enrolled both NPs and PAs, while some only enrolled NPs (eTable). Only 1 (10%) program required NPs to be board certified as a family NP, while another (10%) recommended that applicants have experience in urgent care, emergency medicine, or trauma medicine. Lahey Hospital & Medical Center required experience as an NP in a general setting for 1 to 2 years prior to applying. No program required prior experience in the field of dermatology.

Program length varied from 6 to 24 months, and cohort size typically was limited to 1 to 2 providers (eTable). Although the exact numbers could not be ascertained, most curricula focused on medical dermatology, including clinical and didactic components, but many offered electives such as cosmetic and procedural dermatology. Two institutions (20%) required independent research. Work typically was limited to 40 hours per week, and most paid a full-time employee salary and provided benefits such as health insurance, retirement, and paid leave (eTable). Kansas Medical Clinic (Topeka, Kansas) required at least 3 years of employment in an underserved community following program completion. The Oasis Dermatology private practice group in Texas required a 1-year teaching commitment after program completion. The Advanced Dermatology and Cosmetic Surgery group offered a full-time position upon program completion.

Comment

There is a large difference in the total number of training and credentialing hours when comparing graduate school training and postgraduate credentialing of medical and osteopathic physicians compared with APPs. A new graduate physician has at least twice as many clinical hours as a PA and 10 times as many clinical hours as an NP prior to starting residency. Physicians also typically complete at least 6 times the number of hours of certification examinations compared to NPs and PAs.¹

Nurse practitioner students typically complete the 500 hours of prelicensure clinical training required for NP school in 2 to 4 years.^2,3 The amount of time required for completion is dependent on the degree and experience of the student upon program entry (eg, bachelor of science in nursing vs master of science in nursing as a terminal degree). Physician assistant students are required to complete 2000 prelicensure clinical hours, and most PA programs are 3 years in duration.⁴ Many NP and PA programs require some degree of clinical experience prior to beginning graduate education.⁵

When comparing prelicensure examinations, questions assessing dermatologic knowledge comprise approximately 6% to 10% of the total questions on the United States Medical Licensing Examination Steps 1 and 2.⁶ The Comprehensive Osteopathic Medical Licensing Examination of the United States Level 1 and Level 2-Cognitive Evaluation both have at least 5% of questions dedicated to dermatology.⁷ Approximately 5% of the questions on the Physician Assistant National Certifying Examination are dedicated to dermatology.⁸ The dermatology content on either of the NP certification examinations is unclear.^2,3 In the states of California, Indiana, and New York, national certification through the American Association of Nurse Practitioners or American Nurses Credentialing Center is not required for NPs to practice in their respective states.⁹

Regarding dermatologic board certification, a new graduate NP may obtain certification from the Dermatology Nurse Practitioner Certification Board with 3000 hours of general dermatology practice that may occur during normal working hours.¹⁰ These hours do not have to occur in one of the previously identified postgraduate APP training programs. The National Board of Dermatology Physician Assistants was founded in 2018 and has since dissolved. The National Board of Dermatology Physician Assistants was not accredited and required at least 3 years of training in dermatology with the same dermatologist in addition to completing a 125-question multiple-choice examination.¹¹ Of note, this examination was opposed by both the ABD and the Society for Dermatology Physician Associates.¹² A PA also may become a Diplomate Fellow with the Society of Dermatology Physician Associates after completion of 64.5 hours of online continuing education modules.⁴ Some PAs may choose to obtain a Certificate of Added Qualifications, which is a voluntary credential that helps document specialty experience and expertise in dermatology or other specialties.

In contrast, a dermatology resident physician requires nearly 11,000 to 13,000 hours of clinical training hours, which last 3 to 4 years following medical school.¹³ This training involves direct patient care under supervision in various settings, including hospitals, outpatient clinics, and surgical procedures. In addition to this clinical experience, dermatology residents must pass a 3-step certification examination process administered by the ABD.¹³ This process includes approximately 20 hours of examinations designed to assess both knowledge and practical skills. For those who wish to further specialize, additional fellowship training in areas such as pediatric dermatology, dermatopathology, or Mohs surgery may follow residency; such fellowships involve an extra 2500 to 3500 hours of training and culminate in another certification examination, further refining a resident’s expertise in a specific dermatologic field. Osteopathic physicians may opt out of the ABD 3-step pathway and obtain board certification through the American Osteopathic Board of Dermatology.¹⁴

Many of the programs we evaluated integrate APP trainees into resident education, allowing participation in equivalent didactic curricula, clinical rotations, and departmental academic activities. The salary and benefits associated with these programs are somewhat like those of resident physicians.^15,16 While most tuition-based programs were excluded from our study due to their lack of credibility and alignment with our study criteria, we identified 2 specific programs that stood out as credible despite requiring students to pay tuition. These programs demonstrated a structured and rigorous curriculum with a clear focus on comprehensive dermatologic training, meeting our standards for inclusion. These programs offer dermatologic training for graduates of NP and PA programs at a cost to the student.^15,16 The program at the Florida Atlantic University, Boca Raton, is largely online,¹⁵ and the program at the University of Miami, Florida, offers no direct clinical contact.¹⁶ These programs illustrate the variety of postgraduate dermatology curricula available nationally in comparison to resident salaries; however, they were not included in our formal analysis because they do not provide structured, in-person clinical training consistent with our inclusion criteria. Neither of these programs would enable participants to qualify for credentialing with the Dermatology Nurse Practitioner Certification Board after completion. While this study identified postgraduate training programs for APPs in dermatology advertised online, it is possible some were omitted or not advertised online.

While many of the postgraduate programs we evaluated provide unique educational opportunities for APPs, it is unknown if graduating providers are equipped to handle the care of patients with complex dermatologic needs. Regardless, the increased utilization of APPs by BCDs has been well documented over the past 2 decades.^17-20 It has been suggested that a higher ratio of APPs to dermatologists can decrease the time it takes for a patient to be seen in a clinic.^21-23 However, investigators have expressed concerns that APPs lack standardized surgical training and clinical hour requirements in the field of dermatology.²⁴ Despite these concerns, Medicare claims data show that APPs are performing advanced surgical and cosmetic procedures at increasing rates.^17,18 Other authors have questioned the cost-effectiveness of APPs, as multiple studies have shown that the number of biopsies needed to diagnose 1 case of skin cancer is higher for midlevel providers than for dermatologists.^25-27

Conclusion

With the anticipated expansion of private equity in dermatology and the growth of our Medicare-eligible population, we are likely to see increased utilization of APPs to address the shortage of BCDs.^28,29 Understanding the prelicensure and postlicensure clinical training requirements, examination hours, and extent of dermatology-focused education among APPs and BCDs can help dermatologists collaborate more effectively and ensure safe, high-quality patient care. Standardizing, improving, and providing high-quality education and promoting lifelong learning in the field of dermatology should be celebrated, and dermatologists are the skin experts best equipped to lead dermatologic education forward.

Trauma-focused evidence-based psychotherapies (TF-EBPs), including cognitive processing therapy (CPT) and prolonged exposure therapy (PE), are recommended treatments for posttraumatic stress disorder (PTSD) in clinical practice guidelines.^1-3 To increase initiation of these treatments, the US Department of Veterans Affairs (VA) used a large-scale dissemination and implementation effort to improve access to TF-EBP.^4,5 These efforts achieved modest success, increasing prevalence of TF-EBP from a handful of veterans in 2004 to an annual prevalence of 14.6% for CPT and 4.3% for PE in 2014.⁶

Throughout these efforts, qualitative studies have been used to better understand veterans’ perspectives on receiving TF-EBP care.^7-18 Barriers to initiation of and engagement in TF-EBP and PTSD care have been identified from these qualitative studies. One identified barrier was lack of knowledge—particularly lack of knowledge about what is meant by a PTSD diagnosis and available treatments.^7-10 Stigma (ie, automatic negative associations) toward mental health problems or seeking mental health care also has been identified as a barrier to initiation.^7,10-14 Perceptions of poor alignment between treatment and veteran goals, including lack of buy-in for the rationale, served as barriers to initiation and engagement.^8,15-18

Using prior qualitative work, numerous initiatives have been developed to reduce stigma, facilitate conversations about how treatment aligns with goals, and fill knowledge gaps, particularly through online resources and shared decision-making.^19,20 To better inform the state of veterans’ experiences with TF-EBP, a qualitative investigation was conducted involving veterans who recently initiated TF-EBP. Themes directly related to transitions to TF-EBP were identified; however, all veterans interviewed also described their experiences with TFEBP engagement and mental health care. Consistent with recommendations for qualitative methods, this study extends prior work on transitions to TF-EBP by describing themes with a distinct focus on the experience of engaging with TF-EBP and mental health care.^21,22

Methods

The experiences of veterans who were transitioning into TF-EBPs were collected in semistructured interviews and analyzed. The semistructured interview guide was developed and refined in consultation with both qualitative methods experts and PTSD treatment experts to ensure that 6 content domains were appropriately queried: PTSD treatment options, cultural sensitivity of treatment, PTSD treatment selection, transition criteria, beliefs about stabilization treatment, and treatment needs/preferences.

Participants were identified using the VA Corporate Data Warehouse and included post-9/11 veterans who had recently initiated CPT or PE for the first time between September 1, 2021, and September 1, 2022. More details of participant selection are available in Holder et al.²¹ From a population of 10,814 patients, stratified random sampling generated a recruitment pool of 200 veterans for further outreach. The strata were defined such that this recruitment pool had similar proportions of demographic characteristics (ie, gender, race, ethnicity) to the population of eligible veterans, equivalent distributions of time to CPT or PE initiation (ie, 33.3% < 1 year, 33.3% 1-3 years, and 33.3% > 3 years), and adequate variability in TF-EBP type (ie, 66.7% CPT, 33.3% PE). A manual chart review in the recruitment pool excluded 12 veterans who did not initiate CPT or PE, 1 veteran with evidence of current active psychosis and/or cognitive impairment that would likely preclude comprehension of study materials, and 1 who was deceased.

Eligible veterans from the recruitment pool were contacted in groups of 25. First, a recruitment letter with study information and instructions to opt-out of further contact was mailed or emailed to veterans. After 2 weeks, veterans who had not responded were contacted by phone up to 3 times. Veterans interested in participating were scheduled for a 1-time visit that included verbal consent and the qualitative interview. Metrics were established a priori to ensure an adequately diverse and inclusive sample. Specifically, a minimum number of racial and/or ethnic minority veterans (33%) and women veterans (20%) were sought. Equal distribution across the 3 categories of time from first mental health visit to CPT/PE initiation also was targeted. Throughout enrollment, recruitment efforts were adapted to meet these metrics in the emerging sample. While the goal was to generate a diverse and inclusive sample using these methods, the sample was not intended to be representative of the population.

Of the 186 eligible participants, 21 declined participation and 26 could not be reached. The targeted sample was reached after exhausting contact for 47 veterans and contacting 80 veterans for a final response rate of 40% among fully contacted veterans and 27% among veterans with any contact. The final sample included 30 veterans who received CPT or PE in VA facilities (Table).

After veterans provided verbal consent for study participation, sociodemographic information was verbally reported, and a 30- to 60-minute semistructured qualitative phone interview was recorded and transcribed. Veterans received $40 for participation. All procedures were approved by the University of California San Francisco Institutional Review Board.

Qualitative Data Analysis

Rapid analysis procedures were used to analyze qualitative data. This approach is suitable for focused, moderately structured qualitative analyses in health services research and facilitates rapid dissemination to stakeholders.²³ The qualitative analysts were 2 clinical psychologists with expertise in PTSD treatment (NH primary and RR secondary). Consistent with rapid analysis procedures, analysts prepared a templated summary (including relevant quotations) of each interview, organized by the prespecified content domains. Interviews were summarized independently, compared to ensure consistency, and discrepancies were resolved through review of the interview source materials. Individual summary templates were combined into a master analytic matrix to facilitate the identification of patterns and delineation of themes. Analysts routinely met to identify, discuss, and refine preliminary themes, revisiting source materials to reach consensus as needed.

Results

Fifteen themes were identified and organized into 2 distinct focus areas: themes directly related to the transition to TF-EBP (8 themes) and themes related to veterans’ experiences with TF-EBP and general mental health care with potential process-improvement implications (7 themes).²¹ Seven themes were identified related to experiences with TF-EBP engagement and VA mental health care. The 7 themes related to TF-EBP engagement and VA mental health care themes are summarized with exemplary quotations.

Veterans want a better understanding of psychotherapy and engaging with VA mental health. Veterans reported that they generally had a poor or “nebulous” understanding about the experience of psychotherapy. For example, veterans exhibited confusion about whether certain experiences were equivalent to participating in psychotherapy. They were sometimes unable to distinguish between interactions such as assessment, disability evaluations, peer support, and psychotherapy. One veteran described a conversation with a TFEBP therapist about prior treatment:

She [asked], have you ever been, or gone through a therapy to begin with? And I, I said, well I just chatted with somebody. And she said that’s not, that’s not therapy. So, I was like, oh, it’s not? That’s not what people do?

Veterans were surprised the VA offered a diverse range of psychotherapy interventions, rather than simply therapy. They did not realize there were different types of psychotherapy. As a result, veterans were not aware that some VA mental practitioners have specialty training and certification to provide treatment matched to specific diagnoses or needs. They thought that all clinicians could provide the same care. One veteran described their understanding:

I just figured all mental health people are mental health people. I didn’t have a better understanding of the system and all the different levels and how it plays out and specialties and things like that. Which, I guess, I should have because you have a primary care doctor, but then you have specialists in all these other different sectors that specialize in one particular area. I guess that should’ve been common sense, but it wasn’t.

Stigma was a barrier to seeking and engaging in mental health care. Veterans discovered they had to overcome stigma associated with seeking and engaging in mental health treatment. Military culture was often discussed as promoting stigma regarding mental health treatment. Specifically, veterans described that seeking treatment meant “either, I’m weak or I’m gonna be seen as weak.” In active-duty settings, the strategy for dealing with mental health symptoms was to “leave those feelings, you push ‘em aside,” an approach highly inconsistent with TF-EBP. In some cases, incorrect information about the VA and PTSD was presented as part of discharge from the military, leading to long-term skepticism of the VA and PTSD treatment. One veteran described his experience as part of a class on the VA compensation and pension assessment process for service-connected disabilities during his military discharge:

[A fellow discharging soldier asked] what about like PTSD, gettin’ rated for PTSD. I hear they take our weapons and stuff like we can’t own firearms and all that stuff. And [the instructor] was like, well, yes that’s a thing. He didn’t explain it like if you get compensated for PTSD you don’t lose your rights to carry a firearm or to have, to be able to go hunting.

Importantly, veterans often described how other identities (eg, race, ethnicity, gender, region of origin) interacted with military culture to enhance stigma. Hearing messaging from multiple sources reinforced beliefs that mental health treatment is inappropriate or is associated with weakness:

As a first-generation Italian, I was always taught keep your feelings to yourself. Never talk outside your family. Never bring up problems to other people and stuff like that. Same with the military. And then the old stigma working in [emergency medical services] and public safety, you’re weak if you get help.

The fundamentals of therapy, including rapport and flexibility, were important. Veterans valued nonspecific therapy factors, genuine empathy, building trust, being honest about treatment, personality, and rapport. These characteristics were almost universally described as particularly important:

I liked the fact that she made it personable and she cared. It wasn’t just like, here, we’re gonna start this. She explained it in the ways I could understand, not in medical terms, so to speak, but that’s what I liked about her. She really cared about what she did and helping me.

Flexibility was viewed as an asset, particularly when clinicians acknowledged veteran autonomy. A consistent example was when veterans were able to titrate trauma disclosure. One veteran described this flexible treatment experience: “She was right there in the room, she said, you know, at any time, you know, we could stop, we could debrief.”

Experiences of clinician flexibility and personalization of therapy were contrasted with experiences of overly rigid therapy. Overemphasis on protocols created barriers, often because treatment did not feel personalized. One veteran described how a clinician’s task-oriented approach interfered with their ability to engage in TF-EBP:

They listened, but it just didn’t seem like they were listening, because they really wanted to stay on task… So, I felt like if the person was more concerned, or more sympathetic to the things that was also going on in my life at that present time, I think I would’ve felt more comfortable talking about what was the PTSD part, too.

Veterans valued shared decision-making prior to TF-EBP initiation. Veterans typically described being involved in a shared decision-making process prior to initiating TF-EBP. During these sessions, clinicians discussed treatment options and provided veterans with a variety of materials describing treatments (eg, pamphlets, websites, videos, statistics). Most veterans appreciated being able to reflect on and discuss treatment options with their clinicians. Being given time in and out of session to review was viewed as valuable and increased confidence in treatment choice. One veteran described their experience:

I was given the information, you know, they gave me handouts, PDFs, whatever was available, and let me read over it. I didn’t have to choose anything right then and there, you know, they let me sleep on it. And I got back to them after some thought.

However, some veterans felt overwhelmed by being presented with too much information and did not believe they knew enough to make a final treatment decision. One veteran described being asked to contribute to the treatment decision:

I definitely asked [the clinician] to weigh in on maybe what he thought was best, because—I mean, I don’t know… I’m not necessarily sure I know what I think is best. I think we’re just lucky I’m here, so if you can give me a solid and help me out here by telling me just based on what I’ve said to you and the things that I’ve gone through, what do you think?

Veterans who perceived that their treatment preferences were respected had a positive outlook on TF-EBP. As part of the shared-decision making process, veterans typically described being given choices among PTSD treatments. One way that preferences were respected was through clinicians tailoring treatment descriptions to a veteran’s unique symptoms, experiences, and values. In these cases, clinicians observed specific concerns and clearly linked treatment principles to those concerns. For example, one veteran described their clinician’s recommendation for PE: “The hardest thing for me is to do the normal things like grocery store or getting on a train or anything like that. And so, he suggested that [PE] would be a good idea.”

In other cases, veterans wanted the highest quality of treatment rather than a match between treatment principles and the veteran’s presentation, goals, or strengths. These veterans wanted the best treatment available for PTSD and valued research support, recommendations from clinical practice guidelines, or clinician confidence in the effectiveness of the treatment. One veteran described this perspective:

I just wanted to be able to really tackle it in the best way possible and in the most like aggressive way possible. And it seemed like PE really was going to, they said that it’s a difficult type of therapy, but I really just wanted to kind of do the best that I could to eradicate some of the issues that I was having.

When veterans perceived a lack of respect for their preferences, they were hesitant about TF-EBP. For some veterans, a generic pitch for a TF-EBP was detrimental in the absence of the personal connection between the treatment and their own symptoms, goals, or strengths. These veterans did not question whether the treatment was effective in general but did question whether the treatment was best for them. One veteran described the contrast between their clinician’s perspective and their own.

I felt like they felt very comfortable, very confident in [CPT] being the program, because it was comfortable for them. Because they did it several times. And maybe they had a lot of success with other individuals... but they were very comfortable with that one, as a provider, more than: Is this the best fit for [me]?

Some veterans perceived little concern for their preferences and a lack of choice in available treatments, which tended to perpetuate negative perceptions of TFEBP. These veterans described their lack of choices with frustration. Alternatives to TFEBP were described by these veterans as so undesirable that they did not believe they had a real choice:

[CPT] was the only decision they had. There was nothing else for PTSD. They didn’t offer anything else. So, I mean it wasn’t a decision. It was either … take treatment or don’t take treatment at all… Actually, I need to correct myself. So, there were 2 options, group therapy or CPT. I forgot about that. I’m not a big group guy so I chose the CPT.

Another veteran was offered a choice between therapeutic approaches, but all were delivered via telehealth (consistent with the transition to virtual services during the COVID-19 pandemic). For this veteran, not only was the distinction between approaches unclear, but the choice between approaches was unimportant compared to the mode of delivery.

This happened during COVID-19 and VA stopped seeing anybody physically, face-to-face. So my only option for therapy was [telehealth]… There was like 3 of them, and I tried to figure out, you know, from the layperson’s perspective, like: I don’t know which one to go with.

Veterans wanted to be asked about their cultural identity. Veterans valued when clinicians asked questions about cultural identity as part of their mental health treatment and listened to their cultural context. Cultural identity factors extended beyond factors such as race, ethnicity, gender, and sexual orientation to religion, military culture, and regionality. Veterans often described situations where they wished clinicians would ask the question or initiate conversations about culture. A veteran highlighted the importance of their faith but noted that it was a taboo topic. Their clinician did not say “we don’t go there,” but they “never dove into it either.” Another veteran expressed a desire for their clinician to ask questions about experiences in the National Guard and as an African American veteran:

If a provider was to say like: Oh, you know, it’s a stressful situation being a part of the military, being in the National Guard. You know, just asking questions about that. I think that would really go a long way… Being African American was difficult as well. And more so because of my region, I think… I felt like it would probably be an uncomfortable subject to speak on… I mean, it wasn’t anything that my providers necessarily did, it was more so just because it wasn’t brought up.

One common area of concern for veterans was a match between veteran and therapist demographics. When asked about how their cultural identity influenced treatment, several veterans described the relevance of therapist match. Much like questions about their own cultural identity, veterans valued being asked about identity preferences in clinicians (eg, gender or race matching), rather than having to bring up the preference themselves. One veteran described relief at this question being asked directly: “I was relieved when she had asked [whether I wanted a male or female clinician] primarily because I was going to ask that or bring that up somehow. But her asking that before me was a weight off my shoulders.”

Discussing cultural identity through treatment strengthened veterans’ engagement in therapy. Many veterans appreciated when analogies used in therapy were relevant to their cultural experiences and when clinicians understood their culture (eg, military culture, race, ethnicity, religious beliefs, sexual orientation). One veteran described how their clinician understood military culture and made connections between military culture and the rationale for TF-EBP, which strengthened the veteran’s buy-in for the treatment and alliance with the clinician:

At the beginning when she was explaining PTSD, and I remember she said that your brain needed to think this way when you were in the military because it was a way of protecting and surviving, so your brain was doing that in order for you to survive in whatever areas you were because there was danger. So, your brain had you thinking that way. But now, you’re not in those situations anymore. You’re not in danger. You’re not in the military, but your brain is still thinking you are, and that’s what PTSD generally does to you.

Specific elements of TF-EBP also provided opportunities to discuss and integrate important aspects of identity. This is accomplished in PE by assigning relevant in vivo exercises. In CPT, “connecting the dots” on how prior experiences influenced trauma-related stuck points achieved this element. One veteran described their experience with a clinician who was comfortable discussing the veteran’s sexual orientation and recognized the impacts of prior trauma on intimacy:

They’re very different, and there’s a lot of things that can be accepted in gay relationships that are not in straight ones. With all that said, I think [the PE therapist] did a fantastic job being not—like never once did she laugh or make an uncomfortable comment or say she didn’t wanna talk about something when like part of the reason I wanted to get into therapy is that my partner and I weren’t having sex unless I used alcohol.

Discussion

As part of a larger national qualitative investigation of the experiences of veterans who recently initiated TF-EBP, veterans discussed their experiences with therapy and mental health care that have important implications for continued process improvement.²¹ Three key areas for continued process improvement were identified: (1) providing information about the diverse range of mental health care services at the VA and the implications of this continuum of care; (2) consideration of veteran preferences in treatment decision-making, including the importance of perceived choice; and (3) incorporating cultural assessment and cultural responsiveness into case conceptualization and treatment.

One area of process improvement identified was increasing knowledge about different types of psychotherapy and the continuum of care available at the VA. Veterans in this study confused or conflated participating in psychotherapy with talking about mental health symptoms with a clinician (eg, assessment, disability evaluation). They were sometimes surprised that psychotherapy is an umbrella term referring to a variety of different modalities. The downstream impact of these misunderstandings was a perception of VA mental health care as nebulous. Veterans were surprised that all mental health practitioners were unable to provide the same care. Confusion may have been compounded by highly variable referral processes across VA.²⁴ To address this, clinicians have developed local educational resources and handouts for both veterans and referring clinicians from nonmental health and general mental health specialties.²⁵ Given the variability in referral processes both between and within VA medical centers, national dissemination of these educational materials may be more difficult compared to materials for TF-EBPs.²⁴ The VA started to use behavioral health interdisciplinary program (BHIP) teams, which are designed to be clinical homes for veterans connected with a central clinician who can explain and coordinate their mental health care as well as bring more consistency to the referral process.²⁶ The ongoing transition toward the BHIP model of mental health care at VA may provide the opportunity to consolidate and integrate knowledge about the VA approach to mental health care, potentially filling knowledge gaps.

A second area of process improvement focused on the shared decision-making process. Consistent with mental health initiatives, veterans generally believed they had received sufficient information about TF-EBP and engaged in shared decision-making with clinicians.^20,27 Veterans were given educational materials to review and had the opportunity to discuss these materials with clinicians. However, veterans described variability in the success of shared decision-making. Although veterans valued receiving accurate, comprehensible information to support treatment decisions, some preferred to defer to clinicians’ expertise regarding which treatment to pursue. While these veterans valued information, they also valued the expertise of clinicians in explaining why specific treatments would be beneficial. A key contributor to veterans satisfaction was assessing how veterans wanted to engage in the decision-making process and respecting those preferences.²⁸ Veterans approached shared decision-making differently, from making decisions independently after receiving information to relying solely on clinician recommendation. The process was most successful when clinicians articulated how their recommended treatment aligned with a veteran’s preferences, including recommendations based on specific values (eg, personalized match vs being the best). Another important consideration is ensuring veterans know they can receive a variety of different types of mental health services available in different modalities (eg, virtual vs in-person; group vs individual). When veterans did not perceive choice in treatment aspects important to them (typically despite having choices), they were less satisfied with their TF-EBP experience.

A final area of process improvement identified involves how therapists address important aspects of culture. Veterans often described mental health stigma coming from intersecting cultural identities and expressed appreciation when therapists helped them recognize the impact of these beliefs on treatment. Some veterans did not discuss important aspects of their identity with clinicians, including race/ethnicity, religion, and military culture. Veterans did not report negative interactions with clinicians or experiences suggesting it was inappropriate to discuss identity; however, they were reluctant to independently raise these identity factors. Strategies such as the ADDRESSING framework, a mnemonic acronym that describes a series of potentially relevant characteristics, can help clinicians comprehensively consider different aspects that may be relevant to veterans, modeling that discussion of relevant these characteristics is welcome in TF-EBP.²⁹ Veterans reported that making culturally relevant connections enhanced the TF-EBP experience, most commonly with military culture. These data support that TF-EBP delivery with attention to culture should be an integrated part of treatment, supporting engagement and therapeutic alliance.³⁰ The VA National Center for PTSD consultation program is a resource to support clinicians in assessing and incorporating relevant aspects of cultural identity.³¹ For example, the National Center for PTSD provides a guide for using case conceptualization to address patient reactions to race-based violence during PTSD treatment.³² Both manualized design and therapist certification training can reinforce that assessing and attending to case conceptualization (including identity factors) is an integral component of TF-EBP.^33,34

Limitations

While the current study has numerous strengths (eg, national veteran sampling, robust qualitative methods), results should be considered within the context of study limitations. First, veteran participants all received TF-EBP, and the perspectives of veterans who never initiate TF-EBP may differ. Despite the strong sampling approach, the study design is not intended to be generalizable to all veterans receiving TF-EBP for PTSD. Qualitative analysis yielded 15 themes, described in this study and prior research, consistent with recommendations.^21,22 This approach allows rich description of distinct focus areas that would not be possible in a single manuscript. Nonetheless, all veterans interviewed described their experiences in TF-EBP and general mental health care, the focus of the semistructured interview guide was on the experience of transitioning from other treatment to TF-EBP.

Conclusion

This study describes themes related to general mental health and TF-EBP process improvement as part of a larger study on transitions in PTSD care.^21,22 Veterans valued the fundamentals of therapy, including rapport and flexibility. Treatment-specific rapport (eg, pointing out treatment progress and effort in completing treatment components) and flexibility within the context of fidelity (ie, personalizing treatment while maintaining core treatment elements) may be most effective at engaging veterans in recommended PTSD treatments.^18,34 In addition to successes, themes suggest multiple opportunities for process improvement. Ongoing VA initiatives and priorities (ie, BHIP, shared decision-making, consultation services) aim to improve processes consistent with veteran recommendations. Future research is needed to evaluate the success of these and other programs to optimize access to and engagement in recommended PTSD treatments.

Trauma-focused evidence-based psychotherapies (TF-EBPs), including cognitive processing therapy (CPT) and prolonged exposure therapy (PE), are recommended treatments for posttraumatic stress disorder (PTSD) in clinical practice guidelines.^1-3 To increase initiation of these treatments, the US Department of Veterans Affairs (VA) used a large-scale dissemination and implementation effort to improve access to TF-EBP.^4,5 These efforts achieved modest success, increasing prevalence of TF-EBP from a handful of veterans in 2004 to an annual prevalence of 14.6% for CPT and 4.3% for PE in 2014.⁶

Throughout these efforts, qualitative studies have been used to better understand veterans’ perspectives on receiving TF-EBP care.^7-18 Barriers to initiation of and engagement in TF-EBP and PTSD care have been identified from these qualitative studies. One identified barrier was lack of knowledge—particularly lack of knowledge about what is meant by a PTSD diagnosis and available treatments.^7-10 Stigma (ie, automatic negative associations) toward mental health problems or seeking mental health care also has been identified as a barrier to initiation.^7,10-14 Perceptions of poor alignment between treatment and veteran goals, including lack of buy-in for the rationale, served as barriers to initiation and engagement.^8,15-18

Using prior qualitative work, numerous initiatives have been developed to reduce stigma, facilitate conversations about how treatment aligns with goals, and fill knowledge gaps, particularly through online resources and shared decision-making.^19,20 To better inform the state of veterans’ experiences with TF-EBP, a qualitative investigation was conducted involving veterans who recently initiated TF-EBP. Themes directly related to transitions to TF-EBP were identified; however, all veterans interviewed also described their experiences with TFEBP engagement and mental health care. Consistent with recommendations for qualitative methods, this study extends prior work on transitions to TF-EBP by describing themes with a distinct focus on the experience of engaging with TF-EBP and mental health care.^21,22

Methods

The experiences of veterans who were transitioning into TF-EBPs were collected in semistructured interviews and analyzed. The semistructured interview guide was developed and refined in consultation with both qualitative methods experts and PTSD treatment experts to ensure that 6 content domains were appropriately queried: PTSD treatment options, cultural sensitivity of treatment, PTSD treatment selection, transition criteria, beliefs about stabilization treatment, and treatment needs/preferences.

Participants were identified using the VA Corporate Data Warehouse and included post-9/11 veterans who had recently initiated CPT or PE for the first time between September 1, 2021, and September 1, 2022. More details of participant selection are available in Holder et al.²¹ From a population of 10,814 patients, stratified random sampling generated a recruitment pool of 200 veterans for further outreach. The strata were defined such that this recruitment pool had similar proportions of demographic characteristics (ie, gender, race, ethnicity) to the population of eligible veterans, equivalent distributions of time to CPT or PE initiation (ie, 33.3% < 1 year, 33.3% 1-3 years, and 33.3% > 3 years), and adequate variability in TF-EBP type (ie, 66.7% CPT, 33.3% PE). A manual chart review in the recruitment pool excluded 12 veterans who did not initiate CPT or PE, 1 veteran with evidence of current active psychosis and/or cognitive impairment that would likely preclude comprehension of study materials, and 1 who was deceased.

Eligible veterans from the recruitment pool were contacted in groups of 25. First, a recruitment letter with study information and instructions to opt-out of further contact was mailed or emailed to veterans. After 2 weeks, veterans who had not responded were contacted by phone up to 3 times. Veterans interested in participating were scheduled for a 1-time visit that included verbal consent and the qualitative interview. Metrics were established a priori to ensure an adequately diverse and inclusive sample. Specifically, a minimum number of racial and/or ethnic minority veterans (33%) and women veterans (20%) were sought. Equal distribution across the 3 categories of time from first mental health visit to CPT/PE initiation also was targeted. Throughout enrollment, recruitment efforts were adapted to meet these metrics in the emerging sample. While the goal was to generate a diverse and inclusive sample using these methods, the sample was not intended to be representative of the population.

Of the 186 eligible participants, 21 declined participation and 26 could not be reached. The targeted sample was reached after exhausting contact for 47 veterans and contacting 80 veterans for a final response rate of 40% among fully contacted veterans and 27% among veterans with any contact. The final sample included 30 veterans who received CPT or PE in VA facilities (Table).

After veterans provided verbal consent for study participation, sociodemographic information was verbally reported, and a 30- to 60-minute semistructured qualitative phone interview was recorded and transcribed. Veterans received $40 for participation. All procedures were approved by the University of California San Francisco Institutional Review Board.

Qualitative Data Analysis

Rapid analysis procedures were used to analyze qualitative data. This approach is suitable for focused, moderately structured qualitative analyses in health services research and facilitates rapid dissemination to stakeholders.²³ The qualitative analysts were 2 clinical psychologists with expertise in PTSD treatment (NH primary and RR secondary). Consistent with rapid analysis procedures, analysts prepared a templated summary (including relevant quotations) of each interview, organized by the prespecified content domains. Interviews were summarized independently, compared to ensure consistency, and discrepancies were resolved through review of the interview source materials. Individual summary templates were combined into a master analytic matrix to facilitate the identification of patterns and delineation of themes. Analysts routinely met to identify, discuss, and refine preliminary themes, revisiting source materials to reach consensus as needed.

Results

Fifteen themes were identified and organized into 2 distinct focus areas: themes directly related to the transition to TF-EBP (8 themes) and themes related to veterans’ experiences with TF-EBP and general mental health care with potential process-improvement implications (7 themes).²¹ Seven themes were identified related to experiences with TF-EBP engagement and VA mental health care. The 7 themes related to TF-EBP engagement and VA mental health care themes are summarized with exemplary quotations.

Veterans want a better understanding of psychotherapy and engaging with VA mental health. Veterans reported that they generally had a poor or “nebulous” understanding about the experience of psychotherapy. For example, veterans exhibited confusion about whether certain experiences were equivalent to participating in psychotherapy. They were sometimes unable to distinguish between interactions such as assessment, disability evaluations, peer support, and psychotherapy. One veteran described a conversation with a TFEBP therapist about prior treatment:

She [asked], have you ever been, or gone through a therapy to begin with? And I, I said, well I just chatted with somebody. And she said that’s not, that’s not therapy. So, I was like, oh, it’s not? That’s not what people do?

Veterans were surprised the VA offered a diverse range of psychotherapy interventions, rather than simply therapy. They did not realize there were different types of psychotherapy. As a result, veterans were not aware that some VA mental practitioners have specialty training and certification to provide treatment matched to specific diagnoses or needs. They thought that all clinicians could provide the same care. One veteran described their understanding:

I just figured all mental health people are mental health people. I didn’t have a better understanding of the system and all the different levels and how it plays out and specialties and things like that. Which, I guess, I should have because you have a primary care doctor, but then you have specialists in all these other different sectors that specialize in one particular area. I guess that should’ve been common sense, but it wasn’t.

Stigma was a barrier to seeking and engaging in mental health care. Veterans discovered they had to overcome stigma associated with seeking and engaging in mental health treatment. Military culture was often discussed as promoting stigma regarding mental health treatment. Specifically, veterans described that seeking treatment meant “either, I’m weak or I’m gonna be seen as weak.” In active-duty settings, the strategy for dealing with mental health symptoms was to “leave those feelings, you push ‘em aside,” an approach highly inconsistent with TF-EBP. In some cases, incorrect information about the VA and PTSD was presented as part of discharge from the military, leading to long-term skepticism of the VA and PTSD treatment. One veteran described his experience as part of a class on the VA compensation and pension assessment process for service-connected disabilities during his military discharge:

[A fellow discharging soldier asked] what about like PTSD, gettin’ rated for PTSD. I hear they take our weapons and stuff like we can’t own firearms and all that stuff. And [the instructor] was like, well, yes that’s a thing. He didn’t explain it like if you get compensated for PTSD you don’t lose your rights to carry a firearm or to have, to be able to go hunting.

Importantly, veterans often described how other identities (eg, race, ethnicity, gender, region of origin) interacted with military culture to enhance stigma. Hearing messaging from multiple sources reinforced beliefs that mental health treatment is inappropriate or is associated with weakness:

As a first-generation Italian, I was always taught keep your feelings to yourself. Never talk outside your family. Never bring up problems to other people and stuff like that. Same with the military. And then the old stigma working in [emergency medical services] and public safety, you’re weak if you get help.

The fundamentals of therapy, including rapport and flexibility, were important. Veterans valued nonspecific therapy factors, genuine empathy, building trust, being honest about treatment, personality, and rapport. These characteristics were almost universally described as particularly important:

I liked the fact that she made it personable and she cared. It wasn’t just like, here, we’re gonna start this. She explained it in the ways I could understand, not in medical terms, so to speak, but that’s what I liked about her. She really cared about what she did and helping me.

Flexibility was viewed as an asset, particularly when clinicians acknowledged veteran autonomy. A consistent example was when veterans were able to titrate trauma disclosure. One veteran described this flexible treatment experience: “She was right there in the room, she said, you know, at any time, you know, we could stop, we could debrief.”

Experiences of clinician flexibility and personalization of therapy were contrasted with experiences of overly rigid therapy. Overemphasis on protocols created barriers, often because treatment did not feel personalized. One veteran described how a clinician’s task-oriented approach interfered with their ability to engage in TF-EBP:

They listened, but it just didn’t seem like they were listening, because they really wanted to stay on task… So, I felt like if the person was more concerned, or more sympathetic to the things that was also going on in my life at that present time, I think I would’ve felt more comfortable talking about what was the PTSD part, too.

Veterans valued shared decision-making prior to TF-EBP initiation. Veterans typically described being involved in a shared decision-making process prior to initiating TF-EBP. During these sessions, clinicians discussed treatment options and provided veterans with a variety of materials describing treatments (eg, pamphlets, websites, videos, statistics). Most veterans appreciated being able to reflect on and discuss treatment options with their clinicians. Being given time in and out of session to review was viewed as valuable and increased confidence in treatment choice. One veteran described their experience:

I was given the information, you know, they gave me handouts, PDFs, whatever was available, and let me read over it. I didn’t have to choose anything right then and there, you know, they let me sleep on it. And I got back to them after some thought.

However, some veterans felt overwhelmed by being presented with too much information and did not believe they knew enough to make a final treatment decision. One veteran described being asked to contribute to the treatment decision:

I definitely asked [the clinician] to weigh in on maybe what he thought was best, because—I mean, I don’t know… I’m not necessarily sure I know what I think is best. I think we’re just lucky I’m here, so if you can give me a solid and help me out here by telling me just based on what I’ve said to you and the things that I’ve gone through, what do you think?

Veterans who perceived that their treatment preferences were respected had a positive outlook on TF-EBP. As part of the shared-decision making process, veterans typically described being given choices among PTSD treatments. One way that preferences were respected was through clinicians tailoring treatment descriptions to a veteran’s unique symptoms, experiences, and values. In these cases, clinicians observed specific concerns and clearly linked treatment principles to those concerns. For example, one veteran described their clinician’s recommendation for PE: “The hardest thing for me is to do the normal things like grocery store or getting on a train or anything like that. And so, he suggested that [PE] would be a good idea.”

In other cases, veterans wanted the highest quality of treatment rather than a match between treatment principles and the veteran’s presentation, goals, or strengths. These veterans wanted the best treatment available for PTSD and valued research support, recommendations from clinical practice guidelines, or clinician confidence in the effectiveness of the treatment. One veteran described this perspective:

I just wanted to be able to really tackle it in the best way possible and in the most like aggressive way possible. And it seemed like PE really was going to, they said that it’s a difficult type of therapy, but I really just wanted to kind of do the best that I could to eradicate some of the issues that I was having.

When veterans perceived a lack of respect for their preferences, they were hesitant about TF-EBP. For some veterans, a generic pitch for a TF-EBP was detrimental in the absence of the personal connection between the treatment and their own symptoms, goals, or strengths. These veterans did not question whether the treatment was effective in general but did question whether the treatment was best for them. One veteran described the contrast between their clinician’s perspective and their own.

I felt like they felt very comfortable, very confident in [CPT] being the program, because it was comfortable for them. Because they did it several times. And maybe they had a lot of success with other individuals... but they were very comfortable with that one, as a provider, more than: Is this the best fit for [me]?

Some veterans perceived little concern for their preferences and a lack of choice in available treatments, which tended to perpetuate negative perceptions of TFEBP. These veterans described their lack of choices with frustration. Alternatives to TFEBP were described by these veterans as so undesirable that they did not believe they had a real choice:

[CPT] was the only decision they had. There was nothing else for PTSD. They didn’t offer anything else. So, I mean it wasn’t a decision. It was either … take treatment or don’t take treatment at all… Actually, I need to correct myself. So, there were 2 options, group therapy or CPT. I forgot about that. I’m not a big group guy so I chose the CPT.

Another veteran was offered a choice between therapeutic approaches, but all were delivered via telehealth (consistent with the transition to virtual services during the COVID-19 pandemic). For this veteran, not only was the distinction between approaches unclear, but the choice between approaches was unimportant compared to the mode of delivery.

This happened during COVID-19 and VA stopped seeing anybody physically, face-to-face. So my only option for therapy was [telehealth]… There was like 3 of them, and I tried to figure out, you know, from the layperson’s perspective, like: I don’t know which one to go with.

Veterans wanted to be asked about their cultural identity. Veterans valued when clinicians asked questions about cultural identity as part of their mental health treatment and listened to their cultural context. Cultural identity factors extended beyond factors such as race, ethnicity, gender, and sexual orientation to religion, military culture, and regionality. Veterans often described situations where they wished clinicians would ask the question or initiate conversations about culture. A veteran highlighted the importance of their faith but noted that it was a taboo topic. Their clinician did not say “we don’t go there,” but they “never dove into it either.” Another veteran expressed a desire for their clinician to ask questions about experiences in the National Guard and as an African American veteran:

If a provider was to say like: Oh, you know, it’s a stressful situation being a part of the military, being in the National Guard. You know, just asking questions about that. I think that would really go a long way… Being African American was difficult as well. And more so because of my region, I think… I felt like it would probably be an uncomfortable subject to speak on… I mean, it wasn’t anything that my providers necessarily did, it was more so just because it wasn’t brought up.

One common area of concern for veterans was a match between veteran and therapist demographics. When asked about how their cultural identity influenced treatment, several veterans described the relevance of therapist match. Much like questions about their own cultural identity, veterans valued being asked about identity preferences in clinicians (eg, gender or race matching), rather than having to bring up the preference themselves. One veteran described relief at this question being asked directly: “I was relieved when she had asked [whether I wanted a male or female clinician] primarily because I was going to ask that or bring that up somehow. But her asking that before me was a weight off my shoulders.”

Discussing cultural identity through treatment strengthened veterans’ engagement in therapy. Many veterans appreciated when analogies used in therapy were relevant to their cultural experiences and when clinicians understood their culture (eg, military culture, race, ethnicity, religious beliefs, sexual orientation). One veteran described how their clinician understood military culture and made connections between military culture and the rationale for TF-EBP, which strengthened the veteran’s buy-in for the treatment and alliance with the clinician:

At the beginning when she was explaining PTSD, and I remember she said that your brain needed to think this way when you were in the military because it was a way of protecting and surviving, so your brain was doing that in order for you to survive in whatever areas you were because there was danger. So, your brain had you thinking that way. But now, you’re not in those situations anymore. You’re not in danger. You’re not in the military, but your brain is still thinking you are, and that’s what PTSD generally does to you.

Specific elements of TF-EBP also provided opportunities to discuss and integrate important aspects of identity. This is accomplished in PE by assigning relevant in vivo exercises. In CPT, “connecting the dots” on how prior experiences influenced trauma-related stuck points achieved this element. One veteran described their experience with a clinician who was comfortable discussing the veteran’s sexual orientation and recognized the impacts of prior trauma on intimacy:

They’re very different, and there’s a lot of things that can be accepted in gay relationships that are not in straight ones. With all that said, I think [the PE therapist] did a fantastic job being not—like never once did she laugh or make an uncomfortable comment or say she didn’t wanna talk about something when like part of the reason I wanted to get into therapy is that my partner and I weren’t having sex unless I used alcohol.

Discussion

As part of a larger national qualitative investigation of the experiences of veterans who recently initiated TF-EBP, veterans discussed their experiences with therapy and mental health care that have important implications for continued process improvement.²¹ Three key areas for continued process improvement were identified: (1) providing information about the diverse range of mental health care services at the VA and the implications of this continuum of care; (2) consideration of veteran preferences in treatment decision-making, including the importance of perceived choice; and (3) incorporating cultural assessment and cultural responsiveness into case conceptualization and treatment.

One area of process improvement identified was increasing knowledge about different types of psychotherapy and the continuum of care available at the VA. Veterans in this study confused or conflated participating in psychotherapy with talking about mental health symptoms with a clinician (eg, assessment, disability evaluation). They were sometimes surprised that psychotherapy is an umbrella term referring to a variety of different modalities. The downstream impact of these misunderstandings was a perception of VA mental health care as nebulous. Veterans were surprised that all mental health practitioners were unable to provide the same care. Confusion may have been compounded by highly variable referral processes across VA.²⁴ To address this, clinicians have developed local educational resources and handouts for both veterans and referring clinicians from nonmental health and general mental health specialties.²⁵ Given the variability in referral processes both between and within VA medical centers, national dissemination of these educational materials may be more difficult compared to materials for TF-EBPs.²⁴ The VA started to use behavioral health interdisciplinary program (BHIP) teams, which are designed to be clinical homes for veterans connected with a central clinician who can explain and coordinate their mental health care as well as bring more consistency to the referral process.²⁶ The ongoing transition toward the BHIP model of mental health care at VA may provide the opportunity to consolidate and integrate knowledge about the VA approach to mental health care, potentially filling knowledge gaps.

A second area of process improvement focused on the shared decision-making process. Consistent with mental health initiatives, veterans generally believed they had received sufficient information about TF-EBP and engaged in shared decision-making with clinicians.^20,27 Veterans were given educational materials to review and had the opportunity to discuss these materials with clinicians. However, veterans described variability in the success of shared decision-making. Although veterans valued receiving accurate, comprehensible information to support treatment decisions, some preferred to defer to clinicians’ expertise regarding which treatment to pursue. While these veterans valued information, they also valued the expertise of clinicians in explaining why specific treatments would be beneficial. A key contributor to veterans satisfaction was assessing how veterans wanted to engage in the decision-making process and respecting those preferences.²⁸ Veterans approached shared decision-making differently, from making decisions independently after receiving information to relying solely on clinician recommendation. The process was most successful when clinicians articulated how their recommended treatment aligned with a veteran’s preferences, including recommendations based on specific values (eg, personalized match vs being the best). Another important consideration is ensuring veterans know they can receive a variety of different types of mental health services available in different modalities (eg, virtual vs in-person; group vs individual). When veterans did not perceive choice in treatment aspects important to them (typically despite having choices), they were less satisfied with their TF-EBP experience.

A final area of process improvement identified involves how therapists address important aspects of culture. Veterans often described mental health stigma coming from intersecting cultural identities and expressed appreciation when therapists helped them recognize the impact of these beliefs on treatment. Some veterans did not discuss important aspects of their identity with clinicians, including race/ethnicity, religion, and military culture. Veterans did not report negative interactions with clinicians or experiences suggesting it was inappropriate to discuss identity; however, they were reluctant to independently raise these identity factors. Strategies such as the ADDRESSING framework, a mnemonic acronym that describes a series of potentially relevant characteristics, can help clinicians comprehensively consider different aspects that may be relevant to veterans, modeling that discussion of relevant these characteristics is welcome in TF-EBP.²⁹ Veterans reported that making culturally relevant connections enhanced the TF-EBP experience, most commonly with military culture. These data support that TF-EBP delivery with attention to culture should be an integrated part of treatment, supporting engagement and therapeutic alliance.³⁰ The VA National Center for PTSD consultation program is a resource to support clinicians in assessing and incorporating relevant aspects of cultural identity.³¹ For example, the National Center for PTSD provides a guide for using case conceptualization to address patient reactions to race-based violence during PTSD treatment.³² Both manualized design and therapist certification training can reinforce that assessing and attending to case conceptualization (including identity factors) is an integral component of TF-EBP.^33,34

Limitations

While the current study has numerous strengths (eg, national veteran sampling, robust qualitative methods), results should be considered within the context of study limitations. First, veteran participants all received TF-EBP, and the perspectives of veterans who never initiate TF-EBP may differ. Despite the strong sampling approach, the study design is not intended to be generalizable to all veterans receiving TF-EBP for PTSD. Qualitative analysis yielded 15 themes, described in this study and prior research, consistent with recommendations.^21,22 This approach allows rich description of distinct focus areas that would not be possible in a single manuscript. Nonetheless, all veterans interviewed described their experiences in TF-EBP and general mental health care, the focus of the semistructured interview guide was on the experience of transitioning from other treatment to TF-EBP.

Conclusion

This study describes themes related to general mental health and TF-EBP process improvement as part of a larger study on transitions in PTSD care.^21,22 Veterans valued the fundamentals of therapy, including rapport and flexibility. Treatment-specific rapport (eg, pointing out treatment progress and effort in completing treatment components) and flexibility within the context of fidelity (ie, personalizing treatment while maintaining core treatment elements) may be most effective at engaging veterans in recommended PTSD treatments.^18,34 In addition to successes, themes suggest multiple opportunities for process improvement. Ongoing VA initiatives and priorities (ie, BHIP, shared decision-making, consultation services) aim to improve processes consistent with veteran recommendations. Future research is needed to evaluate the success of these and other programs to optimize access to and engagement in recommended PTSD treatments.

Trauma-focused evidence-based psychotherapies (TF-EBPs), including cognitive processing therapy (CPT) and prolonged exposure therapy (PE), are recommended treatments for posttraumatic stress disorder (PTSD) in clinical practice guidelines.^1-3 To increase initiation of these treatments, the US Department of Veterans Affairs (VA) used a large-scale dissemination and implementation effort to improve access to TF-EBP.^4,5 These efforts achieved modest success, increasing prevalence of TF-EBP from a handful of veterans in 2004 to an annual prevalence of 14.6% for CPT and 4.3% for PE in 2014.⁶

Throughout these efforts, qualitative studies have been used to better understand veterans’ perspectives on receiving TF-EBP care.^7-18 Barriers to initiation of and engagement in TF-EBP and PTSD care have been identified from these qualitative studies. One identified barrier was lack of knowledge—particularly lack of knowledge about what is meant by a PTSD diagnosis and available treatments.^7-10 Stigma (ie, automatic negative associations) toward mental health problems or seeking mental health care also has been identified as a barrier to initiation.^7,10-14 Perceptions of poor alignment between treatment and veteran goals, including lack of buy-in for the rationale, served as barriers to initiation and engagement.^8,15-18

Using prior qualitative work, numerous initiatives have been developed to reduce stigma, facilitate conversations about how treatment aligns with goals, and fill knowledge gaps, particularly through online resources and shared decision-making.^19,20 To better inform the state of veterans’ experiences with TF-EBP, a qualitative investigation was conducted involving veterans who recently initiated TF-EBP. Themes directly related to transitions to TF-EBP were identified; however, all veterans interviewed also described their experiences with TFEBP engagement and mental health care. Consistent with recommendations for qualitative methods, this study extends prior work on transitions to TF-EBP by describing themes with a distinct focus on the experience of engaging with TF-EBP and mental health care.^21,22

Methods

The experiences of veterans who were transitioning into TF-EBPs were collected in semistructured interviews and analyzed. The semistructured interview guide was developed and refined in consultation with both qualitative methods experts and PTSD treatment experts to ensure that 6 content domains were appropriately queried: PTSD treatment options, cultural sensitivity of treatment, PTSD treatment selection, transition criteria, beliefs about stabilization treatment, and treatment needs/preferences.

Participants were identified using the VA Corporate Data Warehouse and included post-9/11 veterans who had recently initiated CPT or PE for the first time between September 1, 2021, and September 1, 2022. More details of participant selection are available in Holder et al.²¹ From a population of 10,814 patients, stratified random sampling generated a recruitment pool of 200 veterans for further outreach. The strata were defined such that this recruitment pool had similar proportions of demographic characteristics (ie, gender, race, ethnicity) to the population of eligible veterans, equivalent distributions of time to CPT or PE initiation (ie, 33.3% < 1 year, 33.3% 1-3 years, and 33.3% > 3 years), and adequate variability in TF-EBP type (ie, 66.7% CPT, 33.3% PE). A manual chart review in the recruitment pool excluded 12 veterans who did not initiate CPT or PE, 1 veteran with evidence of current active psychosis and/or cognitive impairment that would likely preclude comprehension of study materials, and 1 who was deceased.

Eligible veterans from the recruitment pool were contacted in groups of 25. First, a recruitment letter with study information and instructions to opt-out of further contact was mailed or emailed to veterans. After 2 weeks, veterans who had not responded were contacted by phone up to 3 times. Veterans interested in participating were scheduled for a 1-time visit that included verbal consent and the qualitative interview. Metrics were established a priori to ensure an adequately diverse and inclusive sample. Specifically, a minimum number of racial and/or ethnic minority veterans (33%) and women veterans (20%) were sought. Equal distribution across the 3 categories of time from first mental health visit to CPT/PE initiation also was targeted. Throughout enrollment, recruitment efforts were adapted to meet these metrics in the emerging sample. While the goal was to generate a diverse and inclusive sample using these methods, the sample was not intended to be representative of the population.

Of the 186 eligible participants, 21 declined participation and 26 could not be reached. The targeted sample was reached after exhausting contact for 47 veterans and contacting 80 veterans for a final response rate of 40% among fully contacted veterans and 27% among veterans with any contact. The final sample included 30 veterans who received CPT or PE in VA facilities (Table).

After veterans provided verbal consent for study participation, sociodemographic information was verbally reported, and a 30- to 60-minute semistructured qualitative phone interview was recorded and transcribed. Veterans received $40 for participation. All procedures were approved by the University of California San Francisco Institutional Review Board.

Qualitative Data Analysis

Rapid analysis procedures were used to analyze qualitative data. This approach is suitable for focused, moderately structured qualitative analyses in health services research and facilitates rapid dissemination to stakeholders.²³ The qualitative analysts were 2 clinical psychologists with expertise in PTSD treatment (NH primary and RR secondary). Consistent with rapid analysis procedures, analysts prepared a templated summary (including relevant quotations) of each interview, organized by the prespecified content domains. Interviews were summarized independently, compared to ensure consistency, and discrepancies were resolved through review of the interview source materials. Individual summary templates were combined into a master analytic matrix to facilitate the identification of patterns and delineation of themes. Analysts routinely met to identify, discuss, and refine preliminary themes, revisiting source materials to reach consensus as needed.

Results

Fifteen themes were identified and organized into 2 distinct focus areas: themes directly related to the transition to TF-EBP (8 themes) and themes related to veterans’ experiences with TF-EBP and general mental health care with potential process-improvement implications (7 themes).²¹ Seven themes were identified related to experiences with TF-EBP engagement and VA mental health care. The 7 themes related to TF-EBP engagement and VA mental health care themes are summarized with exemplary quotations.

Veterans want a better understanding of psychotherapy and engaging with VA mental health. Veterans reported that they generally had a poor or “nebulous” understanding about the experience of psychotherapy. For example, veterans exhibited confusion about whether certain experiences were equivalent to participating in psychotherapy. They were sometimes unable to distinguish between interactions such as assessment, disability evaluations, peer support, and psychotherapy. One veteran described a conversation with a TFEBP therapist about prior treatment:

She [asked], have you ever been, or gone through a therapy to begin with? And I, I said, well I just chatted with somebody. And she said that’s not, that’s not therapy. So, I was like, oh, it’s not? That’s not what people do?

Veterans were surprised the VA offered a diverse range of psychotherapy interventions, rather than simply therapy. They did not realize there were different types of psychotherapy. As a result, veterans were not aware that some VA mental practitioners have specialty training and certification to provide treatment matched to specific diagnoses or needs. They thought that all clinicians could provide the same care. One veteran described their understanding:

I just figured all mental health people are mental health people. I didn’t have a better understanding of the system and all the different levels and how it plays out and specialties and things like that. Which, I guess, I should have because you have a primary care doctor, but then you have specialists in all these other different sectors that specialize in one particular area. I guess that should’ve been common sense, but it wasn’t.

Stigma was a barrier to seeking and engaging in mental health care. Veterans discovered they had to overcome stigma associated with seeking and engaging in mental health treatment. Military culture was often discussed as promoting stigma regarding mental health treatment. Specifically, veterans described that seeking treatment meant “either, I’m weak or I’m gonna be seen as weak.” In active-duty settings, the strategy for dealing with mental health symptoms was to “leave those feelings, you push ‘em aside,” an approach highly inconsistent with TF-EBP. In some cases, incorrect information about the VA and PTSD was presented as part of discharge from the military, leading to long-term skepticism of the VA and PTSD treatment. One veteran described his experience as part of a class on the VA compensation and pension assessment process for service-connected disabilities during his military discharge:

[A fellow discharging soldier asked] what about like PTSD, gettin’ rated for PTSD. I hear they take our weapons and stuff like we can’t own firearms and all that stuff. And [the instructor] was like, well, yes that’s a thing. He didn’t explain it like if you get compensated for PTSD you don’t lose your rights to carry a firearm or to have, to be able to go hunting.

Importantly, veterans often described how other identities (eg, race, ethnicity, gender, region of origin) interacted with military culture to enhance stigma. Hearing messaging from multiple sources reinforced beliefs that mental health treatment is inappropriate or is associated with weakness:

As a first-generation Italian, I was always taught keep your feelings to yourself. Never talk outside your family. Never bring up problems to other people and stuff like that. Same with the military. And then the old stigma working in [emergency medical services] and public safety, you’re weak if you get help.

The fundamentals of therapy, including rapport and flexibility, were important. Veterans valued nonspecific therapy factors, genuine empathy, building trust, being honest about treatment, personality, and rapport. These characteristics were almost universally described as particularly important:

I liked the fact that she made it personable and she cared. It wasn’t just like, here, we’re gonna start this. She explained it in the ways I could understand, not in medical terms, so to speak, but that’s what I liked about her. She really cared about what she did and helping me.

Flexibility was viewed as an asset, particularly when clinicians acknowledged veteran autonomy. A consistent example was when veterans were able to titrate trauma disclosure. One veteran described this flexible treatment experience: “She was right there in the room, she said, you know, at any time, you know, we could stop, we could debrief.”

Experiences of clinician flexibility and personalization of therapy were contrasted with experiences of overly rigid therapy. Overemphasis on protocols created barriers, often because treatment did not feel personalized. One veteran described how a clinician’s task-oriented approach interfered with their ability to engage in TF-EBP:

They listened, but it just didn’t seem like they were listening, because they really wanted to stay on task… So, I felt like if the person was more concerned, or more sympathetic to the things that was also going on in my life at that present time, I think I would’ve felt more comfortable talking about what was the PTSD part, too.

Veterans valued shared decision-making prior to TF-EBP initiation. Veterans typically described being involved in a shared decision-making process prior to initiating TF-EBP. During these sessions, clinicians discussed treatment options and provided veterans with a variety of materials describing treatments (eg, pamphlets, websites, videos, statistics). Most veterans appreciated being able to reflect on and discuss treatment options with their clinicians. Being given time in and out of session to review was viewed as valuable and increased confidence in treatment choice. One veteran described their experience:

I was given the information, you know, they gave me handouts, PDFs, whatever was available, and let me read over it. I didn’t have to choose anything right then and there, you know, they let me sleep on it. And I got back to them after some thought.

However, some veterans felt overwhelmed by being presented with too much information and did not believe they knew enough to make a final treatment decision. One veteran described being asked to contribute to the treatment decision:

I definitely asked [the clinician] to weigh in on maybe what he thought was best, because—I mean, I don’t know… I’m not necessarily sure I know what I think is best. I think we’re just lucky I’m here, so if you can give me a solid and help me out here by telling me just based on what I’ve said to you and the things that I’ve gone through, what do you think?

Veterans who perceived that their treatment preferences were respected had a positive outlook on TF-EBP. As part of the shared-decision making process, veterans typically described being given choices among PTSD treatments. One way that preferences were respected was through clinicians tailoring treatment descriptions to a veteran’s unique symptoms, experiences, and values. In these cases, clinicians observed specific concerns and clearly linked treatment principles to those concerns. For example, one veteran described their clinician’s recommendation for PE: “The hardest thing for me is to do the normal things like grocery store or getting on a train or anything like that. And so, he suggested that [PE] would be a good idea.”

In other cases, veterans wanted the highest quality of treatment rather than a match between treatment principles and the veteran’s presentation, goals, or strengths. These veterans wanted the best treatment available for PTSD and valued research support, recommendations from clinical practice guidelines, or clinician confidence in the effectiveness of the treatment. One veteran described this perspective:

I just wanted to be able to really tackle it in the best way possible and in the most like aggressive way possible. And it seemed like PE really was going to, they said that it’s a difficult type of therapy, but I really just wanted to kind of do the best that I could to eradicate some of the issues that I was having.

When veterans perceived a lack of respect for their preferences, they were hesitant about TF-EBP. For some veterans, a generic pitch for a TF-EBP was detrimental in the absence of the personal connection between the treatment and their own symptoms, goals, or strengths. These veterans did not question whether the treatment was effective in general but did question whether the treatment was best for them. One veteran described the contrast between their clinician’s perspective and their own.

I felt like they felt very comfortable, very confident in [CPT] being the program, because it was comfortable for them. Because they did it several times. And maybe they had a lot of success with other individuals... but they were very comfortable with that one, as a provider, more than: Is this the best fit for [me]?

Some veterans perceived little concern for their preferences and a lack of choice in available treatments, which tended to perpetuate negative perceptions of TFEBP. These veterans described their lack of choices with frustration. Alternatives to TFEBP were described by these veterans as so undesirable that they did not believe they had a real choice:

[CPT] was the only decision they had. There was nothing else for PTSD. They didn’t offer anything else. So, I mean it wasn’t a decision. It was either … take treatment or don’t take treatment at all… Actually, I need to correct myself. So, there were 2 options, group therapy or CPT. I forgot about that. I’m not a big group guy so I chose the CPT.

Another veteran was offered a choice between therapeutic approaches, but all were delivered via telehealth (consistent with the transition to virtual services during the COVID-19 pandemic). For this veteran, not only was the distinction between approaches unclear, but the choice between approaches was unimportant compared to the mode of delivery.

This happened during COVID-19 and VA stopped seeing anybody physically, face-to-face. So my only option for therapy was [telehealth]… There was like 3 of them, and I tried to figure out, you know, from the layperson’s perspective, like: I don’t know which one to go with.

Veterans wanted to be asked about their cultural identity. Veterans valued when clinicians asked questions about cultural identity as part of their mental health treatment and listened to their cultural context. Cultural identity factors extended beyond factors such as race, ethnicity, gender, and sexual orientation to religion, military culture, and regionality. Veterans often described situations where they wished clinicians would ask the question or initiate conversations about culture. A veteran highlighted the importance of their faith but noted that it was a taboo topic. Their clinician did not say “we don’t go there,” but they “never dove into it either.” Another veteran expressed a desire for their clinician to ask questions about experiences in the National Guard and as an African American veteran:

If a provider was to say like: Oh, you know, it’s a stressful situation being a part of the military, being in the National Guard. You know, just asking questions about that. I think that would really go a long way… Being African American was difficult as well. And more so because of my region, I think… I felt like it would probably be an uncomfortable subject to speak on… I mean, it wasn’t anything that my providers necessarily did, it was more so just because it wasn’t brought up.

One common area of concern for veterans was a match between veteran and therapist demographics. When asked about how their cultural identity influenced treatment, several veterans described the relevance of therapist match. Much like questions about their own cultural identity, veterans valued being asked about identity preferences in clinicians (eg, gender or race matching), rather than having to bring up the preference themselves. One veteran described relief at this question being asked directly: “I was relieved when she had asked [whether I wanted a male or female clinician] primarily because I was going to ask that or bring that up somehow. But her asking that before me was a weight off my shoulders.”

Discussing cultural identity through treatment strengthened veterans’ engagement in therapy. Many veterans appreciated when analogies used in therapy were relevant to their cultural experiences and when clinicians understood their culture (eg, military culture, race, ethnicity, religious beliefs, sexual orientation). One veteran described how their clinician understood military culture and made connections between military culture and the rationale for TF-EBP, which strengthened the veteran’s buy-in for the treatment and alliance with the clinician:

At the beginning when she was explaining PTSD, and I remember she said that your brain needed to think this way when you were in the military because it was a way of protecting and surviving, so your brain was doing that in order for you to survive in whatever areas you were because there was danger. So, your brain had you thinking that way. But now, you’re not in those situations anymore. You’re not in danger. You’re not in the military, but your brain is still thinking you are, and that’s what PTSD generally does to you.

Specific elements of TF-EBP also provided opportunities to discuss and integrate important aspects of identity. This is accomplished in PE by assigning relevant in vivo exercises. In CPT, “connecting the dots” on how prior experiences influenced trauma-related stuck points achieved this element. One veteran described their experience with a clinician who was comfortable discussing the veteran’s sexual orientation and recognized the impacts of prior trauma on intimacy:

They’re very different, and there’s a lot of things that can be accepted in gay relationships that are not in straight ones. With all that said, I think [the PE therapist] did a fantastic job being not—like never once did she laugh or make an uncomfortable comment or say she didn’t wanna talk about something when like part of the reason I wanted to get into therapy is that my partner and I weren’t having sex unless I used alcohol.

Discussion

As part of a larger national qualitative investigation of the experiences of veterans who recently initiated TF-EBP, veterans discussed their experiences with therapy and mental health care that have important implications for continued process improvement.²¹ Three key areas for continued process improvement were identified: (1) providing information about the diverse range of mental health care services at the VA and the implications of this continuum of care; (2) consideration of veteran preferences in treatment decision-making, including the importance of perceived choice; and (3) incorporating cultural assessment and cultural responsiveness into case conceptualization and treatment.

One area of process improvement identified was increasing knowledge about different types of psychotherapy and the continuum of care available at the VA. Veterans in this study confused or conflated participating in psychotherapy with talking about mental health symptoms with a clinician (eg, assessment, disability evaluation). They were sometimes surprised that psychotherapy is an umbrella term referring to a variety of different modalities. The downstream impact of these misunderstandings was a perception of VA mental health care as nebulous. Veterans were surprised that all mental health practitioners were unable to provide the same care. Confusion may have been compounded by highly variable referral processes across VA.²⁴ To address this, clinicians have developed local educational resources and handouts for both veterans and referring clinicians from nonmental health and general mental health specialties.²⁵ Given the variability in referral processes both between and within VA medical centers, national dissemination of these educational materials may be more difficult compared to materials for TF-EBPs.²⁴ The VA started to use behavioral health interdisciplinary program (BHIP) teams, which are designed to be clinical homes for veterans connected with a central clinician who can explain and coordinate their mental health care as well as bring more consistency to the referral process.²⁶ The ongoing transition toward the BHIP model of mental health care at VA may provide the opportunity to consolidate and integrate knowledge about the VA approach to mental health care, potentially filling knowledge gaps.

A second area of process improvement focused on the shared decision-making process. Consistent with mental health initiatives, veterans generally believed they had received sufficient information about TF-EBP and engaged in shared decision-making with clinicians.^20,27 Veterans were given educational materials to review and had the opportunity to discuss these materials with clinicians. However, veterans described variability in the success of shared decision-making. Although veterans valued receiving accurate, comprehensible information to support treatment decisions, some preferred to defer to clinicians’ expertise regarding which treatment to pursue. While these veterans valued information, they also valued the expertise of clinicians in explaining why specific treatments would be beneficial. A key contributor to veterans satisfaction was assessing how veterans wanted to engage in the decision-making process and respecting those preferences.²⁸ Veterans approached shared decision-making differently, from making decisions independently after receiving information to relying solely on clinician recommendation. The process was most successful when clinicians articulated how their recommended treatment aligned with a veteran’s preferences, including recommendations based on specific values (eg, personalized match vs being the best). Another important consideration is ensuring veterans know they can receive a variety of different types of mental health services available in different modalities (eg, virtual vs in-person; group vs individual). When veterans did not perceive choice in treatment aspects important to them (typically despite having choices), they were less satisfied with their TF-EBP experience.

A final area of process improvement identified involves how therapists address important aspects of culture. Veterans often described mental health stigma coming from intersecting cultural identities and expressed appreciation when therapists helped them recognize the impact of these beliefs on treatment. Some veterans did not discuss important aspects of their identity with clinicians, including race/ethnicity, religion, and military culture. Veterans did not report negative interactions with clinicians or experiences suggesting it was inappropriate to discuss identity; however, they were reluctant to independently raise these identity factors. Strategies such as the ADDRESSING framework, a mnemonic acronym that describes a series of potentially relevant characteristics, can help clinicians comprehensively consider different aspects that may be relevant to veterans, modeling that discussion of relevant these characteristics is welcome in TF-EBP.²⁹ Veterans reported that making culturally relevant connections enhanced the TF-EBP experience, most commonly with military culture. These data support that TF-EBP delivery with attention to culture should be an integrated part of treatment, supporting engagement and therapeutic alliance.³⁰ The VA National Center for PTSD consultation program is a resource to support clinicians in assessing and incorporating relevant aspects of cultural identity.³¹ For example, the National Center for PTSD provides a guide for using case conceptualization to address patient reactions to race-based violence during PTSD treatment.³² Both manualized design and therapist certification training can reinforce that assessing and attending to case conceptualization (including identity factors) is an integral component of TF-EBP.^33,34

Limitations

While the current study has numerous strengths (eg, national veteran sampling, robust qualitative methods), results should be considered within the context of study limitations. First, veteran participants all received TF-EBP, and the perspectives of veterans who never initiate TF-EBP may differ. Despite the strong sampling approach, the study design is not intended to be generalizable to all veterans receiving TF-EBP for PTSD. Qualitative analysis yielded 15 themes, described in this study and prior research, consistent with recommendations.^21,22 This approach allows rich description of distinct focus areas that would not be possible in a single manuscript. Nonetheless, all veterans interviewed described their experiences in TF-EBP and general mental health care, the focus of the semistructured interview guide was on the experience of transitioning from other treatment to TF-EBP.

Conclusion

This study describes themes related to general mental health and TF-EBP process improvement as part of a larger study on transitions in PTSD care.^21,22 Veterans valued the fundamentals of therapy, including rapport and flexibility. Treatment-specific rapport (eg, pointing out treatment progress and effort in completing treatment components) and flexibility within the context of fidelity (ie, personalizing treatment while maintaining core treatment elements) may be most effective at engaging veterans in recommended PTSD treatments.^18,34 In addition to successes, themes suggest multiple opportunities for process improvement. Ongoing VA initiatives and priorities (ie, BHIP, shared decision-making, consultation services) aim to improve processes consistent with veteran recommendations. Future research is needed to evaluate the success of these and other programs to optimize access to and engagement in recommended PTSD treatments.

Deconditioning among hospitalized older adults contributes to significant decline in posthospitalization functional ability, physical performance, and physical activity.^1-10 Previous hospital-to-home interventions have targeted improving function and physical activity, including recent programs leveraging home telehealth as a feasible and potentially effective mode of delivering in-home exercise and rehabilitation.^11-14 However, pilot interventions have shown mixed effectiveness.^11,12,14 This study expands on a previously published intervention describing a pilot home telehealth program for veterans posthospital discharge that demonstrated significant 6-month improvement in physical activity as well as trends in physical function improvement, including among those with cognitive impairment.¹⁵ Factors that contribute to improved outcomes are the focus of the present study.

Key factors underlying the complexity of hospital-to-home transitions include hospitalization elements (ie, reason for admission and length of stay), associated posthospital syndromes (ie, postdischarge falls, medication changes, cognitive impairment, and pain), and postdischarge health care application (ie, physical therapy and hospital readmission).^16-18 These factors may be associated with postdischarge functional ability, physical performance, and physical activity, but their direct influence on intervention outcomes is unclear (Figure 1).^5,7,9,16-20 The objective of this study was to examine the influence of hospitalization, posthospital syndrome, and postdischarge health care application factors on outcomes of a US Department of Veterans Affairs (VA) Video Connect (VVC) intervention to enhance function and physical activity in older adults posthospital discharge.

Methods

The previous analysis reported on patient characteristics, program feasibility, and preliminary outcomes.^13,15 The current study reports on relationships between hospitalization, posthospital syndrome, and postdischarge health care application factors and change in key outcomes, namely postdischarge self-reported functional ability, physical performance, and physical activity from baseline to endpoint.

Participants provided written informed consent. The protocol and consent forms were approved by the VA Ann Arbor Healthcare System (VAAAHS) Research and Development Committee, and the project was registered on clinicaltrials.gov (NCT04045054).

Intervention

The pilot program targeted older adults following recent hospital discharge from VAAAHS. Participants were eligible if they were aged ≥ 50 years, had been discharged following an inpatient stay in the past 1 to 2 weeks, evaluated by physical therapy during hospitalization with stated rehabilitation goals on discharge, and followed by a VAAAHS primary care physician. Participants were either recruited during hospital admission or shortly after discharge.¹³

An experienced physical activity trainer (PAT) supported the progression of participants’ rehabilitation goals via a home exercise program and coached the patient and caregiver to optimize functional ability, physical performance, and physical activity. The PAT was a nonlicensed research assistant with extensive experience in applying standard physical activity enhancement protocols (eg, increased walking) to older adults with comorbidities. Participation in the program lasted about 6 months. Initiation of the PAT program was delayed if the patient was already receiving postdischarge home-based or outpatient physical therapy. The PAT contacted the patient weekly via VVC for the first 6 weeks, then monthly for a total of 6 months. Each contact included information on optimal walking form, injury prevention, program progression, and ways to incorporate sit-to-stand transitions, nonsitting behavior, and walking into daily routines. The initial VVC contact lasted about 60 minutes and subsequent sessions lasted about 30 minutes.¹³

Demographic characteristics were self-reported by participants and included age, sex, race, years of education, and marital status. Clinical characteristics were obtained from each participant’s electronic health record (EHR), including copay status, index hospitalization length of stay, admission diagnosis, and postsurgery status (postsurgery vs nonpostsurgery). Intervention adherence was tracked as the number of PAT sessions attended.

Posthospital Syndrome Factors

Participant falls (categorized as those who reported a fall vs those who did not) and medication changes (number of changes reported, including new medication, discontinued medication, dose changes, medication changes, or changes in medication schedule) were reported by participants or caregivers during each VVC contact. Participants completed the Montreal Cognitive Assessment (MoCA) at baseline, and were dichotomized into 2 groups: no cognitive impairment (MoCA score ≥ 26) and mild to moderate cognitive impairment (MoCA score 10-25).^13,21

Participants rated how much pain interfered with their normal daily activities since the previous VVC session on a 5-point Likert scale (1, not at all; to 5, extremely).²² Similar to prior research, participants were placed into 2 groups based on their mean pain interference score (individuals with scores from 1.0 to 2.0 in 1 group, and individuals with > 2.0 in another).^23-25 Participants were separated into a no or mild pain interference group and a moderate to severe pain interference group. Hospital readmissions (VA and non-VA) and postdischarge physical therapy outcomes were obtained from the participant’s EHR, including primary care visits.

Outcomes

Outcomes were collected at baseline (posthospital discharge) and 6 months postenrollment.

Self-Reported Functional Ability. This measure is provided by participants or caregivers and measured by the Katz Index of Independence in Activities of Daily Living (ADL), Lawton and Brody Instrumental ADL Scale (IADL), Nagi Disability Model, and Rosow-Breslau Scale. The Katz ADL assesses the ability to complete 6 self-care activities and awards 1 point for independence and 0 if the individual is dependent (total score range, 0-6).²⁶ The Lawton and Brody IADL measures an individual’s independence in 8 instrumental ADLs; it awards 1 point for independence and 0 if the individual is dependent (total score range, 0-8).²⁷ The Nagi Disability Model evaluates an individual’s difficulty performing 5 tasks (total score range, 0-5) and tallies the number of items with a response other than “no difficulty at all” (higher total score indicates greater difficulty). ²⁸ The Rosow-Breslau Scale is a 3-item measure of mobility disability; individual responses are 0 (no help) and 1 (requires help or unable); higher total score (range, 0-3) indicates greater disability.²⁹

Physical Performance. Measured using the Short Physical Performance Battery (SPPB), which evaluates standing balance, sit to stand, and walking performance. Scores range from 0 to 4 on the balance, gait speed, and chair stand tests, for a total composite score between 0 and 12 (higher score indicates better performance).³⁰

Physical Activity. Measured using actigraphy, namely a physical activity monitor adherent to the thigh (activ-PAL3TM, PAL Technologies Ltd., Glasgow, UK).³¹ Participants were instructed to wear the activPal for ≥ 1 week. Participants with a minimum of 5 days of wear were included in this analysis.

Data Analyses

Analyses were performed using SPSS software version 29.0.³² Continuous variables were summarized using mean (SD) or median and IQR using the weighted average method; categorical variables were summarized using frequencies and percentages. Baseline scores on outcome variables were compared by categorical hospitalization, posthospital syndrome, and postdischarge health care application factor variables using Mann-Whitney U tests. The differences between outcome variables from baseline to endpoint were then calculated to produce change scores. Relationships between the number of PAT sessions attended and baseline outcomes and outcome change scores were estimated using Spearman correlations. Relationships between categorical factors (hospitalization, posthospital syndrome, and postdischarge health care application) and outcome variable change scores (which were normally distributed) were examined using Mann-Whitney U tests. Relationships with continuous hospitalization (length of stay) and posthospital syndrome factors (medication changes) were estimated using Spearman correlations. Effect sizes (ES) were estimated with Cohen d; small (d = 0.2), medium (d = 0.5), or large (d ≥ 0.8). Missing data were handled using pairwise deletion.³³ Therefore, sample sizes were reported for each analysis. For all statistical tests, P < .05 was considered significant.

Results

Twenty-four individuals completed the pilot intervention.¹⁵ Mean (SD) age was 73.6 (8.1) years (range, 64-93 years) and participants were predominantly White males (Table 1). Eight participants had a high school education only and 13 had more than a high school education. Diagnoses at admission included 9 patients with orthopedic/musculoskeletal conditions (6 were for joint replacement), 6 patients with vascular/pulmonary conditions, and 4 with gastrointestinal/renal/urological conditions. Of the 11 postsurgery participants, 7 were orthopedic, 4 were gastrointestinal, and 1 was peripheral vascular.

Baseline outcome scores did not differ significantly between groups, except individuals with moderate to severe pain interference reported a significantly lower IADL score (median [IQR] 4 [2-7]) than individuals with mild or moderate pain interference (median [IQR] 8 [7-8]; P = .02) (Table 2). The mean (SD) number of PAT sessions attended was 9.3 (3.7) (range, 3-19). There were no significant relationships between number of sessions attended and any baseline outcome variables or outcome change scores.

Hospitalization Factors

Participants who were postsurgery tended to have greater improvement than individuals who were nonpostsurgery in ADLs (median [IQR] 0 [0-1.5]; ES, 0.6; P = .10) and SPPB (median [IQR] 2 [1.5-9]; ES, 0.9; P = .07), but the improvements were not statistically significant (Table 3). Mean (SD) length of stay of the index hospitalization was 6.7 (6.1) days. Longer length of stay was significantly correlated with an increase in Nagi score (ρ, 0.45; 95% CI, 0.01-0.75). There were no other significant or trending relationships between length of stay and outcome variables.

Posthospital Syndrome Factors

The 16 participants with mild to moderate cognitive impairment had less improvement in ADLs (median [IQR] 0 [0-1]) than the 8 participants with no impairment (median [IQR] 0 [-0.75 to 0]; ES, -1.1; P = .04). Change in outcome variables from baseline to endpoint did not significantly differ between the 8 patients who reported a fall compared with the 13 who did not, nor were any trends observed. Change in outcome variables from baseline to endpoint also did not significantly differ between the 8 participants who reported no or mild pain interference compared with the 10 patients with moderate to severe pain interference, nor were any trends observed. Mean (SD) number of medication changes was 2.5 (1.6). Higher number of medication changes was significantly correlated with a decrease in Rosow-Breslau score (ρ, -0.47; 95% CI, -0.76 to -0.02). There were no other significant or trending relationships between number of medication changes and outcome variables.

Postdischarge Health Care Application Factors

The 16 participants who attended posthospital physical therapy trended towards less improvement in IADLs (median [IQR] 0 [-0.5 to 1.5]; ES, -0.7; P = .11) and SPPB (median [IQR] 2 [-3.0 to 4.5]; ES, -0.5; P = .15) than the 8 patients with no postdischarge physical therapy. Eleven participants were readmitted, while 13 had no readmissions in their medical records between baseline and endpoint. Participants with ≥ 1 readmission experienced a greater increase in Rosow-Breslau score (median [IQR] 0 [-0.5 to 1.0]) than those not readmitted (median [IQR] 0 [-1.25 to 0.25]; ES, 1.0; P = .03). Borderline greater improvement in number of steps was found in those not readmitted (median [IQR] 3365.6 [274.4-7710.9]) compared with those readmitted (median [IQR] 319.9 [-136.1 to 774.5]; ES, -1.3; P = .05). Patients who were readmitted also tended to have lower and not statistically significant improvements in SPPB (median [IQR] 1 [-4.0 to 5.3]) compared with those not readmitted (median [IQR] 2 [0.3-3.8]; ES, -0.5; P = .17) (Table 3).

Discussion

This study examined the association between hospitalization, posthospital syndrome, and postdischarge health care use in patients undergoing a VVC-based intervention following hospital discharge. Participants who had no or mild cognitive impairment, no readmissions, higher medication changes, and a shorter hospital length of stay tended to experience lower disability, including in mobility and ADLs. This suggests individuals who are less clinically complex may be more likely to benefit from this type of virtual rehabilitation program. These findings are consistent with clinical experiences; home-based programs to improve physical activity posthospital discharge can be challenging for those who were medically ill (and did not undergo a specific surgical procedure), cognitively impaired, and become acutely ill and trigger hospital readmission. ¹⁵ For example, the sample in this study had higher rates of falls, pain, and readmissions compared to previous research.^2,3,34-39

The importance of posthospital syndrome in the context of recovery of function and health at home following hospitalization is well documented.^16-18 The potential impact of posthospital syndrome on physical activity-focused interventions is less understood. In our analysis, participants with mild or moderate cognitive impairment tended to become more dependent in their ADLs, while those with no cognitive impairment tended to become more independent in their ADLs. This functional decline over time is perhaps expected in persons with cognitive impairment, but the significant difference with a large ES warrants further consideration on how to tailor interventions to better promote functional recovery in these individuals.^40,41 While some cognitive decline may not be preventable, this finding supports the need to promote healthy cognitive aging, identify declines in cognition, and work to mitigate additional decline. Programs specifically designed to promote function and physical activity in older adults with cognitive impairment are needed, especially during care transitions.^41-43

While participants reported that falls and pain interference did not have a significant impact on change in outcomes between baseline and endpoint, these areas need further investigation. Falls and pain have been associated with function and physical activity in older adults.^42-46 Pain is common, yet underappreciated during older adult hospital-to-home transitions.^11,12,45,46 There is a need for more comprehensive assessment of pain (including pain intensity) and qualitative research.

Hospitalization and postdischarge health care application factors may have a significant impact on home-telehealth physical activity intervention success. Individuals who were postsurgery tended to have greater improvements in ADLs and physical performance. Most postsurgery participants had joint replacement surgery. Postsurgery status may not be modifiable, but it is important to note expected differences in recovery between medical and surgical admissions and the need to tailor care based on admission diagnosis. Those with a longer length of hospital stay may be considered at higher risk of suboptimal outcomes postdischarge, which indicates an opportunity for targeting resources and support, in addition to efforts of reducing length of stay where possible.⁴⁷

Readmissions were significantly related to a change in Rosow-Breslau mobility disability score. This may indicate the detrimental impact a readmission can have on increasing mobility and physical activity postdischarge, or the potential of this pilot program to impact readmissions by increasing mobility and physical activity, contrary to prior physical exercise interventions.^5,7,9,48 With 5% to 79% of readmissions considered preventable, continued efforts and program dissemination and implementation to address preventable readmissions are warranted.⁴⁹ Individuals with postdischarge physical therapy (prior to beginning the pilot program) tended to demonstrate less improvement in disability and physical performance. This relationship needs further investigation; the 2 groups did not appear to have significant differences at baseline, albeit with a small sample size. It is possible they experienced initial improvements with postdischarge physical therapy and plateaued or had little further reserve to improve upon entering the VVC program.

Strengths and Limitations

This pilot program provided evaluative data on the use of VVC to enhance function and physical activity in older adults posthospital discharge. It included individual (eg, fall, pain, cognitive impairment) and health service (eg, readmission, physical therapy) level factors as predictors of function and physical activity posthospitalization.^5,7,9,15-19

The results of this pilot project stem from a small sample lacking diversity in terms of race, ethnicity, and sex. There was some variation in baseline and endpoints between participants, and when hospitalization, posthospital syndrome, and postdischarge health care application factors were collected. The majority of participants were recruited within a month postdischarge, and the program lasted about 6 months. Data collection was attempted at regular PAT contacts, but there was some variation in when visits occurred based on participant availability and preference. Some participants had missing data, which was handled using pairwise deletion.³³ Larger studies are needed to confirm the findings of this study, particularly the trends that did not reach statistical significance. Home health services other than physical therapy (eg, nursing, occupational therapy) were not fully accounted for and should be considered in future research.

Conclusions

In patients undergoing a 6-month pilot VVC-based physical activity intervention posthospital discharge, improvements in mobility and disability were most likely in those who had no cognitive impairment and were not readmitted. Larger sample and qualitative investigations are necessary to optimize outcomes for patients who meet these clinical profiles.

Deconditioning among hospitalized older adults contributes to significant decline in posthospitalization functional ability, physical performance, and physical activity.^1-10 Previous hospital-to-home interventions have targeted improving function and physical activity, including recent programs leveraging home telehealth as a feasible and potentially effective mode of delivering in-home exercise and rehabilitation.^11-14 However, pilot interventions have shown mixed effectiveness.^11,12,14 This study expands on a previously published intervention describing a pilot home telehealth program for veterans posthospital discharge that demonstrated significant 6-month improvement in physical activity as well as trends in physical function improvement, including among those with cognitive impairment.¹⁵ Factors that contribute to improved outcomes are the focus of the present study.

Key factors underlying the complexity of hospital-to-home transitions include hospitalization elements (ie, reason for admission and length of stay), associated posthospital syndromes (ie, postdischarge falls, medication changes, cognitive impairment, and pain), and postdischarge health care application (ie, physical therapy and hospital readmission).^16-18 These factors may be associated with postdischarge functional ability, physical performance, and physical activity, but their direct influence on intervention outcomes is unclear (Figure 1).^5,7,9,16-20 The objective of this study was to examine the influence of hospitalization, posthospital syndrome, and postdischarge health care application factors on outcomes of a US Department of Veterans Affairs (VA) Video Connect (VVC) intervention to enhance function and physical activity in older adults posthospital discharge.

Methods

The previous analysis reported on patient characteristics, program feasibility, and preliminary outcomes.^13,15 The current study reports on relationships between hospitalization, posthospital syndrome, and postdischarge health care application factors and change in key outcomes, namely postdischarge self-reported functional ability, physical performance, and physical activity from baseline to endpoint.

Participants provided written informed consent. The protocol and consent forms were approved by the VA Ann Arbor Healthcare System (VAAAHS) Research and Development Committee, and the project was registered on clinicaltrials.gov (NCT04045054).

Intervention

The pilot program targeted older adults following recent hospital discharge from VAAAHS. Participants were eligible if they were aged ≥ 50 years, had been discharged following an inpatient stay in the past 1 to 2 weeks, evaluated by physical therapy during hospitalization with stated rehabilitation goals on discharge, and followed by a VAAAHS primary care physician. Participants were either recruited during hospital admission or shortly after discharge.¹³

An experienced physical activity trainer (PAT) supported the progression of participants’ rehabilitation goals via a home exercise program and coached the patient and caregiver to optimize functional ability, physical performance, and physical activity. The PAT was a nonlicensed research assistant with extensive experience in applying standard physical activity enhancement protocols (eg, increased walking) to older adults with comorbidities. Participation in the program lasted about 6 months. Initiation of the PAT program was delayed if the patient was already receiving postdischarge home-based or outpatient physical therapy. The PAT contacted the patient weekly via VVC for the first 6 weeks, then monthly for a total of 6 months. Each contact included information on optimal walking form, injury prevention, program progression, and ways to incorporate sit-to-stand transitions, nonsitting behavior, and walking into daily routines. The initial VVC contact lasted about 60 minutes and subsequent sessions lasted about 30 minutes.¹³

Demographic characteristics were self-reported by participants and included age, sex, race, years of education, and marital status. Clinical characteristics were obtained from each participant’s electronic health record (EHR), including copay status, index hospitalization length of stay, admission diagnosis, and postsurgery status (postsurgery vs nonpostsurgery). Intervention adherence was tracked as the number of PAT sessions attended.

Posthospital Syndrome Factors

Participant falls (categorized as those who reported a fall vs those who did not) and medication changes (number of changes reported, including new medication, discontinued medication, dose changes, medication changes, or changes in medication schedule) were reported by participants or caregivers during each VVC contact. Participants completed the Montreal Cognitive Assessment (MoCA) at baseline, and were dichotomized into 2 groups: no cognitive impairment (MoCA score ≥ 26) and mild to moderate cognitive impairment (MoCA score 10-25).^13,21

Participants rated how much pain interfered with their normal daily activities since the previous VVC session on a 5-point Likert scale (1, not at all; to 5, extremely).²² Similar to prior research, participants were placed into 2 groups based on their mean pain interference score (individuals with scores from 1.0 to 2.0 in 1 group, and individuals with > 2.0 in another).^23-25 Participants were separated into a no or mild pain interference group and a moderate to severe pain interference group. Hospital readmissions (VA and non-VA) and postdischarge physical therapy outcomes were obtained from the participant’s EHR, including primary care visits.

Outcomes

Outcomes were collected at baseline (posthospital discharge) and 6 months postenrollment.

Self-Reported Functional Ability. This measure is provided by participants or caregivers and measured by the Katz Index of Independence in Activities of Daily Living (ADL), Lawton and Brody Instrumental ADL Scale (IADL), Nagi Disability Model, and Rosow-Breslau Scale. The Katz ADL assesses the ability to complete 6 self-care activities and awards 1 point for independence and 0 if the individual is dependent (total score range, 0-6).²⁶ The Lawton and Brody IADL measures an individual’s independence in 8 instrumental ADLs; it awards 1 point for independence and 0 if the individual is dependent (total score range, 0-8).²⁷ The Nagi Disability Model evaluates an individual’s difficulty performing 5 tasks (total score range, 0-5) and tallies the number of items with a response other than “no difficulty at all” (higher total score indicates greater difficulty). ²⁸ The Rosow-Breslau Scale is a 3-item measure of mobility disability; individual responses are 0 (no help) and 1 (requires help or unable); higher total score (range, 0-3) indicates greater disability.²⁹

Physical Performance. Measured using the Short Physical Performance Battery (SPPB), which evaluates standing balance, sit to stand, and walking performance. Scores range from 0 to 4 on the balance, gait speed, and chair stand tests, for a total composite score between 0 and 12 (higher score indicates better performance).³⁰

Physical Activity. Measured using actigraphy, namely a physical activity monitor adherent to the thigh (activ-PAL3TM, PAL Technologies Ltd., Glasgow, UK).³¹ Participants were instructed to wear the activPal for ≥ 1 week. Participants with a minimum of 5 days of wear were included in this analysis.

Data Analyses

Analyses were performed using SPSS software version 29.0.³² Continuous variables were summarized using mean (SD) or median and IQR using the weighted average method; categorical variables were summarized using frequencies and percentages. Baseline scores on outcome variables were compared by categorical hospitalization, posthospital syndrome, and postdischarge health care application factor variables using Mann-Whitney U tests. The differences between outcome variables from baseline to endpoint were then calculated to produce change scores. Relationships between the number of PAT sessions attended and baseline outcomes and outcome change scores were estimated using Spearman correlations. Relationships between categorical factors (hospitalization, posthospital syndrome, and postdischarge health care application) and outcome variable change scores (which were normally distributed) were examined using Mann-Whitney U tests. Relationships with continuous hospitalization (length of stay) and posthospital syndrome factors (medication changes) were estimated using Spearman correlations. Effect sizes (ES) were estimated with Cohen d; small (d = 0.2), medium (d = 0.5), or large (d ≥ 0.8). Missing data were handled using pairwise deletion.³³ Therefore, sample sizes were reported for each analysis. For all statistical tests, P < .05 was considered significant.

Results

Twenty-four individuals completed the pilot intervention.¹⁵ Mean (SD) age was 73.6 (8.1) years (range, 64-93 years) and participants were predominantly White males (Table 1). Eight participants had a high school education only and 13 had more than a high school education. Diagnoses at admission included 9 patients with orthopedic/musculoskeletal conditions (6 were for joint replacement), 6 patients with vascular/pulmonary conditions, and 4 with gastrointestinal/renal/urological conditions. Of the 11 postsurgery participants, 7 were orthopedic, 4 were gastrointestinal, and 1 was peripheral vascular.

Baseline outcome scores did not differ significantly between groups, except individuals with moderate to severe pain interference reported a significantly lower IADL score (median [IQR] 4 [2-7]) than individuals with mild or moderate pain interference (median [IQR] 8 [7-8]; P = .02) (Table 2). The mean (SD) number of PAT sessions attended was 9.3 (3.7) (range, 3-19). There were no significant relationships between number of sessions attended and any baseline outcome variables or outcome change scores.

Hospitalization Factors

Participants who were postsurgery tended to have greater improvement than individuals who were nonpostsurgery in ADLs (median [IQR] 0 [0-1.5]; ES, 0.6; P = .10) and SPPB (median [IQR] 2 [1.5-9]; ES, 0.9; P = .07), but the improvements were not statistically significant (Table 3). Mean (SD) length of stay of the index hospitalization was 6.7 (6.1) days. Longer length of stay was significantly correlated with an increase in Nagi score (ρ, 0.45; 95% CI, 0.01-0.75). There were no other significant or trending relationships between length of stay and outcome variables.

Posthospital Syndrome Factors

The 16 participants with mild to moderate cognitive impairment had less improvement in ADLs (median [IQR] 0 [0-1]) than the 8 participants with no impairment (median [IQR] 0 [-0.75 to 0]; ES, -1.1; P = .04). Change in outcome variables from baseline to endpoint did not significantly differ between the 8 patients who reported a fall compared with the 13 who did not, nor were any trends observed. Change in outcome variables from baseline to endpoint also did not significantly differ between the 8 participants who reported no or mild pain interference compared with the 10 patients with moderate to severe pain interference, nor were any trends observed. Mean (SD) number of medication changes was 2.5 (1.6). Higher number of medication changes was significantly correlated with a decrease in Rosow-Breslau score (ρ, -0.47; 95% CI, -0.76 to -0.02). There were no other significant or trending relationships between number of medication changes and outcome variables.

Postdischarge Health Care Application Factors

The 16 participants who attended posthospital physical therapy trended towards less improvement in IADLs (median [IQR] 0 [-0.5 to 1.5]; ES, -0.7; P = .11) and SPPB (median [IQR] 2 [-3.0 to 4.5]; ES, -0.5; P = .15) than the 8 patients with no postdischarge physical therapy. Eleven participants were readmitted, while 13 had no readmissions in their medical records between baseline and endpoint. Participants with ≥ 1 readmission experienced a greater increase in Rosow-Breslau score (median [IQR] 0 [-0.5 to 1.0]) than those not readmitted (median [IQR] 0 [-1.25 to 0.25]; ES, 1.0; P = .03). Borderline greater improvement in number of steps was found in those not readmitted (median [IQR] 3365.6 [274.4-7710.9]) compared with those readmitted (median [IQR] 319.9 [-136.1 to 774.5]; ES, -1.3; P = .05). Patients who were readmitted also tended to have lower and not statistically significant improvements in SPPB (median [IQR] 1 [-4.0 to 5.3]) compared with those not readmitted (median [IQR] 2 [0.3-3.8]; ES, -0.5; P = .17) (Table 3).

Discussion

This study examined the association between hospitalization, posthospital syndrome, and postdischarge health care use in patients undergoing a VVC-based intervention following hospital discharge. Participants who had no or mild cognitive impairment, no readmissions, higher medication changes, and a shorter hospital length of stay tended to experience lower disability, including in mobility and ADLs. This suggests individuals who are less clinically complex may be more likely to benefit from this type of virtual rehabilitation program. These findings are consistent with clinical experiences; home-based programs to improve physical activity posthospital discharge can be challenging for those who were medically ill (and did not undergo a specific surgical procedure), cognitively impaired, and become acutely ill and trigger hospital readmission. ¹⁵ For example, the sample in this study had higher rates of falls, pain, and readmissions compared to previous research.^2,3,34-39

The importance of posthospital syndrome in the context of recovery of function and health at home following hospitalization is well documented.^16-18 The potential impact of posthospital syndrome on physical activity-focused interventions is less understood. In our analysis, participants with mild or moderate cognitive impairment tended to become more dependent in their ADLs, while those with no cognitive impairment tended to become more independent in their ADLs. This functional decline over time is perhaps expected in persons with cognitive impairment, but the significant difference with a large ES warrants further consideration on how to tailor interventions to better promote functional recovery in these individuals.^40,41 While some cognitive decline may not be preventable, this finding supports the need to promote healthy cognitive aging, identify declines in cognition, and work to mitigate additional decline. Programs specifically designed to promote function and physical activity in older adults with cognitive impairment are needed, especially during care transitions.^41-43

While participants reported that falls and pain interference did not have a significant impact on change in outcomes between baseline and endpoint, these areas need further investigation. Falls and pain have been associated with function and physical activity in older adults.^42-46 Pain is common, yet underappreciated during older adult hospital-to-home transitions.^11,12,45,46 There is a need for more comprehensive assessment of pain (including pain intensity) and qualitative research.

Hospitalization and postdischarge health care application factors may have a significant impact on home-telehealth physical activity intervention success. Individuals who were postsurgery tended to have greater improvements in ADLs and physical performance. Most postsurgery participants had joint replacement surgery. Postsurgery status may not be modifiable, but it is important to note expected differences in recovery between medical and surgical admissions and the need to tailor care based on admission diagnosis. Those with a longer length of hospital stay may be considered at higher risk of suboptimal outcomes postdischarge, which indicates an opportunity for targeting resources and support, in addition to efforts of reducing length of stay where possible.⁴⁷

Readmissions were significantly related to a change in Rosow-Breslau mobility disability score. This may indicate the detrimental impact a readmission can have on increasing mobility and physical activity postdischarge, or the potential of this pilot program to impact readmissions by increasing mobility and physical activity, contrary to prior physical exercise interventions.^5,7,9,48 With 5% to 79% of readmissions considered preventable, continued efforts and program dissemination and implementation to address preventable readmissions are warranted.⁴⁹ Individuals with postdischarge physical therapy (prior to beginning the pilot program) tended to demonstrate less improvement in disability and physical performance. This relationship needs further investigation; the 2 groups did not appear to have significant differences at baseline, albeit with a small sample size. It is possible they experienced initial improvements with postdischarge physical therapy and plateaued or had little further reserve to improve upon entering the VVC program.

Strengths and Limitations

This pilot program provided evaluative data on the use of VVC to enhance function and physical activity in older adults posthospital discharge. It included individual (eg, fall, pain, cognitive impairment) and health service (eg, readmission, physical therapy) level factors as predictors of function and physical activity posthospitalization.^5,7,9,15-19

The results of this pilot project stem from a small sample lacking diversity in terms of race, ethnicity, and sex. There was some variation in baseline and endpoints between participants, and when hospitalization, posthospital syndrome, and postdischarge health care application factors were collected. The majority of participants were recruited within a month postdischarge, and the program lasted about 6 months. Data collection was attempted at regular PAT contacts, but there was some variation in when visits occurred based on participant availability and preference. Some participants had missing data, which was handled using pairwise deletion.³³ Larger studies are needed to confirm the findings of this study, particularly the trends that did not reach statistical significance. Home health services other than physical therapy (eg, nursing, occupational therapy) were not fully accounted for and should be considered in future research.

Conclusions

In patients undergoing a 6-month pilot VVC-based physical activity intervention posthospital discharge, improvements in mobility and disability were most likely in those who had no cognitive impairment and were not readmitted. Larger sample and qualitative investigations are necessary to optimize outcomes for patients who meet these clinical profiles.

Deconditioning among hospitalized older adults contributes to significant decline in posthospitalization functional ability, physical performance, and physical activity.^1-10 Previous hospital-to-home interventions have targeted improving function and physical activity, including recent programs leveraging home telehealth as a feasible and potentially effective mode of delivering in-home exercise and rehabilitation.^11-14 However, pilot interventions have shown mixed effectiveness.^11,12,14 This study expands on a previously published intervention describing a pilot home telehealth program for veterans posthospital discharge that demonstrated significant 6-month improvement in physical activity as well as trends in physical function improvement, including among those with cognitive impairment.¹⁵ Factors that contribute to improved outcomes are the focus of the present study.

Key factors underlying the complexity of hospital-to-home transitions include hospitalization elements (ie, reason for admission and length of stay), associated posthospital syndromes (ie, postdischarge falls, medication changes, cognitive impairment, and pain), and postdischarge health care application (ie, physical therapy and hospital readmission).^16-18 These factors may be associated with postdischarge functional ability, physical performance, and physical activity, but their direct influence on intervention outcomes is unclear (Figure 1).^5,7,9,16-20 The objective of this study was to examine the influence of hospitalization, posthospital syndrome, and postdischarge health care application factors on outcomes of a US Department of Veterans Affairs (VA) Video Connect (VVC) intervention to enhance function and physical activity in older adults posthospital discharge.

Methods

The previous analysis reported on patient characteristics, program feasibility, and preliminary outcomes.^13,15 The current study reports on relationships between hospitalization, posthospital syndrome, and postdischarge health care application factors and change in key outcomes, namely postdischarge self-reported functional ability, physical performance, and physical activity from baseline to endpoint.

Participants provided written informed consent. The protocol and consent forms were approved by the VA Ann Arbor Healthcare System (VAAAHS) Research and Development Committee, and the project was registered on clinicaltrials.gov (NCT04045054).

Intervention

The pilot program targeted older adults following recent hospital discharge from VAAAHS. Participants were eligible if they were aged ≥ 50 years, had been discharged following an inpatient stay in the past 1 to 2 weeks, evaluated by physical therapy during hospitalization with stated rehabilitation goals on discharge, and followed by a VAAAHS primary care physician. Participants were either recruited during hospital admission or shortly after discharge.¹³

An experienced physical activity trainer (PAT) supported the progression of participants’ rehabilitation goals via a home exercise program and coached the patient and caregiver to optimize functional ability, physical performance, and physical activity. The PAT was a nonlicensed research assistant with extensive experience in applying standard physical activity enhancement protocols (eg, increased walking) to older adults with comorbidities. Participation in the program lasted about 6 months. Initiation of the PAT program was delayed if the patient was already receiving postdischarge home-based or outpatient physical therapy. The PAT contacted the patient weekly via VVC for the first 6 weeks, then monthly for a total of 6 months. Each contact included information on optimal walking form, injury prevention, program progression, and ways to incorporate sit-to-stand transitions, nonsitting behavior, and walking into daily routines. The initial VVC contact lasted about 60 minutes and subsequent sessions lasted about 30 minutes.¹³

Demographic characteristics were self-reported by participants and included age, sex, race, years of education, and marital status. Clinical characteristics were obtained from each participant’s electronic health record (EHR), including copay status, index hospitalization length of stay, admission diagnosis, and postsurgery status (postsurgery vs nonpostsurgery). Intervention adherence was tracked as the number of PAT sessions attended.

Posthospital Syndrome Factors

Participant falls (categorized as those who reported a fall vs those who did not) and medication changes (number of changes reported, including new medication, discontinued medication, dose changes, medication changes, or changes in medication schedule) were reported by participants or caregivers during each VVC contact. Participants completed the Montreal Cognitive Assessment (MoCA) at baseline, and were dichotomized into 2 groups: no cognitive impairment (MoCA score ≥ 26) and mild to moderate cognitive impairment (MoCA score 10-25).^13,21

Participants rated how much pain interfered with their normal daily activities since the previous VVC session on a 5-point Likert scale (1, not at all; to 5, extremely).²² Similar to prior research, participants were placed into 2 groups based on their mean pain interference score (individuals with scores from 1.0 to 2.0 in 1 group, and individuals with > 2.0 in another).^23-25 Participants were separated into a no or mild pain interference group and a moderate to severe pain interference group. Hospital readmissions (VA and non-VA) and postdischarge physical therapy outcomes were obtained from the participant’s EHR, including primary care visits.

Outcomes

Outcomes were collected at baseline (posthospital discharge) and 6 months postenrollment.

Self-Reported Functional Ability. This measure is provided by participants or caregivers and measured by the Katz Index of Independence in Activities of Daily Living (ADL), Lawton and Brody Instrumental ADL Scale (IADL), Nagi Disability Model, and Rosow-Breslau Scale. The Katz ADL assesses the ability to complete 6 self-care activities and awards 1 point for independence and 0 if the individual is dependent (total score range, 0-6).²⁶ The Lawton and Brody IADL measures an individual’s independence in 8 instrumental ADLs; it awards 1 point for independence and 0 if the individual is dependent (total score range, 0-8).²⁷ The Nagi Disability Model evaluates an individual’s difficulty performing 5 tasks (total score range, 0-5) and tallies the number of items with a response other than “no difficulty at all” (higher total score indicates greater difficulty). ²⁸ The Rosow-Breslau Scale is a 3-item measure of mobility disability; individual responses are 0 (no help) and 1 (requires help or unable); higher total score (range, 0-3) indicates greater disability.²⁹

Physical Performance. Measured using the Short Physical Performance Battery (SPPB), which evaluates standing balance, sit to stand, and walking performance. Scores range from 0 to 4 on the balance, gait speed, and chair stand tests, for a total composite score between 0 and 12 (higher score indicates better performance).³⁰

Physical Activity. Measured using actigraphy, namely a physical activity monitor adherent to the thigh (activ-PAL3TM, PAL Technologies Ltd., Glasgow, UK).³¹ Participants were instructed to wear the activPal for ≥ 1 week. Participants with a minimum of 5 days of wear were included in this analysis.

Data Analyses

Analyses were performed using SPSS software version 29.0.³² Continuous variables were summarized using mean (SD) or median and IQR using the weighted average method; categorical variables were summarized using frequencies and percentages. Baseline scores on outcome variables were compared by categorical hospitalization, posthospital syndrome, and postdischarge health care application factor variables using Mann-Whitney U tests. The differences between outcome variables from baseline to endpoint were then calculated to produce change scores. Relationships between the number of PAT sessions attended and baseline outcomes and outcome change scores were estimated using Spearman correlations. Relationships between categorical factors (hospitalization, posthospital syndrome, and postdischarge health care application) and outcome variable change scores (which were normally distributed) were examined using Mann-Whitney U tests. Relationships with continuous hospitalization (length of stay) and posthospital syndrome factors (medication changes) were estimated using Spearman correlations. Effect sizes (ES) were estimated with Cohen d; small (d = 0.2), medium (d = 0.5), or large (d ≥ 0.8). Missing data were handled using pairwise deletion.³³ Therefore, sample sizes were reported for each analysis. For all statistical tests, P < .05 was considered significant.

Results

Twenty-four individuals completed the pilot intervention.¹⁵ Mean (SD) age was 73.6 (8.1) years (range, 64-93 years) and participants were predominantly White males (Table 1). Eight participants had a high school education only and 13 had more than a high school education. Diagnoses at admission included 9 patients with orthopedic/musculoskeletal conditions (6 were for joint replacement), 6 patients with vascular/pulmonary conditions, and 4 with gastrointestinal/renal/urological conditions. Of the 11 postsurgery participants, 7 were orthopedic, 4 were gastrointestinal, and 1 was peripheral vascular.

Baseline outcome scores did not differ significantly between groups, except individuals with moderate to severe pain interference reported a significantly lower IADL score (median [IQR] 4 [2-7]) than individuals with mild or moderate pain interference (median [IQR] 8 [7-8]; P = .02) (Table 2). The mean (SD) number of PAT sessions attended was 9.3 (3.7) (range, 3-19). There were no significant relationships between number of sessions attended and any baseline outcome variables or outcome change scores.

Hospitalization Factors

Participants who were postsurgery tended to have greater improvement than individuals who were nonpostsurgery in ADLs (median [IQR] 0 [0-1.5]; ES, 0.6; P = .10) and SPPB (median [IQR] 2 [1.5-9]; ES, 0.9; P = .07), but the improvements were not statistically significant (Table 3). Mean (SD) length of stay of the index hospitalization was 6.7 (6.1) days. Longer length of stay was significantly correlated with an increase in Nagi score (ρ, 0.45; 95% CI, 0.01-0.75). There were no other significant or trending relationships between length of stay and outcome variables.

Posthospital Syndrome Factors

The 16 participants with mild to moderate cognitive impairment had less improvement in ADLs (median [IQR] 0 [0-1]) than the 8 participants with no impairment (median [IQR] 0 [-0.75 to 0]; ES, -1.1; P = .04). Change in outcome variables from baseline to endpoint did not significantly differ between the 8 patients who reported a fall compared with the 13 who did not, nor were any trends observed. Change in outcome variables from baseline to endpoint also did not significantly differ between the 8 participants who reported no or mild pain interference compared with the 10 patients with moderate to severe pain interference, nor were any trends observed. Mean (SD) number of medication changes was 2.5 (1.6). Higher number of medication changes was significantly correlated with a decrease in Rosow-Breslau score (ρ, -0.47; 95% CI, -0.76 to -0.02). There were no other significant or trending relationships between number of medication changes and outcome variables.

Postdischarge Health Care Application Factors

The 16 participants who attended posthospital physical therapy trended towards less improvement in IADLs (median [IQR] 0 [-0.5 to 1.5]; ES, -0.7; P = .11) and SPPB (median [IQR] 2 [-3.0 to 4.5]; ES, -0.5; P = .15) than the 8 patients with no postdischarge physical therapy. Eleven participants were readmitted, while 13 had no readmissions in their medical records between baseline and endpoint. Participants with ≥ 1 readmission experienced a greater increase in Rosow-Breslau score (median [IQR] 0 [-0.5 to 1.0]) than those not readmitted (median [IQR] 0 [-1.25 to 0.25]; ES, 1.0; P = .03). Borderline greater improvement in number of steps was found in those not readmitted (median [IQR] 3365.6 [274.4-7710.9]) compared with those readmitted (median [IQR] 319.9 [-136.1 to 774.5]; ES, -1.3; P = .05). Patients who were readmitted also tended to have lower and not statistically significant improvements in SPPB (median [IQR] 1 [-4.0 to 5.3]) compared with those not readmitted (median [IQR] 2 [0.3-3.8]; ES, -0.5; P = .17) (Table 3).

Discussion

This study examined the association between hospitalization, posthospital syndrome, and postdischarge health care use in patients undergoing a VVC-based intervention following hospital discharge. Participants who had no or mild cognitive impairment, no readmissions, higher medication changes, and a shorter hospital length of stay tended to experience lower disability, including in mobility and ADLs. This suggests individuals who are less clinically complex may be more likely to benefit from this type of virtual rehabilitation program. These findings are consistent with clinical experiences; home-based programs to improve physical activity posthospital discharge can be challenging for those who were medically ill (and did not undergo a specific surgical procedure), cognitively impaired, and become acutely ill and trigger hospital readmission. ¹⁵ For example, the sample in this study had higher rates of falls, pain, and readmissions compared to previous research.^2,3,34-39

The importance of posthospital syndrome in the context of recovery of function and health at home following hospitalization is well documented.^16-18 The potential impact of posthospital syndrome on physical activity-focused interventions is less understood. In our analysis, participants with mild or moderate cognitive impairment tended to become more dependent in their ADLs, while those with no cognitive impairment tended to become more independent in their ADLs. This functional decline over time is perhaps expected in persons with cognitive impairment, but the significant difference with a large ES warrants further consideration on how to tailor interventions to better promote functional recovery in these individuals.^40,41 While some cognitive decline may not be preventable, this finding supports the need to promote healthy cognitive aging, identify declines in cognition, and work to mitigate additional decline. Programs specifically designed to promote function and physical activity in older adults with cognitive impairment are needed, especially during care transitions.^41-43

While participants reported that falls and pain interference did not have a significant impact on change in outcomes between baseline and endpoint, these areas need further investigation. Falls and pain have been associated with function and physical activity in older adults.^42-46 Pain is common, yet underappreciated during older adult hospital-to-home transitions.^11,12,45,46 There is a need for more comprehensive assessment of pain (including pain intensity) and qualitative research.

Hospitalization and postdischarge health care application factors may have a significant impact on home-telehealth physical activity intervention success. Individuals who were postsurgery tended to have greater improvements in ADLs and physical performance. Most postsurgery participants had joint replacement surgery. Postsurgery status may not be modifiable, but it is important to note expected differences in recovery between medical and surgical admissions and the need to tailor care based on admission diagnosis. Those with a longer length of hospital stay may be considered at higher risk of suboptimal outcomes postdischarge, which indicates an opportunity for targeting resources and support, in addition to efforts of reducing length of stay where possible.⁴⁷

Readmissions were significantly related to a change in Rosow-Breslau mobility disability score. This may indicate the detrimental impact a readmission can have on increasing mobility and physical activity postdischarge, or the potential of this pilot program to impact readmissions by increasing mobility and physical activity, contrary to prior physical exercise interventions.^5,7,9,48 With 5% to 79% of readmissions considered preventable, continued efforts and program dissemination and implementation to address preventable readmissions are warranted.⁴⁹ Individuals with postdischarge physical therapy (prior to beginning the pilot program) tended to demonstrate less improvement in disability and physical performance. This relationship needs further investigation; the 2 groups did not appear to have significant differences at baseline, albeit with a small sample size. It is possible they experienced initial improvements with postdischarge physical therapy and plateaued or had little further reserve to improve upon entering the VVC program.

Strengths and Limitations

This pilot program provided evaluative data on the use of VVC to enhance function and physical activity in older adults posthospital discharge. It included individual (eg, fall, pain, cognitive impairment) and health service (eg, readmission, physical therapy) level factors as predictors of function and physical activity posthospitalization.^5,7,9,15-19

The results of this pilot project stem from a small sample lacking diversity in terms of race, ethnicity, and sex. There was some variation in baseline and endpoints between participants, and when hospitalization, posthospital syndrome, and postdischarge health care application factors were collected. The majority of participants were recruited within a month postdischarge, and the program lasted about 6 months. Data collection was attempted at regular PAT contacts, but there was some variation in when visits occurred based on participant availability and preference. Some participants had missing data, which was handled using pairwise deletion.³³ Larger studies are needed to confirm the findings of this study, particularly the trends that did not reach statistical significance. Home health services other than physical therapy (eg, nursing, occupational therapy) were not fully accounted for and should be considered in future research.

Conclusions

In patients undergoing a 6-month pilot VVC-based physical activity intervention posthospital discharge, improvements in mobility and disability were most likely in those who had no cognitive impairment and were not readmitted. Larger sample and qualitative investigations are necessary to optimize outcomes for patients who meet these clinical profiles.

More than 38 million people in the United States (12%) have diabetes mellitus (DM), though 1 in 5 are unaware they have DM.¹ The prevalence among veterans is even more substantial, impacting nearly 25% of those who received care from the US Department of Veterans Affairs (VA).² DM can lead to increased health care costs in addition to various complications (eg, cardiovascular, renal), especially if left uncontrolled.^1,3 similar impact is found in the perioperative period (defined as at or around the time of an operation), as multiple studies have found that uncontrolled preoperative DM can result in worsened surgical outcomes, including longer hospital stays, more infectious complications, and higher perioperative mortality.^4-6

In contrast, adequate glycemic control assessed with blood glucose levels has been shown to decrease the incidence of postoperative infections.⁷ Optimizing glycemic control during hospital stays, especially postsurgery, has become the standard of care, with most health systems establishing specific protocols. In current literature, most studies examining DM management in the perioperative period are focused on postoperative care, with little attention to the preoperative period.^4,6,7

One study found that patients with poor presurgery glycemic control assessed by hemoglobin A_1c (HbA_1c) levels were more likely to remain hyperglycemic during and after surgery. 8 Blood glucose levels < 200 mg/dL can lead to an increased risk of infection and impaired wound healing, meaning a well-controlled HbA_1c before a procedure serves as a potential factor for success.⁹ The 2025 American Diabetes Association (ADA) Standards of Care (SOC) recommendation is to target HbA_1c < 8% whenever possible, and some health systems require lower levels (eg, < 7% or 7.5%).¹⁰ With that goal in mind and knowing that preoperative hyperglycemia has been shown to be a contributing factor in the delay or cancellation of surgical cases, an argument can be made that attention to preoperative DM management also should be a focus for health care systems performing surgeries.^8,9,11

Attention to glucose control during preoperative care offers an opportunity to screen for DM in patients who may not have been screened otherwise and to standardize perioperative DM management. Since DM disproportionately impacts veterans, this is a pertinent issue to the VA. Veterans can be more susceptible to complications if DM is left uncontrolled prior to surgery. To determine readiness for surgery and control of comorbid conditions such as DM before a planned surgery, facilities often perform a preoperative clinic assessment, often in a multidisciplinary clinic.

At Veteran Health Indiana (VHI), a presurgery clinic visit involving the primary surgery service (physician, nurse practitioner, and/or a physician assistant) is conducted 1 to 2 months prior to the planned procedure to determine whether a patient is ready for surgery. During this visit, patients receive a packet with instructions for various tasks and medications, such as applying topical antibiotic prophylaxis on the anticipated surgical site. This is documented in the form of a note in the VHI Computerized Patient Record System (CPRS). The medication instructions are provided according to the preferences of the surgical team. These may be templated notes that contain general directions on the timing and dosing of specific medications, in addition to instructions for holding or reducing doses when appropriate. The instructions can be tailored by the team conducting the preoperative visit (eg, “Take 20 units of insulin glargine the day before surgery” vs “Take half of your long-acting insulin the night before surgery”). Specific to DM, VHI has a nurse-driven day of surgery glucose assessment where point-of-care blood glucose is collected during preoperative holding for most patients.

There is limited research assessing the level of preoperative glycemic control and the incidence of complications in a veteran population. The objective of this study was to gain a baseline understanding of what, if any, standardization exists for preoperative instructions for DM medications and to assess the level of preoperative glycemic control and postoperative complications in patients with DM undergoing major elective surgical procedures.

Methods

This retrospective, single-center chart review was conducted at VHI. The Indiana University and VHI institutional review boards determined that this quality improvement project was exempt from review.

The primary outcome was the number of patients with surgical procedures delayed or canceled due to hyperglycemia or hypoglycemia. Hyperglycemia was defined as blood glucose > 180 mg/dL and hypoglycemia was defined as < 70 mg/dL, slight variations from the current ADA SOC preoperative specific recommendation of a blood glucose reading of 100 to 180 mg/dL within 4 hours of surgery.¹⁰ The standard outpatient hypoglycemia definition of blood glucose < 70 mg/dL was chosen because the current goal (< 100 mg/dL) was not the standard in previous ADA SOCs that were in place during the study period. Specifically, the 2018 ADA SOC did not provide preoperative recommendations and the 2019-2021 ADA SOC recommended 80 to 180 mg/dL.^10,12-18 For patients who had multiple preoperative blood glucose measurements, the first recorded glucose on the day of the procedure was used.

The secondary outcomes of this study were focused on the preoperative process/care at VHI and postoperative glycemic control. The preoperative process included examining whether medication instructions were given and their quality. Additionally, the number of interventions for hyperglycemia and hypoglycemia were required immediately prior to surgery and the average preoperative HbA_1c (measured within 3 months prior to surgery) were collected and analyzed. For postoperative glycemic control, average blood glucose measurements and number of hypoglycemic (< 70 mg/dL) and hyperglycemic (> 180 mg/dL) events were measured in addition to the frequency of changes made at discharge to patients’ DM medication regimens.

The safety outcome of this study assessed commonly observed postoperative complications and was examined up to 30 days postsurgery. These included acute kidney injury (defined using Kidney Disease: Improving Global Outcomes 2012, the standard during the study period), nonfatal myocardial infarction, nonfatal stroke, and surgical site infections, which were identified from the discharge summary written by the primary surgery service.¹⁹ All-cause mortality also was collected.

Patients were included if they were admitted for major elective surgeries and had a diagnosis of either type 1 or type 2 DM on their problem list, determined by International Classification of Diseases, Tenth Revision codes. Major elective surgery was defined as a procedure that would likely result in a hospital admission of > 24 hours. Of note, patients may have been included in this study more than once if they had > 1 procedure at least 30 days apart and met inclusion criteria within the time frame. Patients were excluded if they were taking no DM medications or chronic steroids (at any dose), residing in a long-term care facility, being managed by a non-VA clinician prior to surgery, or missing a preoperative blood glucose measurement.

All data were collected from the CPRS. A list of surgical cases involving patients with DM who were scheduled to undergo major elective surgeries from January 1, 2018, to December 31, 2021, at VHI was generated. The list was randomized to a smaller number (N = 394) for data collection due to the time and resource constraints for a pharmacy residency project. All data were deidentified and stored in a secured VA server to protect patient confidentiality. Descriptive statistics were used for all results.

Results

Initially, 2362 surgeries were identified. A randomized sample of 394 charts were reviewed and 131 cases met inclusion criteria. Each case involved a unique patient (Figure). The most common reasons for exclusion were 143 patients with diet-controlled DM and 78 nonelective surgeries. The mean (SD) age of patients was 68 (8) years, and the most were male (98.5%) and White (76.3%) (Table 1). 

At baseline, 45 of 131 patients (34.4%) had coronary artery disease and 29 (22.1%) each had autonomic neuropathy and chronic kidney disease. Most surgeries were conducted by orthopedic (32.1%) and peripheral vascular (21.4%) specialties. The mean (SD) length of surgery was 4.6 (2.6) hours and of hospital length of stay was 4 (4) days. No patients stayed longer than the 30-day safety outcome follow-up period. All patients had type 2 DM and took a mean 2 DM medications. The 63 patients taking insulin had a mean (SD) total daily dose of 99 (77) U (Table 2). A preoperative HbA1c was collected in 116 patients within 3 months of surgery, with a mean HbA_1c of 7.0% (range, 5.3-10.7).

No patients had surgeries delayed or canceled because of uncontrolled DM on the day of surgery. The mean preoperative blood glucose level was 146 mg/dL (range, 73-365) (Table 3). No patients had a preoperative blood glucose level of < 70 mg/dL and 19 (14.5%) had a blood glucose level > 180 mg/dL. Among patients with hyperglycemia immediately prior to surgery, 6 (31.6%) had documentation of insulin being provided.

For this sample of patients, the preoperative clinic visit was conducted a mean 22 days prior to the planned surgery date. Among the 131 included patients, 122 (93.1%) had documentation of receiving instructions for DM medications. Among patients who had documented receipt of instructions, only 30 (24.6%) had instructions specifically tailored to their regimen rather than a generic templated form. The mean (SD) preoperative blood glucose was similar for those who received specific perioperative DM instructions at 146 (50) mg/dL when compared with those who did not at 147 (45) mg/dL. The mean (SD) preoperative blood glucose reading for those who had no documentation of receipt of perioperative instructions was 126 (54) mg/dL compared with 147 (46) mg/dL for those who did.

The mean number of postoperative blood glucose events per day was negligible for hypoglycemia and more frequent for hyperglycemia with a mean of 2 events per day. The mean postoperative blood glucose range was 121 to 247 mg/dL with most readings < 180 mg/dL. Upon discharge, most patients continued their home DM regimen with 5 patients (3.8%) having changes made to their regimen upon discharge.

Very few postoperative complications were identified from chart review. The most frequently observed postoperative complications were acute kidney injury, surgical site infections, and nonfatal stroke. There were no documented nonfatal myocardial infarctions. Two patients (1.5%) died within 30 days of the surgery; neither death was deemed to have been related to poor perioperative glycemic control.

Discussion

To our knowledge, this retrospective chart review was the first study to assess preoperative DM management and postoperative complications in a veteran population. VHI is a large, tertiary, level 1a, academic medical center that serves approximately 62,000 veterans annually and performs about 5000 to 6000 surgeries annually, a total that is increasing following the COVID-19 pandemic.²⁰ This study found that the current process of a presurgery clinic visit and day of surgery glucose assessment has prevented surgical delays or cancellations.

Most patients included in this study were well controlled at baseline in accordance with the 2025 ADA SOC HbA_1c recommendation of a preoperative HbA1c of < 8%, which may have contributed to no surgical delays or cancellations.¹⁰ However, not all patients had HbA1c collected within 3 months of surgery or even had one collected at all. Despite the ADA SOC providing no explicit recommendation for universal HbA_1c screening prior to elective procedures, its importance cannot be understated given the body of evidence demonstrating poor outcomes with uncontrolled preoperative DM.^8,10 The glycemic control at baseline may have contributed to the very few postsurgical complications observed in this study.

Although the current process at VHI prevented surgical delays and cancellations in this sample, there are still identified areas for improvement. One area is the instructions the patients received. Patients with DM are often prescribed ≥ 1 medication or a combination of insulins, noninsulin injectables, and oral DM medications, and this study population was no different. Because these medications may influence the anesthesia and perioperative periods, the ADA has specific guidance for altering administration schedules in the days leading up to surgery.¹⁰

Inappropriate administration of DM medications could lead to perioperative hypoglycemia or hyperglycemia, possibly causing surgical delays, case cancellations, and/or postoperative complications.21 Although these data reveal the specificity and documented receipt that the preoperative DM instructions did not impact the first recorded preoperative blood glucose, future studies should examine patient confidence in how to properly administer their DM medications prior to surgery. It is vital that patients receive clear instructions in accordance with the ADA SOC on whether to continue, hold, or adjust the dose of their medications to prevent fluctuations in blood glucose levels in the perioperative period, ensure safety with anesthesia, and prevent postoperative complications such as acute kidney injury. Of note, compliance with guideline recommendations for medication instructions was not examined because the data collection time frame expanded over multiple years and the recommendations have evolved each year as new data emerge.

Preoperative DM Management

The first key takeaway from this study is to ensure patients are ready for surgery with a formal assessment (typically in the form of a clinic visit) prior to the surgery. One private sector health system published their approach to this by administering an automatic preoperative HbA_1c screening for those with a DM diagnosis and all patients with a random plasma glucose ≥ 200 mg/dL.²² Additionally, if the patient's HbA_1c level was not at goal prior to surgery (≥ 8% for those with known DM and ≥ 6.5% with no known DM), patients were referred to endocrinology for further management. Increasing attention to the preoperative visit and extending HbA_1c testing to all patients regardless of DM status also provides an opportunity to identify individuals living with undiagnosed DM.¹

Even though there was no difference in the mean preoperative blood glucose level based on receipt or specificity of preoperative DM instructions, a second takeaway from this study is the importance of ensuring patients receive clear instructions on their DM medication schedule in the perioperative period. A practical first step may be updating the templates used by the primary surgery teams and providing education to the clinicians in the clinic on how to personalize the visits. Because the current preoperative DM process at VHI is managed by the primary surgical team in a clinic visit, there is an opportunity to shift this responsibility to other health care professionals, such as pharmacists—a change shown to reduce unintended omission of home medications following surgery during hospitalization and reduce costs.^23,24

Limitations

This study relied on data included in the patient chart. These data include medication interventions made immediately prior to surgery, which can sometimes be inaccurately charted or difficult to find as they are not documented in the typical medication administration record. Also, the safety outcomes were collected from a discharge summary written by different clinicians, which may lead to information bias. Special attention was taken to ensure these data points were collected as accurately as possible, but it is possible some data may be inaccurate from unintentional human error. Additionally, the safety outcome was limited to a 30-day follow-up, but encompassed the entire length of postoperative stay for all included patients. Finally, given this study was retrospective with no comparison group and the intent was to improve processes at VHI, only hypotheses and potential interventions can be generated from this study. Future prospective studies with larger sample sizes and comparator groups are needed to draw further conclusions.

Conclusions

This study found that the current presurgery process at VHI appears to be successful in preventing surgical delays or cancellations due to hyperglycemia or hypoglycemia. Optimizing DM management can improve surgical outcomes by decreasing rates of postoperative complications, and this study added additional evidence in support of that in a unique population: veterans. Insight on the awareness of preoperative blood glucose management should be gleaned from this study, and based on this sample and site, the preadmission screening process and instructions provided to patients can serve as 2 starting points for optimizing elective surgery.

More than 38 million people in the United States (12%) have diabetes mellitus (DM), though 1 in 5 are unaware they have DM.¹ The prevalence among veterans is even more substantial, impacting nearly 25% of those who received care from the US Department of Veterans Affairs (VA).² DM can lead to increased health care costs in addition to various complications (eg, cardiovascular, renal), especially if left uncontrolled.^1,3 similar impact is found in the perioperative period (defined as at or around the time of an operation), as multiple studies have found that uncontrolled preoperative DM can result in worsened surgical outcomes, including longer hospital stays, more infectious complications, and higher perioperative mortality.^4-6

In contrast, adequate glycemic control assessed with blood glucose levels has been shown to decrease the incidence of postoperative infections.⁷ Optimizing glycemic control during hospital stays, especially postsurgery, has become the standard of care, with most health systems establishing specific protocols. In current literature, most studies examining DM management in the perioperative period are focused on postoperative care, with little attention to the preoperative period.^4,6,7

One study found that patients with poor presurgery glycemic control assessed by hemoglobin A_1c (HbA_1c) levels were more likely to remain hyperglycemic during and after surgery. 8 Blood glucose levels < 200 mg/dL can lead to an increased risk of infection and impaired wound healing, meaning a well-controlled HbA_1c before a procedure serves as a potential factor for success.⁹ The 2025 American Diabetes Association (ADA) Standards of Care (SOC) recommendation is to target HbA_1c < 8% whenever possible, and some health systems require lower levels (eg, < 7% or 7.5%).¹⁰ With that goal in mind and knowing that preoperative hyperglycemia has been shown to be a contributing factor in the delay or cancellation of surgical cases, an argument can be made that attention to preoperative DM management also should be a focus for health care systems performing surgeries.^8,9,11

Attention to glucose control during preoperative care offers an opportunity to screen for DM in patients who may not have been screened otherwise and to standardize perioperative DM management. Since DM disproportionately impacts veterans, this is a pertinent issue to the VA. Veterans can be more susceptible to complications if DM is left uncontrolled prior to surgery. To determine readiness for surgery and control of comorbid conditions such as DM before a planned surgery, facilities often perform a preoperative clinic assessment, often in a multidisciplinary clinic.

At Veteran Health Indiana (VHI), a presurgery clinic visit involving the primary surgery service (physician, nurse practitioner, and/or a physician assistant) is conducted 1 to 2 months prior to the planned procedure to determine whether a patient is ready for surgery. During this visit, patients receive a packet with instructions for various tasks and medications, such as applying topical antibiotic prophylaxis on the anticipated surgical site. This is documented in the form of a note in the VHI Computerized Patient Record System (CPRS). The medication instructions are provided according to the preferences of the surgical team. These may be templated notes that contain general directions on the timing and dosing of specific medications, in addition to instructions for holding or reducing doses when appropriate. The instructions can be tailored by the team conducting the preoperative visit (eg, “Take 20 units of insulin glargine the day before surgery” vs “Take half of your long-acting insulin the night before surgery”). Specific to DM, VHI has a nurse-driven day of surgery glucose assessment where point-of-care blood glucose is collected during preoperative holding for most patients.

There is limited research assessing the level of preoperative glycemic control and the incidence of complications in a veteran population. The objective of this study was to gain a baseline understanding of what, if any, standardization exists for preoperative instructions for DM medications and to assess the level of preoperative glycemic control and postoperative complications in patients with DM undergoing major elective surgical procedures.

Methods

This retrospective, single-center chart review was conducted at VHI. The Indiana University and VHI institutional review boards determined that this quality improvement project was exempt from review.

The primary outcome was the number of patients with surgical procedures delayed or canceled due to hyperglycemia or hypoglycemia. Hyperglycemia was defined as blood glucose > 180 mg/dL and hypoglycemia was defined as < 70 mg/dL, slight variations from the current ADA SOC preoperative specific recommendation of a blood glucose reading of 100 to 180 mg/dL within 4 hours of surgery.¹⁰ The standard outpatient hypoglycemia definition of blood glucose < 70 mg/dL was chosen because the current goal (< 100 mg/dL) was not the standard in previous ADA SOCs that were in place during the study period. Specifically, the 2018 ADA SOC did not provide preoperative recommendations and the 2019-2021 ADA SOC recommended 80 to 180 mg/dL.^10,12-18 For patients who had multiple preoperative blood glucose measurements, the first recorded glucose on the day of the procedure was used.

The secondary outcomes of this study were focused on the preoperative process/care at VHI and postoperative glycemic control. The preoperative process included examining whether medication instructions were given and their quality. Additionally, the number of interventions for hyperglycemia and hypoglycemia were required immediately prior to surgery and the average preoperative HbA_1c (measured within 3 months prior to surgery) were collected and analyzed. For postoperative glycemic control, average blood glucose measurements and number of hypoglycemic (< 70 mg/dL) and hyperglycemic (> 180 mg/dL) events were measured in addition to the frequency of changes made at discharge to patients’ DM medication regimens.

The safety outcome of this study assessed commonly observed postoperative complications and was examined up to 30 days postsurgery. These included acute kidney injury (defined using Kidney Disease: Improving Global Outcomes 2012, the standard during the study period), nonfatal myocardial infarction, nonfatal stroke, and surgical site infections, which were identified from the discharge summary written by the primary surgery service.¹⁹ All-cause mortality also was collected.

Patients were included if they were admitted for major elective surgeries and had a diagnosis of either type 1 or type 2 DM on their problem list, determined by International Classification of Diseases, Tenth Revision codes. Major elective surgery was defined as a procedure that would likely result in a hospital admission of > 24 hours. Of note, patients may have been included in this study more than once if they had > 1 procedure at least 30 days apart and met inclusion criteria within the time frame. Patients were excluded if they were taking no DM medications or chronic steroids (at any dose), residing in a long-term care facility, being managed by a non-VA clinician prior to surgery, or missing a preoperative blood glucose measurement.

All data were collected from the CPRS. A list of surgical cases involving patients with DM who were scheduled to undergo major elective surgeries from January 1, 2018, to December 31, 2021, at VHI was generated. The list was randomized to a smaller number (N = 394) for data collection due to the time and resource constraints for a pharmacy residency project. All data were deidentified and stored in a secured VA server to protect patient confidentiality. Descriptive statistics were used for all results.

Results

Initially, 2362 surgeries were identified. A randomized sample of 394 charts were reviewed and 131 cases met inclusion criteria. Each case involved a unique patient (Figure). The most common reasons for exclusion were 143 patients with diet-controlled DM and 78 nonelective surgeries. The mean (SD) age of patients was 68 (8) years, and the most were male (98.5%) and White (76.3%) (Table 1). 

At baseline, 45 of 131 patients (34.4%) had coronary artery disease and 29 (22.1%) each had autonomic neuropathy and chronic kidney disease. Most surgeries were conducted by orthopedic (32.1%) and peripheral vascular (21.4%) specialties. The mean (SD) length of surgery was 4.6 (2.6) hours and of hospital length of stay was 4 (4) days. No patients stayed longer than the 30-day safety outcome follow-up period. All patients had type 2 DM and took a mean 2 DM medications. The 63 patients taking insulin had a mean (SD) total daily dose of 99 (77) U (Table 2). A preoperative HbA1c was collected in 116 patients within 3 months of surgery, with a mean HbA_1c of 7.0% (range, 5.3-10.7).

No patients had surgeries delayed or canceled because of uncontrolled DM on the day of surgery. The mean preoperative blood glucose level was 146 mg/dL (range, 73-365) (Table 3). No patients had a preoperative blood glucose level of < 70 mg/dL and 19 (14.5%) had a blood glucose level > 180 mg/dL. Among patients with hyperglycemia immediately prior to surgery, 6 (31.6%) had documentation of insulin being provided.

For this sample of patients, the preoperative clinic visit was conducted a mean 22 days prior to the planned surgery date. Among the 131 included patients, 122 (93.1%) had documentation of receiving instructions for DM medications. Among patients who had documented receipt of instructions, only 30 (24.6%) had instructions specifically tailored to their regimen rather than a generic templated form. The mean (SD) preoperative blood glucose was similar for those who received specific perioperative DM instructions at 146 (50) mg/dL when compared with those who did not at 147 (45) mg/dL. The mean (SD) preoperative blood glucose reading for those who had no documentation of receipt of perioperative instructions was 126 (54) mg/dL compared with 147 (46) mg/dL for those who did.

The mean number of postoperative blood glucose events per day was negligible for hypoglycemia and more frequent for hyperglycemia with a mean of 2 events per day. The mean postoperative blood glucose range was 121 to 247 mg/dL with most readings < 180 mg/dL. Upon discharge, most patients continued their home DM regimen with 5 patients (3.8%) having changes made to their regimen upon discharge.

Very few postoperative complications were identified from chart review. The most frequently observed postoperative complications were acute kidney injury, surgical site infections, and nonfatal stroke. There were no documented nonfatal myocardial infarctions. Two patients (1.5%) died within 30 days of the surgery; neither death was deemed to have been related to poor perioperative glycemic control.

Discussion

To our knowledge, this retrospective chart review was the first study to assess preoperative DM management and postoperative complications in a veteran population. VHI is a large, tertiary, level 1a, academic medical center that serves approximately 62,000 veterans annually and performs about 5000 to 6000 surgeries annually, a total that is increasing following the COVID-19 pandemic.²⁰ This study found that the current process of a presurgery clinic visit and day of surgery glucose assessment has prevented surgical delays or cancellations.

Most patients included in this study were well controlled at baseline in accordance with the 2025 ADA SOC HbA_1c recommendation of a preoperative HbA1c of < 8%, which may have contributed to no surgical delays or cancellations.¹⁰ However, not all patients had HbA1c collected within 3 months of surgery or even had one collected at all. Despite the ADA SOC providing no explicit recommendation for universal HbA_1c screening prior to elective procedures, its importance cannot be understated given the body of evidence demonstrating poor outcomes with uncontrolled preoperative DM.^8,10 The glycemic control at baseline may have contributed to the very few postsurgical complications observed in this study.

Although the current process at VHI prevented surgical delays and cancellations in this sample, there are still identified areas for improvement. One area is the instructions the patients received. Patients with DM are often prescribed ≥ 1 medication or a combination of insulins, noninsulin injectables, and oral DM medications, and this study population was no different. Because these medications may influence the anesthesia and perioperative periods, the ADA has specific guidance for altering administration schedules in the days leading up to surgery.¹⁰

Inappropriate administration of DM medications could lead to perioperative hypoglycemia or hyperglycemia, possibly causing surgical delays, case cancellations, and/or postoperative complications.21 Although these data reveal the specificity and documented receipt that the preoperative DM instructions did not impact the first recorded preoperative blood glucose, future studies should examine patient confidence in how to properly administer their DM medications prior to surgery. It is vital that patients receive clear instructions in accordance with the ADA SOC on whether to continue, hold, or adjust the dose of their medications to prevent fluctuations in blood glucose levels in the perioperative period, ensure safety with anesthesia, and prevent postoperative complications such as acute kidney injury. Of note, compliance with guideline recommendations for medication instructions was not examined because the data collection time frame expanded over multiple years and the recommendations have evolved each year as new data emerge.

Preoperative DM Management

The first key takeaway from this study is to ensure patients are ready for surgery with a formal assessment (typically in the form of a clinic visit) prior to the surgery. One private sector health system published their approach to this by administering an automatic preoperative HbA_1c screening for those with a DM diagnosis and all patients with a random plasma glucose ≥ 200 mg/dL.²² Additionally, if the patient's HbA_1c level was not at goal prior to surgery (≥ 8% for those with known DM and ≥ 6.5% with no known DM), patients were referred to endocrinology for further management. Increasing attention to the preoperative visit and extending HbA_1c testing to all patients regardless of DM status also provides an opportunity to identify individuals living with undiagnosed DM.¹

Even though there was no difference in the mean preoperative blood glucose level based on receipt or specificity of preoperative DM instructions, a second takeaway from this study is the importance of ensuring patients receive clear instructions on their DM medication schedule in the perioperative period. A practical first step may be updating the templates used by the primary surgery teams and providing education to the clinicians in the clinic on how to personalize the visits. Because the current preoperative DM process at VHI is managed by the primary surgical team in a clinic visit, there is an opportunity to shift this responsibility to other health care professionals, such as pharmacists—a change shown to reduce unintended omission of home medications following surgery during hospitalization and reduce costs.^23,24

Limitations

This study relied on data included in the patient chart. These data include medication interventions made immediately prior to surgery, which can sometimes be inaccurately charted or difficult to find as they are not documented in the typical medication administration record. Also, the safety outcomes were collected from a discharge summary written by different clinicians, which may lead to information bias. Special attention was taken to ensure these data points were collected as accurately as possible, but it is possible some data may be inaccurate from unintentional human error. Additionally, the safety outcome was limited to a 30-day follow-up, but encompassed the entire length of postoperative stay for all included patients. Finally, given this study was retrospective with no comparison group and the intent was to improve processes at VHI, only hypotheses and potential interventions can be generated from this study. Future prospective studies with larger sample sizes and comparator groups are needed to draw further conclusions.

Conclusions

This study found that the current presurgery process at VHI appears to be successful in preventing surgical delays or cancellations due to hyperglycemia or hypoglycemia. Optimizing DM management can improve surgical outcomes by decreasing rates of postoperative complications, and this study added additional evidence in support of that in a unique population: veterans. Insight on the awareness of preoperative blood glucose management should be gleaned from this study, and based on this sample and site, the preadmission screening process and instructions provided to patients can serve as 2 starting points for optimizing elective surgery.

More than 38 million people in the United States (12%) have diabetes mellitus (DM), though 1 in 5 are unaware they have DM.¹ The prevalence among veterans is even more substantial, impacting nearly 25% of those who received care from the US Department of Veterans Affairs (VA).² DM can lead to increased health care costs in addition to various complications (eg, cardiovascular, renal), especially if left uncontrolled.^1,3 similar impact is found in the perioperative period (defined as at or around the time of an operation), as multiple studies have found that uncontrolled preoperative DM can result in worsened surgical outcomes, including longer hospital stays, more infectious complications, and higher perioperative mortality.^4-6

In contrast, adequate glycemic control assessed with blood glucose levels has been shown to decrease the incidence of postoperative infections.⁷ Optimizing glycemic control during hospital stays, especially postsurgery, has become the standard of care, with most health systems establishing specific protocols. In current literature, most studies examining DM management in the perioperative period are focused on postoperative care, with little attention to the preoperative period.^4,6,7

One study found that patients with poor presurgery glycemic control assessed by hemoglobin A_1c (HbA_1c) levels were more likely to remain hyperglycemic during and after surgery. 8 Blood glucose levels < 200 mg/dL can lead to an increased risk of infection and impaired wound healing, meaning a well-controlled HbA_1c before a procedure serves as a potential factor for success.⁹ The 2025 American Diabetes Association (ADA) Standards of Care (SOC) recommendation is to target HbA_1c < 8% whenever possible, and some health systems require lower levels (eg, < 7% or 7.5%).¹⁰ With that goal in mind and knowing that preoperative hyperglycemia has been shown to be a contributing factor in the delay or cancellation of surgical cases, an argument can be made that attention to preoperative DM management also should be a focus for health care systems performing surgeries.^8,9,11

Attention to glucose control during preoperative care offers an opportunity to screen for DM in patients who may not have been screened otherwise and to standardize perioperative DM management. Since DM disproportionately impacts veterans, this is a pertinent issue to the VA. Veterans can be more susceptible to complications if DM is left uncontrolled prior to surgery. To determine readiness for surgery and control of comorbid conditions such as DM before a planned surgery, facilities often perform a preoperative clinic assessment, often in a multidisciplinary clinic.

At Veteran Health Indiana (VHI), a presurgery clinic visit involving the primary surgery service (physician, nurse practitioner, and/or a physician assistant) is conducted 1 to 2 months prior to the planned procedure to determine whether a patient is ready for surgery. During this visit, patients receive a packet with instructions for various tasks and medications, such as applying topical antibiotic prophylaxis on the anticipated surgical site. This is documented in the form of a note in the VHI Computerized Patient Record System (CPRS). The medication instructions are provided according to the preferences of the surgical team. These may be templated notes that contain general directions on the timing and dosing of specific medications, in addition to instructions for holding or reducing doses when appropriate. The instructions can be tailored by the team conducting the preoperative visit (eg, “Take 20 units of insulin glargine the day before surgery” vs “Take half of your long-acting insulin the night before surgery”). Specific to DM, VHI has a nurse-driven day of surgery glucose assessment where point-of-care blood glucose is collected during preoperative holding for most patients.

There is limited research assessing the level of preoperative glycemic control and the incidence of complications in a veteran population. The objective of this study was to gain a baseline understanding of what, if any, standardization exists for preoperative instructions for DM medications and to assess the level of preoperative glycemic control and postoperative complications in patients with DM undergoing major elective surgical procedures.

Methods

This retrospective, single-center chart review was conducted at VHI. The Indiana University and VHI institutional review boards determined that this quality improvement project was exempt from review.

The primary outcome was the number of patients with surgical procedures delayed or canceled due to hyperglycemia or hypoglycemia. Hyperglycemia was defined as blood glucose > 180 mg/dL and hypoglycemia was defined as < 70 mg/dL, slight variations from the current ADA SOC preoperative specific recommendation of a blood glucose reading of 100 to 180 mg/dL within 4 hours of surgery.¹⁰ The standard outpatient hypoglycemia definition of blood glucose < 70 mg/dL was chosen because the current goal (< 100 mg/dL) was not the standard in previous ADA SOCs that were in place during the study period. Specifically, the 2018 ADA SOC did not provide preoperative recommendations and the 2019-2021 ADA SOC recommended 80 to 180 mg/dL.^10,12-18 For patients who had multiple preoperative blood glucose measurements, the first recorded glucose on the day of the procedure was used.

The secondary outcomes of this study were focused on the preoperative process/care at VHI and postoperative glycemic control. The preoperative process included examining whether medication instructions were given and their quality. Additionally, the number of interventions for hyperglycemia and hypoglycemia were required immediately prior to surgery and the average preoperative HbA_1c (measured within 3 months prior to surgery) were collected and analyzed. For postoperative glycemic control, average blood glucose measurements and number of hypoglycemic (< 70 mg/dL) and hyperglycemic (> 180 mg/dL) events were measured in addition to the frequency of changes made at discharge to patients’ DM medication regimens.

The safety outcome of this study assessed commonly observed postoperative complications and was examined up to 30 days postsurgery. These included acute kidney injury (defined using Kidney Disease: Improving Global Outcomes 2012, the standard during the study period), nonfatal myocardial infarction, nonfatal stroke, and surgical site infections, which were identified from the discharge summary written by the primary surgery service.¹⁹ All-cause mortality also was collected.

Patients were included if they were admitted for major elective surgeries and had a diagnosis of either type 1 or type 2 DM on their problem list, determined by International Classification of Diseases, Tenth Revision codes. Major elective surgery was defined as a procedure that would likely result in a hospital admission of > 24 hours. Of note, patients may have been included in this study more than once if they had > 1 procedure at least 30 days apart and met inclusion criteria within the time frame. Patients were excluded if they were taking no DM medications or chronic steroids (at any dose), residing in a long-term care facility, being managed by a non-VA clinician prior to surgery, or missing a preoperative blood glucose measurement.

All data were collected from the CPRS. A list of surgical cases involving patients with DM who were scheduled to undergo major elective surgeries from January 1, 2018, to December 31, 2021, at VHI was generated. The list was randomized to a smaller number (N = 394) for data collection due to the time and resource constraints for a pharmacy residency project. All data were deidentified and stored in a secured VA server to protect patient confidentiality. Descriptive statistics were used for all results.

Results

Initially, 2362 surgeries were identified. A randomized sample of 394 charts were reviewed and 131 cases met inclusion criteria. Each case involved a unique patient (Figure). The most common reasons for exclusion were 143 patients with diet-controlled DM and 78 nonelective surgeries. The mean (SD) age of patients was 68 (8) years, and the most were male (98.5%) and White (76.3%) (Table 1). 

At baseline, 45 of 131 patients (34.4%) had coronary artery disease and 29 (22.1%) each had autonomic neuropathy and chronic kidney disease. Most surgeries were conducted by orthopedic (32.1%) and peripheral vascular (21.4%) specialties. The mean (SD) length of surgery was 4.6 (2.6) hours and of hospital length of stay was 4 (4) days. No patients stayed longer than the 30-day safety outcome follow-up period. All patients had type 2 DM and took a mean 2 DM medications. The 63 patients taking insulin had a mean (SD) total daily dose of 99 (77) U (Table 2). A preoperative HbA1c was collected in 116 patients within 3 months of surgery, with a mean HbA_1c of 7.0% (range, 5.3-10.7).

No patients had surgeries delayed or canceled because of uncontrolled DM on the day of surgery. The mean preoperative blood glucose level was 146 mg/dL (range, 73-365) (Table 3). No patients had a preoperative blood glucose level of < 70 mg/dL and 19 (14.5%) had a blood glucose level > 180 mg/dL. Among patients with hyperglycemia immediately prior to surgery, 6 (31.6%) had documentation of insulin being provided.

For this sample of patients, the preoperative clinic visit was conducted a mean 22 days prior to the planned surgery date. Among the 131 included patients, 122 (93.1%) had documentation of receiving instructions for DM medications. Among patients who had documented receipt of instructions, only 30 (24.6%) had instructions specifically tailored to their regimen rather than a generic templated form. The mean (SD) preoperative blood glucose was similar for those who received specific perioperative DM instructions at 146 (50) mg/dL when compared with those who did not at 147 (45) mg/dL. The mean (SD) preoperative blood glucose reading for those who had no documentation of receipt of perioperative instructions was 126 (54) mg/dL compared with 147 (46) mg/dL for those who did.

The mean number of postoperative blood glucose events per day was negligible for hypoglycemia and more frequent for hyperglycemia with a mean of 2 events per day. The mean postoperative blood glucose range was 121 to 247 mg/dL with most readings < 180 mg/dL. Upon discharge, most patients continued their home DM regimen with 5 patients (3.8%) having changes made to their regimen upon discharge.

Very few postoperative complications were identified from chart review. The most frequently observed postoperative complications were acute kidney injury, surgical site infections, and nonfatal stroke. There were no documented nonfatal myocardial infarctions. Two patients (1.5%) died within 30 days of the surgery; neither death was deemed to have been related to poor perioperative glycemic control.

Discussion

To our knowledge, this retrospective chart review was the first study to assess preoperative DM management and postoperative complications in a veteran population. VHI is a large, tertiary, level 1a, academic medical center that serves approximately 62,000 veterans annually and performs about 5000 to 6000 surgeries annually, a total that is increasing following the COVID-19 pandemic.²⁰ This study found that the current process of a presurgery clinic visit and day of surgery glucose assessment has prevented surgical delays or cancellations.

Most patients included in this study were well controlled at baseline in accordance with the 2025 ADA SOC HbA_1c recommendation of a preoperative HbA1c of < 8%, which may have contributed to no surgical delays or cancellations.¹⁰ However, not all patients had HbA1c collected within 3 months of surgery or even had one collected at all. Despite the ADA SOC providing no explicit recommendation for universal HbA_1c screening prior to elective procedures, its importance cannot be understated given the body of evidence demonstrating poor outcomes with uncontrolled preoperative DM.^8,10 The glycemic control at baseline may have contributed to the very few postsurgical complications observed in this study.

Although the current process at VHI prevented surgical delays and cancellations in this sample, there are still identified areas for improvement. One area is the instructions the patients received. Patients with DM are often prescribed ≥ 1 medication or a combination of insulins, noninsulin injectables, and oral DM medications, and this study population was no different. Because these medications may influence the anesthesia and perioperative periods, the ADA has specific guidance for altering administration schedules in the days leading up to surgery.¹⁰

Inappropriate administration of DM medications could lead to perioperative hypoglycemia or hyperglycemia, possibly causing surgical delays, case cancellations, and/or postoperative complications.21 Although these data reveal the specificity and documented receipt that the preoperative DM instructions did not impact the first recorded preoperative blood glucose, future studies should examine patient confidence in how to properly administer their DM medications prior to surgery. It is vital that patients receive clear instructions in accordance with the ADA SOC on whether to continue, hold, or adjust the dose of their medications to prevent fluctuations in blood glucose levels in the perioperative period, ensure safety with anesthesia, and prevent postoperative complications such as acute kidney injury. Of note, compliance with guideline recommendations for medication instructions was not examined because the data collection time frame expanded over multiple years and the recommendations have evolved each year as new data emerge.

Preoperative DM Management

The first key takeaway from this study is to ensure patients are ready for surgery with a formal assessment (typically in the form of a clinic visit) prior to the surgery. One private sector health system published their approach to this by administering an automatic preoperative HbA_1c screening for those with a DM diagnosis and all patients with a random plasma glucose ≥ 200 mg/dL.²² Additionally, if the patient's HbA_1c level was not at goal prior to surgery (≥ 8% for those with known DM and ≥ 6.5% with no known DM), patients were referred to endocrinology for further management. Increasing attention to the preoperative visit and extending HbA_1c testing to all patients regardless of DM status also provides an opportunity to identify individuals living with undiagnosed DM.¹

Even though there was no difference in the mean preoperative blood glucose level based on receipt or specificity of preoperative DM instructions, a second takeaway from this study is the importance of ensuring patients receive clear instructions on their DM medication schedule in the perioperative period. A practical first step may be updating the templates used by the primary surgery teams and providing education to the clinicians in the clinic on how to personalize the visits. Because the current preoperative DM process at VHI is managed by the primary surgical team in a clinic visit, there is an opportunity to shift this responsibility to other health care professionals, such as pharmacists—a change shown to reduce unintended omission of home medications following surgery during hospitalization and reduce costs.^23,24

Limitations

This study relied on data included in the patient chart. These data include medication interventions made immediately prior to surgery, which can sometimes be inaccurately charted or difficult to find as they are not documented in the typical medication administration record. Also, the safety outcomes were collected from a discharge summary written by different clinicians, which may lead to information bias. Special attention was taken to ensure these data points were collected as accurately as possible, but it is possible some data may be inaccurate from unintentional human error. Additionally, the safety outcome was limited to a 30-day follow-up, but encompassed the entire length of postoperative stay for all included patients. Finally, given this study was retrospective with no comparison group and the intent was to improve processes at VHI, only hypotheses and potential interventions can be generated from this study. Future prospective studies with larger sample sizes and comparator groups are needed to draw further conclusions.

Conclusions

This study found that the current presurgery process at VHI appears to be successful in preventing surgical delays or cancellations due to hyperglycemia or hypoglycemia. Optimizing DM management can improve surgical outcomes by decreasing rates of postoperative complications, and this study added additional evidence in support of that in a unique population: veterans. Insight on the awareness of preoperative blood glucose management should be gleaned from this study, and based on this sample and site, the preadmission screening process and instructions provided to patients can serve as 2 starting points for optimizing elective surgery.

Type 2 diabetes mellitus (T2DM) is a chronic disease becoming more prevalent each year and is the seventh-leading cause of death in the United States.¹ The most common reason for hospitalization for patients with T2DM is uncontrolled glycemic levels.² Nearly 25% of the US Department of Veterans Affairs (VA) patient population has T2DM.³ T2DM is the leading cause of blindness, end-stage renal disease, and amputation for VA patients.⁴

According to the 2023 American Diabetes Association (ADA) guidelines, treatment goals of T2DM include eliminating symptoms, preventing or delaying complications, and attaining glycemic goals. A typical hemoglobin A_1c (HbA_1c) goal range is < 7%, but individual goals can vary up to < 9% due to a multitude of factors, including patient comorbidities and clinical status.⁵

Initial treatment recommendations are nonpharmacologic and include comprehensive lifestyle interventions such as optimizing nutrition, physical activity, and behavioral therapy. When pharmacologic therapy is required, metformin is the preferred first-line treatment for the majority of newly diagnosed patients with T2DM and should be added to continued lifestyle management.⁵ If HbA_1c levels remains above goal, the 2023 ADA guidelines recommend adding a second medication, including but not limited to insulin, a glucagonlike peptide-1 receptor agonist (GLP-1RA), or a sodium-glucose cotransporter 2 inhibitor. Medication choice is largely based on the patient’s concomitant conditions (eg, atherosclerotic cardiovascular disease, heart failure, or chronic kidney disease). The 2023 ADA guidelines suggest initiating insulin therapy when a patient's blood glucose ≥ 300 mg/dL, HbA_1c > 10%, or if the patient has symptoms of hyperglycemia, even at initial diagnosis. Initiating medications to minimize or avoid hypoglycemia is a priority, especially in high-risk individuals.⁵

Clinical evidence shows that GLP-1RAs may provide similar glycemic control to insulin with lower risk of hypoglycemia.⁶ Other reported benefits of GLP-1RAs include weight loss, blood pressure reduction, and improved lipid levels. The most common adverse events (AEs) with GLP-1RAs are gastrointestinal. Including GLP-1RAs in T2DM pharmacotherapy may lower the risk of hypoglycemia, especially in patients at high risk of hypoglycemia.

The 2023 ADA guidelines indicate that it is appropriate to initiate GLP-]1RAs in patients on insulin.⁵ However, while GLP-1RAs do not increase the risk of hypoglycemia independently, combination treatment with GLP-1RAs and insulin can still result in hypoglycemia.⁶ Insulin is the key suspect of this hypoglycemic risk.⁷ Thus, if insulin dosage can be reduced or discontinued, this might reduce the risk of hypoglycemia.

The literature is limited on how the addition of a GLP-1RA to insulin treatment will affect the patient's daily insulin doses, particularly for the veteran population. The goal of this study is to examine this gap in current research by examining semaglutide, which is the current formulary preferred GLP-1RA at the VA.

Semaglutide is subcutaneously initiated at a dose of 0.25 mg once weekly for 4 weeks to reduce gastrointestinal symptoms, then increased to 0.5 mg weekly. Additional increases to a maintenance dose of 1 mg or 2 mg weekly can occur to achieve glycemic goals. The SUSTAIN-FORTE randomized controlled trial sought to determine whether there was a difference in HbA_1c level reduction and significant weight loss with the 2-mg vs 1-mg dose.⁸ Patients in the trial were taking metformin but needed additional medication to control their HbA_1c. They were not using insulin and may or may not have been taking sulfonylureas prior to semaglutide initiation. Semaglutide 2 mg was found to significantly improve HbA_1c control and promote weight loss compared with semaglutide 1 mg, while maintaining a similar safety profile.

Because this study involved patients who required additional HbA_1c control, although semaglutide reduced HbA_1c, not all patients were able to reduce their other diabetes medications, which depended on the baseline HbA_1c level and the level upon completion of semaglutide titration. Dose reductions for the patients’ other T2DM medications were not reported at trial end. SUSTAIN-FORTE established titration up to semaglutide 2 mg as effective for HbA_1c reduction, although it did not study patients also on insulin.⁸

Insulin is associated with hypoglycemic risk, weight gain, and other AEs.^7,8 This study analyzed whether increasing semaglutide could reduce insulin doses and therefore reduce risk of AEs in patients with T2DM.

Methods

A retrospective, single-center, chart review was conducted at VA Sioux Falls Health Care System (VASFHCS). Data were collected through manual review of VASFHCS electronic medical records. Patients aged ≥ 18 years with active prescriptions for at least once-daily insulin who were initiated on 2-mg weekly dose of semaglutide at the VASFHCS clinical pharmacy practitioner medication management clinic between January 1, 2021, and September 1, 2023, were included. VASFHCS clinical pharmacy practitioners have a scope of practice that allows them to initiate, modify, or discontinue medication therapy within medication management clinics.

The most frequently used prandial insulin at VASFHCS is insulin aspart, and the most frequently used basal insulin is insulin glargine. Patients were retrospectively monitored as they progressed from baseline (the point in time where semaglutide 0.5 mg was initiated) to ≥ 3 months on semaglutide 2-mg therapy. Patients were excluded if they previously used a GLP-1RA or if they were on sliding scale insulin without an exact daily dosage.

The primary endpoint was the percent change in total daily insulin dose from baseline to each dose increase after receiving semaglutide 2 mg for ≥ 3 months. Secondary endpoints included changes in daily prandial insulin dose, daily basal insulin dose, HbA_1c, and number of hypoglycemic events reported. Data collected included age, race, weight, body mass index, total daily prandial insulin dose, total daily basal insulin dose, HbA_1c, and hypoglycemic events reported at the visit when semaglutide was initiated.

Statistical Analysis

The sample size was calculated prior to data collection, and it was determined that for α = .05, 47 patients were needed to achieve 95% power. The primary endpoint was assessed using a paired t test, as were each secondary endpoint. Results with P < .05 were considered statistically significant.

Results

Sixty-two patients were included. The mean HbA_1c level at baseline was 7.7%, the baseline mean prandial and insulin daily doses were 41.5 units and 85.1 units, respectively (Table 1) From baseline to initiation of a semaglutide 1-mg dose, the daily insulin dose changed –5.6% (95% CI, 2.2-14.0; P = .008). From baseline to 2-mg dose initiation daily insulin changed -22.2% (95% CI, 22.0-35.1; P < .001) and for patients receiving semaglutide 2 mg for ≥ 3 months it changed -36.9% (95% CI, 37.4-56.5; P < .001) (Figure).

After receiving the 2-mg dose for ≥ 3 months, the mean daily dose of prandial insulin decreased from 41.5 units to 24.6 units (95% CI, 12.6-21.2; P < .001); mean daily dose of basal insulin decreased from 85.1 units to 52.1 units (95% CI, 23.9-42.0; P < .001); and mean HbA_1c level decreased from 7.7% to 7.1% (95% CI, 0.3-0.8; P < .001). Mean number of hypoglycemic events reported was not statistically significant, changing from 3.6 to 3.2 (95% CI, –0.6 to 0.1; P = .21) (Table 2).

Discussion

This study investigated the effect of subcutaneous semaglutide dose escalation on total daily insulin dose for patients with T2DM. There was a statistically significant decrease in total daily insulin dose from baseline to 1 mg initiation; this decrease continued with further insulin dose reduction seen at the 2-mg dose initiation and additional insulin dose reduction at ≥ 3 months at this dose. It was hypothesized there would be a significant total daily insulin dose reduction at some point, especially when transitioning from the semaglutide 1-mg to the 2-mg dose, based on previous research. ^9,10 The additional reduction in daily insulin dose when continuing on semaglutide 2 mg for ≥ 3 months was an unanticipated but added benefit, showing that if tolerated, maintaining the 2-mg dose will help patients reduce their insulin doses.

In terms of secondary endpoints, there was a statistically significant decrease in mean total daily dose individually for prandial and basal insulin from baseline to ≥ 3 months after semaglutide 2 mg initiation. The change in HbA_1c level was also statistically significant and decreased from baseline, even as insulin doses were reduced. This change in HbA_1c level was expected; previous literature has shown a significant link between improving HbA_1c control when semaglutide doses are increased to 2 mg weekly.¹⁰ Due to having been shown in previous trials, it was expected that HbA_1c levels would decrease even when the insulin doses were being reduced.¹⁰ Insulin dose reduction can potentially be added to the growing evidence of semaglutide benefits. The change in the number of hypoglycemic events was not statistically significant, which was unexpected since previous research show a trend in patients taking GLP-1RAs having fewer hypoglycemic events than those taking insulin.⁶ Further investigation with a larger sample size and prospective trial could determine whether this result is an outlier. In this study, there was no increase in HbA_1c or hypoglycemic events reported with increasing semaglutide doses, which provides further evidence of the safety of semaglutide even at higher doses.

These data suggest that for a patient with T2DM who is already taking insulin, the recommended titration of semaglutide is to start with 0.5 mg and titrate up to a 2-mg subcutaneous weekly dose and to then continue at that dose. As long as the 2-mg dose is tolerated, it will provide patients with the most HbA_1c control and lead to a reduction of their total daily insulin doses according to these results.

Strengths and Limitations

This study compared patient data at different points. This method did not require a second distinct control group, which would potentially introduce confounding factors, such as different baseline characteristics. Another strength is that documentation was available for all patients throughout the study so no one was lost to follow-up. This allowed comprehensive data collection and provided a stronger conclusion given the completeness of the data from baseline to follow-up.

Limitations include the retrospective design and small sample size. In addition, the study design did not allow for randomization. There is no documentation of adherence to medication regimen, which was difficult to determine due to the retrospective nature. Other changes to the patients’ medication regimen were not collected in aggregate and thus, it is possible the total daily insulin dose was impacted by other medication changes. There is also potential for inconsistent documentation of the patients’ true total daily insulin dose in the medical record, thus leading to inaccuracy of recorded data.

Conclusions

A small sample of veterans with T2DM had statistically significant reductions in total daily insulin dose when subcutaneous semaglutide was initiated, as well as after each dose increase. There was also a statistically significant reduction in HbA_1c levels from baseline even as patient insulin doses were reduced. These results support the current practice of using semaglutide to treat T2DM, suggesting it may be safe and effective at reducing HbA_1c levels as the dose is titrated up to 2 mg. There was no statistically significant change in the number of hypoglycemic events reported as semaglutide was titrated up. Thus, when semaglutide is increased to the maximum recommended dose of 2 mg for T2DM, patients may experience a reduction of their total daily dose of insulin and HbA_1c levels. These benefits may reduce the risk of insulin-related AEs while maintaining appropriate glycemic control.

Type 2 diabetes mellitus (T2DM) is a chronic disease becoming more prevalent each year and is the seventh-leading cause of death in the United States.¹ The most common reason for hospitalization for patients with T2DM is uncontrolled glycemic levels.² Nearly 25% of the US Department of Veterans Affairs (VA) patient population has T2DM.³ T2DM is the leading cause of blindness, end-stage renal disease, and amputation for VA patients.⁴

According to the 2023 American Diabetes Association (ADA) guidelines, treatment goals of T2DM include eliminating symptoms, preventing or delaying complications, and attaining glycemic goals. A typical hemoglobin A_1c (HbA_1c) goal range is < 7%, but individual goals can vary up to < 9% due to a multitude of factors, including patient comorbidities and clinical status.⁵

Initial treatment recommendations are nonpharmacologic and include comprehensive lifestyle interventions such as optimizing nutrition, physical activity, and behavioral therapy. When pharmacologic therapy is required, metformin is the preferred first-line treatment for the majority of newly diagnosed patients with T2DM and should be added to continued lifestyle management.⁵ If HbA_1c levels remains above goal, the 2023 ADA guidelines recommend adding a second medication, including but not limited to insulin, a glucagonlike peptide-1 receptor agonist (GLP-1RA), or a sodium-glucose cotransporter 2 inhibitor. Medication choice is largely based on the patient’s concomitant conditions (eg, atherosclerotic cardiovascular disease, heart failure, or chronic kidney disease). The 2023 ADA guidelines suggest initiating insulin therapy when a patient's blood glucose ≥ 300 mg/dL, HbA_1c > 10%, or if the patient has symptoms of hyperglycemia, even at initial diagnosis. Initiating medications to minimize or avoid hypoglycemia is a priority, especially in high-risk individuals.⁵

Clinical evidence shows that GLP-1RAs may provide similar glycemic control to insulin with lower risk of hypoglycemia.⁶ Other reported benefits of GLP-1RAs include weight loss, blood pressure reduction, and improved lipid levels. The most common adverse events (AEs) with GLP-1RAs are gastrointestinal. Including GLP-1RAs in T2DM pharmacotherapy may lower the risk of hypoglycemia, especially in patients at high risk of hypoglycemia.

The 2023 ADA guidelines indicate that it is appropriate to initiate GLP-]1RAs in patients on insulin.⁵ However, while GLP-1RAs do not increase the risk of hypoglycemia independently, combination treatment with GLP-1RAs and insulin can still result in hypoglycemia.⁶ Insulin is the key suspect of this hypoglycemic risk.⁷ Thus, if insulin dosage can be reduced or discontinued, this might reduce the risk of hypoglycemia.

The literature is limited on how the addition of a GLP-1RA to insulin treatment will affect the patient's daily insulin doses, particularly for the veteran population. The goal of this study is to examine this gap in current research by examining semaglutide, which is the current formulary preferred GLP-1RA at the VA.

Semaglutide is subcutaneously initiated at a dose of 0.25 mg once weekly for 4 weeks to reduce gastrointestinal symptoms, then increased to 0.5 mg weekly. Additional increases to a maintenance dose of 1 mg or 2 mg weekly can occur to achieve glycemic goals. The SUSTAIN-FORTE randomized controlled trial sought to determine whether there was a difference in HbA_1c level reduction and significant weight loss with the 2-mg vs 1-mg dose.⁸ Patients in the trial were taking metformin but needed additional medication to control their HbA_1c. They were not using insulin and may or may not have been taking sulfonylureas prior to semaglutide initiation. Semaglutide 2 mg was found to significantly improve HbA_1c control and promote weight loss compared with semaglutide 1 mg, while maintaining a similar safety profile.

Because this study involved patients who required additional HbA_1c control, although semaglutide reduced HbA_1c, not all patients were able to reduce their other diabetes medications, which depended on the baseline HbA_1c level and the level upon completion of semaglutide titration. Dose reductions for the patients’ other T2DM medications were not reported at trial end. SUSTAIN-FORTE established titration up to semaglutide 2 mg as effective for HbA_1c reduction, although it did not study patients also on insulin.⁸

Insulin is associated with hypoglycemic risk, weight gain, and other AEs.^7,8 This study analyzed whether increasing semaglutide could reduce insulin doses and therefore reduce risk of AEs in patients with T2DM.

Methods

A retrospective, single-center, chart review was conducted at VA Sioux Falls Health Care System (VASFHCS). Data were collected through manual review of VASFHCS electronic medical records. Patients aged ≥ 18 years with active prescriptions for at least once-daily insulin who were initiated on 2-mg weekly dose of semaglutide at the VASFHCS clinical pharmacy practitioner medication management clinic between January 1, 2021, and September 1, 2023, were included. VASFHCS clinical pharmacy practitioners have a scope of practice that allows them to initiate, modify, or discontinue medication therapy within medication management clinics.

The most frequently used prandial insulin at VASFHCS is insulin aspart, and the most frequently used basal insulin is insulin glargine. Patients were retrospectively monitored as they progressed from baseline (the point in time where semaglutide 0.5 mg was initiated) to ≥ 3 months on semaglutide 2-mg therapy. Patients were excluded if they previously used a GLP-1RA or if they were on sliding scale insulin without an exact daily dosage.

The primary endpoint was the percent change in total daily insulin dose from baseline to each dose increase after receiving semaglutide 2 mg for ≥ 3 months. Secondary endpoints included changes in daily prandial insulin dose, daily basal insulin dose, HbA_1c, and number of hypoglycemic events reported. Data collected included age, race, weight, body mass index, total daily prandial insulin dose, total daily basal insulin dose, HbA_1c, and hypoglycemic events reported at the visit when semaglutide was initiated.

Statistical Analysis

The sample size was calculated prior to data collection, and it was determined that for α = .05, 47 patients were needed to achieve 95% power. The primary endpoint was assessed using a paired t test, as were each secondary endpoint. Results with P < .05 were considered statistically significant.

Results

Sixty-two patients were included. The mean HbA_1c level at baseline was 7.7%, the baseline mean prandial and insulin daily doses were 41.5 units and 85.1 units, respectively (Table 1) From baseline to initiation of a semaglutide 1-mg dose, the daily insulin dose changed –5.6% (95% CI, 2.2-14.0; P = .008). From baseline to 2-mg dose initiation daily insulin changed -22.2% (95% CI, 22.0-35.1; P < .001) and for patients receiving semaglutide 2 mg for ≥ 3 months it changed -36.9% (95% CI, 37.4-56.5; P < .001) (Figure).

After receiving the 2-mg dose for ≥ 3 months, the mean daily dose of prandial insulin decreased from 41.5 units to 24.6 units (95% CI, 12.6-21.2; P < .001); mean daily dose of basal insulin decreased from 85.1 units to 52.1 units (95% CI, 23.9-42.0; P < .001); and mean HbA_1c level decreased from 7.7% to 7.1% (95% CI, 0.3-0.8; P < .001). Mean number of hypoglycemic events reported was not statistically significant, changing from 3.6 to 3.2 (95% CI, –0.6 to 0.1; P = .21) (Table 2).

Discussion

This study investigated the effect of subcutaneous semaglutide dose escalation on total daily insulin dose for patients with T2DM. There was a statistically significant decrease in total daily insulin dose from baseline to 1 mg initiation; this decrease continued with further insulin dose reduction seen at the 2-mg dose initiation and additional insulin dose reduction at ≥ 3 months at this dose. It was hypothesized there would be a significant total daily insulin dose reduction at some point, especially when transitioning from the semaglutide 1-mg to the 2-mg dose, based on previous research. ^9,10 The additional reduction in daily insulin dose when continuing on semaglutide 2 mg for ≥ 3 months was an unanticipated but added benefit, showing that if tolerated, maintaining the 2-mg dose will help patients reduce their insulin doses.

In terms of secondary endpoints, there was a statistically significant decrease in mean total daily dose individually for prandial and basal insulin from baseline to ≥ 3 months after semaglutide 2 mg initiation. The change in HbA_1c level was also statistically significant and decreased from baseline, even as insulin doses were reduced. This change in HbA_1c level was expected; previous literature has shown a significant link between improving HbA_1c control when semaglutide doses are increased to 2 mg weekly.¹⁰ Due to having been shown in previous trials, it was expected that HbA_1c levels would decrease even when the insulin doses were being reduced.¹⁰ Insulin dose reduction can potentially be added to the growing evidence of semaglutide benefits. The change in the number of hypoglycemic events was not statistically significant, which was unexpected since previous research show a trend in patients taking GLP-1RAs having fewer hypoglycemic events than those taking insulin.⁶ Further investigation with a larger sample size and prospective trial could determine whether this result is an outlier. In this study, there was no increase in HbA_1c or hypoglycemic events reported with increasing semaglutide doses, which provides further evidence of the safety of semaglutide even at higher doses.

These data suggest that for a patient with T2DM who is already taking insulin, the recommended titration of semaglutide is to start with 0.5 mg and titrate up to a 2-mg subcutaneous weekly dose and to then continue at that dose. As long as the 2-mg dose is tolerated, it will provide patients with the most HbA_1c control and lead to a reduction of their total daily insulin doses according to these results.

Strengths and Limitations

This study compared patient data at different points. This method did not require a second distinct control group, which would potentially introduce confounding factors, such as different baseline characteristics. Another strength is that documentation was available for all patients throughout the study so no one was lost to follow-up. This allowed comprehensive data collection and provided a stronger conclusion given the completeness of the data from baseline to follow-up.

Limitations include the retrospective design and small sample size. In addition, the study design did not allow for randomization. There is no documentation of adherence to medication regimen, which was difficult to determine due to the retrospective nature. Other changes to the patients’ medication regimen were not collected in aggregate and thus, it is possible the total daily insulin dose was impacted by other medication changes. There is also potential for inconsistent documentation of the patients’ true total daily insulin dose in the medical record, thus leading to inaccuracy of recorded data.

Conclusions

A small sample of veterans with T2DM had statistically significant reductions in total daily insulin dose when subcutaneous semaglutide was initiated, as well as after each dose increase. There was also a statistically significant reduction in HbA_1c levels from baseline even as patient insulin doses were reduced. These results support the current practice of using semaglutide to treat T2DM, suggesting it may be safe and effective at reducing HbA_1c levels as the dose is titrated up to 2 mg. There was no statistically significant change in the number of hypoglycemic events reported as semaglutide was titrated up. Thus, when semaglutide is increased to the maximum recommended dose of 2 mg for T2DM, patients may experience a reduction of their total daily dose of insulin and HbA_1c levels. These benefits may reduce the risk of insulin-related AEs while maintaining appropriate glycemic control.

Type 2 diabetes mellitus (T2DM) is a chronic disease becoming more prevalent each year and is the seventh-leading cause of death in the United States.¹ The most common reason for hospitalization for patients with T2DM is uncontrolled glycemic levels.² Nearly 25% of the US Department of Veterans Affairs (VA) patient population has T2DM.³ T2DM is the leading cause of blindness, end-stage renal disease, and amputation for VA patients.⁴

According to the 2023 American Diabetes Association (ADA) guidelines, treatment goals of T2DM include eliminating symptoms, preventing or delaying complications, and attaining glycemic goals. A typical hemoglobin A_1c (HbA_1c) goal range is < 7%, but individual goals can vary up to < 9% due to a multitude of factors, including patient comorbidities and clinical status.⁵

Initial treatment recommendations are nonpharmacologic and include comprehensive lifestyle interventions such as optimizing nutrition, physical activity, and behavioral therapy. When pharmacologic therapy is required, metformin is the preferred first-line treatment for the majority of newly diagnosed patients with T2DM and should be added to continued lifestyle management.⁵ If HbA_1c levels remains above goal, the 2023 ADA guidelines recommend adding a second medication, including but not limited to insulin, a glucagonlike peptide-1 receptor agonist (GLP-1RA), or a sodium-glucose cotransporter 2 inhibitor. Medication choice is largely based on the patient’s concomitant conditions (eg, atherosclerotic cardiovascular disease, heart failure, or chronic kidney disease). The 2023 ADA guidelines suggest initiating insulin therapy when a patient's blood glucose ≥ 300 mg/dL, HbA_1c > 10%, or if the patient has symptoms of hyperglycemia, even at initial diagnosis. Initiating medications to minimize or avoid hypoglycemia is a priority, especially in high-risk individuals.⁵

Clinical evidence shows that GLP-1RAs may provide similar glycemic control to insulin with lower risk of hypoglycemia.⁶ Other reported benefits of GLP-1RAs include weight loss, blood pressure reduction, and improved lipid levels. The most common adverse events (AEs) with GLP-1RAs are gastrointestinal. Including GLP-1RAs in T2DM pharmacotherapy may lower the risk of hypoglycemia, especially in patients at high risk of hypoglycemia.

The 2023 ADA guidelines indicate that it is appropriate to initiate GLP-]1RAs in patients on insulin.⁵ However, while GLP-1RAs do not increase the risk of hypoglycemia independently, combination treatment with GLP-1RAs and insulin can still result in hypoglycemia.⁶ Insulin is the key suspect of this hypoglycemic risk.⁷ Thus, if insulin dosage can be reduced or discontinued, this might reduce the risk of hypoglycemia.

The literature is limited on how the addition of a GLP-1RA to insulin treatment will affect the patient's daily insulin doses, particularly for the veteran population. The goal of this study is to examine this gap in current research by examining semaglutide, which is the current formulary preferred GLP-1RA at the VA.

Semaglutide is subcutaneously initiated at a dose of 0.25 mg once weekly for 4 weeks to reduce gastrointestinal symptoms, then increased to 0.5 mg weekly. Additional increases to a maintenance dose of 1 mg or 2 mg weekly can occur to achieve glycemic goals. The SUSTAIN-FORTE randomized controlled trial sought to determine whether there was a difference in HbA_1c level reduction and significant weight loss with the 2-mg vs 1-mg dose.⁸ Patients in the trial were taking metformin but needed additional medication to control their HbA_1c. They were not using insulin and may or may not have been taking sulfonylureas prior to semaglutide initiation. Semaglutide 2 mg was found to significantly improve HbA_1c control and promote weight loss compared with semaglutide 1 mg, while maintaining a similar safety profile.

Because this study involved patients who required additional HbA_1c control, although semaglutide reduced HbA_1c, not all patients were able to reduce their other diabetes medications, which depended on the baseline HbA_1c level and the level upon completion of semaglutide titration. Dose reductions for the patients’ other T2DM medications were not reported at trial end. SUSTAIN-FORTE established titration up to semaglutide 2 mg as effective for HbA_1c reduction, although it did not study patients also on insulin.⁸

Insulin is associated with hypoglycemic risk, weight gain, and other AEs.^7,8 This study analyzed whether increasing semaglutide could reduce insulin doses and therefore reduce risk of AEs in patients with T2DM.

Methods

A retrospective, single-center, chart review was conducted at VA Sioux Falls Health Care System (VASFHCS). Data were collected through manual review of VASFHCS electronic medical records. Patients aged ≥ 18 years with active prescriptions for at least once-daily insulin who were initiated on 2-mg weekly dose of semaglutide at the VASFHCS clinical pharmacy practitioner medication management clinic between January 1, 2021, and September 1, 2023, were included. VASFHCS clinical pharmacy practitioners have a scope of practice that allows them to initiate, modify, or discontinue medication therapy within medication management clinics.

The most frequently used prandial insulin at VASFHCS is insulin aspart, and the most frequently used basal insulin is insulin glargine. Patients were retrospectively monitored as they progressed from baseline (the point in time where semaglutide 0.5 mg was initiated) to ≥ 3 months on semaglutide 2-mg therapy. Patients were excluded if they previously used a GLP-1RA or if they were on sliding scale insulin without an exact daily dosage.

The primary endpoint was the percent change in total daily insulin dose from baseline to each dose increase after receiving semaglutide 2 mg for ≥ 3 months. Secondary endpoints included changes in daily prandial insulin dose, daily basal insulin dose, HbA_1c, and number of hypoglycemic events reported. Data collected included age, race, weight, body mass index, total daily prandial insulin dose, total daily basal insulin dose, HbA_1c, and hypoglycemic events reported at the visit when semaglutide was initiated.

Statistical Analysis

The sample size was calculated prior to data collection, and it was determined that for α = .05, 47 patients were needed to achieve 95% power. The primary endpoint was assessed using a paired t test, as were each secondary endpoint. Results with P < .05 were considered statistically significant.

Results

Sixty-two patients were included. The mean HbA_1c level at baseline was 7.7%, the baseline mean prandial and insulin daily doses were 41.5 units and 85.1 units, respectively (Table 1) From baseline to initiation of a semaglutide 1-mg dose, the daily insulin dose changed –5.6% (95% CI, 2.2-14.0; P = .008). From baseline to 2-mg dose initiation daily insulin changed -22.2% (95% CI, 22.0-35.1; P < .001) and for patients receiving semaglutide 2 mg for ≥ 3 months it changed -36.9% (95% CI, 37.4-56.5; P < .001) (Figure).

After receiving the 2-mg dose for ≥ 3 months, the mean daily dose of prandial insulin decreased from 41.5 units to 24.6 units (95% CI, 12.6-21.2; P < .001); mean daily dose of basal insulin decreased from 85.1 units to 52.1 units (95% CI, 23.9-42.0; P < .001); and mean HbA_1c level decreased from 7.7% to 7.1% (95% CI, 0.3-0.8; P < .001). Mean number of hypoglycemic events reported was not statistically significant, changing from 3.6 to 3.2 (95% CI, –0.6 to 0.1; P = .21) (Table 2).

Discussion

This study investigated the effect of subcutaneous semaglutide dose escalation on total daily insulin dose for patients with T2DM. There was a statistically significant decrease in total daily insulin dose from baseline to 1 mg initiation; this decrease continued with further insulin dose reduction seen at the 2-mg dose initiation and additional insulin dose reduction at ≥ 3 months at this dose. It was hypothesized there would be a significant total daily insulin dose reduction at some point, especially when transitioning from the semaglutide 1-mg to the 2-mg dose, based on previous research. ^9,10 The additional reduction in daily insulin dose when continuing on semaglutide 2 mg for ≥ 3 months was an unanticipated but added benefit, showing that if tolerated, maintaining the 2-mg dose will help patients reduce their insulin doses.

In terms of secondary endpoints, there was a statistically significant decrease in mean total daily dose individually for prandial and basal insulin from baseline to ≥ 3 months after semaglutide 2 mg initiation. The change in HbA_1c level was also statistically significant and decreased from baseline, even as insulin doses were reduced. This change in HbA_1c level was expected; previous literature has shown a significant link between improving HbA_1c control when semaglutide doses are increased to 2 mg weekly.¹⁰ Due to having been shown in previous trials, it was expected that HbA_1c levels would decrease even when the insulin doses were being reduced.¹⁰ Insulin dose reduction can potentially be added to the growing evidence of semaglutide benefits. The change in the number of hypoglycemic events was not statistically significant, which was unexpected since previous research show a trend in patients taking GLP-1RAs having fewer hypoglycemic events than those taking insulin.⁶ Further investigation with a larger sample size and prospective trial could determine whether this result is an outlier. In this study, there was no increase in HbA_1c or hypoglycemic events reported with increasing semaglutide doses, which provides further evidence of the safety of semaglutide even at higher doses.

These data suggest that for a patient with T2DM who is already taking insulin, the recommended titration of semaglutide is to start with 0.5 mg and titrate up to a 2-mg subcutaneous weekly dose and to then continue at that dose. As long as the 2-mg dose is tolerated, it will provide patients with the most HbA_1c control and lead to a reduction of their total daily insulin doses according to these results.

Strengths and Limitations

This study compared patient data at different points. This method did not require a second distinct control group, which would potentially introduce confounding factors, such as different baseline characteristics. Another strength is that documentation was available for all patients throughout the study so no one was lost to follow-up. This allowed comprehensive data collection and provided a stronger conclusion given the completeness of the data from baseline to follow-up.

Limitations include the retrospective design and small sample size. In addition, the study design did not allow for randomization. There is no documentation of adherence to medication regimen, which was difficult to determine due to the retrospective nature. Other changes to the patients’ medication regimen were not collected in aggregate and thus, it is possible the total daily insulin dose was impacted by other medication changes. There is also potential for inconsistent documentation of the patients’ true total daily insulin dose in the medical record, thus leading to inaccuracy of recorded data.

Conclusions

A small sample of veterans with T2DM had statistically significant reductions in total daily insulin dose when subcutaneous semaglutide was initiated, as well as after each dose increase. There was also a statistically significant reduction in HbA_1c levels from baseline even as patient insulin doses were reduced. These results support the current practice of using semaglutide to treat T2DM, suggesting it may be safe and effective at reducing HbA_1c levels as the dose is titrated up to 2 mg. There was no statistically significant change in the number of hypoglycemic events reported as semaglutide was titrated up. Thus, when semaglutide is increased to the maximum recommended dose of 2 mg for T2DM, patients may experience a reduction of their total daily dose of insulin and HbA_1c levels. These benefits may reduce the risk of insulin-related AEs while maintaining appropriate glycemic control.

Diabetes mellitus (DM) is a national health crisis affecting > 38 million people (11.6%) in the United States.¹ American Indian and Alaska Native (AI/AN) adults are disproportionately affected, with a prevalence of 14.5%—the highest among all racial and ethnic groups.¹ Type 2 DM (T2DM) accounts for 90% to 95% of all DM cases and is a leading cause of morbidity and mortality due to its association with cardiovascular disease, kidney failure, and other complications.²

Maintaining glycemic control is important for managing T2DM and preventing microvascular and macrovascular complications.³ The cornerstone of diabetes self-management has been patient self-monitored blood glucose (SMBG) using finger-stick glucometers.⁴ However, SMBG provides measurements from a single point in time and requires frequent, painful, and inconvenient finger pricks, leading to decreased adherence.^5,6 These limitations negatively affect patient engagement and overall glycemic control.⁷

Continuous glucose monitors (CGMs) offer real-time, continuous glucose readings and trends.⁸ CGMs improve glycemic control and reduce hypoglycemic episodes in patients who are insulin-dependent.^9,10 Flash glucose monitors, a type of CGM that requires scanning to obtain glucose readings, provide similar benefits.¹¹ Despite these demonstrated advantages, research has primarily focused on insulin-dependent populations, leaving a significant gap in understanding the effect of CGMs on patients with T2DM who are not insulin-dependent.¹²

Given the high prevalence of T2DM among AI/AN populations and the potential benefits of CGMs, this study sought to evaluate the effect of CGM use on glycemic control and other health metrics in patients with non–insulin-dependent T2DM in an AI/AN population. This focus addresses a critical knowledge gap and may inform clinical practices and policies to improve diabetes management in this high-risk group.

Methods

A retrospective observational study was conducted using deidentified electronic health records (EHRs) from 2019 to 2024 at a federally operated outpatient Indian Health Service (IHS) clinic serving an AI/AN population in the IHS Portland Area (Oregon, Washington, Idaho). The study protocol was reviewed and deemed exempt by institutional review boards at Washington State University and the Portland Area IHS.

Study Population

This study included patients diagnosed with non–insulin-dependent T2DM, had used a CGM for ≥ 1 year, and had hemoglobin A_1c (HbA_1c) measurements within 4 months prior to CGM initiation (baseline) and within ± 4 months after 1 year of CGM use. For other health metrics, including blood pressure (BP), weight, low-density lipoprotein cholesterol (LDL-C), and estimated glomerular filtration rate (eGFR), this study required measurements within 6 months before CGM initiation and within 6 months after 1 year of CGM use. The baseline HbA_1c in the dataset ranged from 5.3% to > 14%.

Patients were excluded if they used insulin during the study period, had incomplete laboratory or clinical data for the required time frame, or had < 1 year of CGM use. The dataset did not include detailed information on oral DM medications; thus, we could not report or account for the type or number of oral hypoglycemic agents used by the patients. The IHS clinical applications coordinator compiled the dataset from the EHR, identifying patients who were prescribed and received a CGM at the clinic. All patients used the Abbott Freestyle Libre CGM, the only formulary CGM available at the clinic during the study period.

A 1-year follow-up endpoint was selected for several reasons: (1) to capture potential seasonal variations in diet and activity; (2) to align with the clinic’s standard practice of annual comprehensive diabetes evaluations; and (3) to allow sufficient time for patients to adapt to CGM use and reflect any meaningful changes in glycemic control.

All patients received standard DM care according to clinic protocols, which included DM self-management education and training. Patients met with the diabetes educator at least once, during which the educator emphasized making informed decisions using CGM data, such as adjusting dietary choices and physical activity levels to manage blood glucose concentrations effectively.

A total of 302 patients were initially identified. After applying exclusion criteria, 132 were excluded due to insulin use, and 77 were excluded due to incomplete HbA_1c data within the specified time frames (Figure 1). The final sample included 93 patients.

Measures

The primary outcome was the change in HbA_1c levels from baseline to 1 year after CGM initiation. Secondary outcomes included changes in weight, systolic and diastolic BP, LDL-C concentrations, and eGFR. For the primary outcome, HbA_1c values were collected within a grace period of ± 4 months from the baseline and 1-year time points. The laboratory’s upper reporting limit for HbA_1c was 14%; values reported as “> 14%” were recorded as 14.1% for data analysis, although the actual values could have been higher.

For secondary outcomes, data were included if measurements were obtained within ± 6 months of the baseline and 1-year time points. Patients who did not have measurements within these time frames for specific metrics were excluded from secondary outcome analysis but remained in the overall study if they met the criteria for HbA_1c and CGM use.

Statistical Analysis

Statistical analysis was performed using R statistical software version 4.4.2. Paired t tests were conducted to compare baseline and 1-year follow- up measurements for variables with parametric distributions. Wilcoxon signed-rank test was used for nonparametric data. A linear regression analysis was conducted to examine the relationship between baseline HbA_1c levels and the change in HbA_1c after 1 year of CGM use. Differences were considered significant at P < .05 set a priori. To guide future research, a posthoc power analysis was performed using Cohen’s d to estimate the required sample sizes for detecting significant effects, assuming a similar population.

Results

The study included 93 patients, with a mean (SD) age of 55 (13) years (range, 29-83 years). Of the participants, 56 were female (60%) and 37 were male (40%). All participants were identified as AI/AN and had non–insulin-dependent T2DM.

Primary Outcomes

A significant reduction in HbA_1c levels was observed after 1 year of CGM use. The mean (SD) baseline HbA1c was 9.5% (2.4%), which decreased to 7.6% (2.2%) at 1-year follow-up (Table 1). This difference represents a mean change of -1.86% (2.4%) (95% CI, -2.35 to -1.37; P < .001 [paired t test, -7.53]).

A linear regression model evaluated the relationship between baseline HbA_1c (predictor) and the change in HbA_1c after 1 year (outcome). The change in HbA_1c was calculated as the difference between 1-year follow-up and baseline values. The regression model revealed a significant negative association between baseline HbA_1c and the change in HbA_1c (Β = -0.576; P < .001), indicating that higher baseline HbA_1c values were associated with greater reductions in HbA_1c over the year. The regression equation was: Change in HbA_1c = 3.587 – 0.576 × Baseline HbA_1c

The regression coefficient for baseline HbA_1c was -0.576 (standard error, 0.083; t = -6.931; P < .001), indicating that for each 1% increase in baseline HbA_1c, the reduction of HbA_1c after 1 year increased by approximately 0.576% (Figure 2). The model explained 34.6% of the variance in HbA_1c change (R² = .345; adjusted R² = .338).

Secondary Outcomes

Systolic BP decreased by a mean (SD) -4.9 (17) mm Hg; 95% CI, -8.6 to -1.11; P = .01, paired t test). However, no significant change was observed for diastolic BP (P = .77, paired t test). Similarly, no significant changes were observed in weight, LDL-C concentrations, or eGFR after 1 year of CGM use. A posthoc power analysis indicated that the study was underpowered to detect smaller effect sizes in secondary outcomes. For example, sample size estimates indicated that detecting significant changes in weight and LDL-C concentrations would require sample sizes of 152 and 220 patients, respectively (Table 2).

Discussion

This study found a clinically significant reduction in HbA_1c levels after 1 year among AI/AN patients with non–insulin-dependent T2DM who used CGMs. The mean HbA_1c decreased 1.9%, from 9.5% at baseline to 7.6% after 1 year. This reduction is not only statistically significant (P < .001), it is clinically meaningful—even a 1% decrease in HbA_1c is associated with substantial reductions in the risk of microvascular complications.³ The magnitude of the HbA_1c reduction observed suggests CGM use may be associated with improved glycemic control in this high-risk population. By achieving lower HbA_1c levels, patients may experience improved long-term health outcomes and a reduced burden of DM-related complications.

Changes in oral DM medications during the study period may have contributed to the observed improvements in HbA_1c levels. While the dataset lacked detailed information on types or dosages of oral hypoglycemic agents used, adjustments in medication regimens are common in DM management and could significantly affect glycemic control. The inability to account for these changes results in an inability to attribute the improvements in HbA_1c solely to CGM use. Future studies should collect comprehensive medication data to better isolate the effects of CGM use from other treatment modifications.

Another factor that may have contributed to the improved glycemic control is the DM self-management education and training patients received as part of standard care. Patients met with diabetes educators at least once and learned how to use the CGM device and interpret the data for self-management decisions. This education may have enhanced patient engagement and empowerment, enabling them to make informed choices about diet, physical activity, and medication adherence. Studies have shown that DM self-management education can significantly improve glycemic control and patient outcomes.¹³ By combining the CGM technology with targeted education, patients may have been better equipped to manage their condition, contributing to the observed reduction in HbA_1c levels. Future studies should consider synergistic effects of CGM use and DM education when evaluating interventions for glycemic control.

The significant reduction in HbA_1c indicates CGM use is associated with improved glycemic control in non–insulin-dependent T2DM. The linear regression analysis suggests patients with poorer glycemic control at baseline experienced greater reductions in HbA_1c over the course of 1 year. This finding aligns with previous studies that have shown greater HbA_1c reductions in patients with higher initial levels when using CGMs. Yaron et al reported similar findings: higher baseline HbA_1c levels predicted more substantial improvements with CGM use in patients with T2DM on insulin therapy.¹⁴

This study contributes to existing research by examining the association between CGM use and glycemic control in patients with non– insulin-dependent T2DM within an AI/AN population, a group that has been underreported in previous studies. Most prior research has focused on insulin-dependent patients or populations with different ethnic backgrounds.12 By focusing on patients with non–insulin-dependent T2DM, this study highlights the broader applicability of CGMs beyond traditional use, showcasing their potential association with benefits in earlier stages of DM management. Targeting the AI/AN population addresses a critical knowledge gap, given the disproportionately high prevalence of T2DM and associated complications in this group. The findings of this study suggest integrating CGM technology into the standard care of AI/AN patients with non–insulin-dependent T2DM may be associated with improved glycemic control and may help reduce health disparities.

The modest decrease in systolic BP observed in this study may indicate potential cardiovascular benefits associated with CGM use, possibly due to improved glycemic control and increased patient engagement in self-management. However, given the limited sample size and exclusion criteria, the study lacked sufficient power to detect significant associations between CGM use and other secondary outcomes such as BP, weight, LDL-C, and eGFR. Therefore, the significant finding with systolic BP should be interpreted with caution.

The lack of significant changes in secondary outcomes may be attributed to the study’s limited sample size and the relatively short duration for observing changes in these parameters. Larger studies are needed to assess the full impact of CGM on these variables. The required sample sizes for achieving adequate power in future studies were calculated, highlighting the utility of our study as a pilot, providing critical data for the design of larger, adequately powered studies.

Limitations

The retrospective design of this study limits causal inferences. Moreover, potential confounding variables were not controlled, such as changes in medication regimens (other than insulin use), dietary counseling, or physical activity. Additionally, we could not account for the type or number of oral DM medications prescribed to patients. The dataset included only information on insulin use, without detailed records of other antidiabetic medications. This limitation may have influenced the observed change in glycemic control, as variations in medication regimens could affect HbA_1c levels.

Because this study lacked a comparator group, the effect of CGM use cannot be definitively isolated from other factors (eg, medication changes, dietary modifications, or physical activity). Moreover, CGM devices can be costly and are not universally covered by all insurance or IHS programs, potentially limiting widespread implementation. Policy-level restrictions and patient-specific barriers may also hinder feasibility in other settings.

The small sample size may limit the generalizability of the findings. Of the initial 302 patients, about 69% were excluded due to insulin use or incomplete laboratory data. A ± 4-month window was selected to balance data quality with real-world practices. Extending this window further (eg, ± 6 months) might have included more participants but risked diluting the 1-year endpoint consistency. The lack of statistical significance in secondary metrics may be due to insufficient power rather than the absence of an effect.

Exclusion of patients due to incomplete data may have introduced selection bias. However, patients were included in the overall analysis if they met the criteria for HbA_1c and CGM use, even if they lacked data for secondary outcomes. Additionally, the laboratory’s upper reporting limit for HbA_1c was 14%, with values above this reported as “> 14%.” For analysis, these were recorded as 14.1%, which may underestimate the true baseline HbA_1c levels and impact of the assessment of change. This occurred for 4 of the 93 patients included.

All patients used the Freestyle Libre CGM, which may limit the generalizability of the findings to other CGM brands or models. Differences in device features, accuracy, scanning frequency, and user experience may influence outcomes, and results might differ with other CGM technologies. The dataset did not include patients’ scanning frequency because this metric was not consistently included in the EHRs.

Conclusions

This study found that CGM use was significantly associated with improved glycemic control in patients with non–insulin-dependent T2DM within an AI/AN population, particularly among patients with higher baseline HbA_1c levels. The findings suggest that CGMs may be a valuable tool for managing T2DM beyond insulin-dependent populations.

Additional research with larger sample sizes, control groups, and extended follow-up periods is recommended to explore long-term benefits and impacts on other health metrics. The sample size estimates derived from this study serve as a valuable resource for researchers designing future studies aimed at addressing these gaps. Future research that expands on our findings by including larger, more diverse cohorts, accounting for medication use, and exploring different CGM technologies will enhance understanding and contribute to more effective diabetes management strategies for varied populations.

Diabetes mellitus (DM) is a national health crisis affecting > 38 million people (11.6%) in the United States.¹ American Indian and Alaska Native (AI/AN) adults are disproportionately affected, with a prevalence of 14.5%—the highest among all racial and ethnic groups.¹ Type 2 DM (T2DM) accounts for 90% to 95% of all DM cases and is a leading cause of morbidity and mortality due to its association with cardiovascular disease, kidney failure, and other complications.²

Maintaining glycemic control is important for managing T2DM and preventing microvascular and macrovascular complications.³ The cornerstone of diabetes self-management has been patient self-monitored blood glucose (SMBG) using finger-stick glucometers.⁴ However, SMBG provides measurements from a single point in time and requires frequent, painful, and inconvenient finger pricks, leading to decreased adherence.^5,6 These limitations negatively affect patient engagement and overall glycemic control.⁷

Continuous glucose monitors (CGMs) offer real-time, continuous glucose readings and trends.⁸ CGMs improve glycemic control and reduce hypoglycemic episodes in patients who are insulin-dependent.^9,10 Flash glucose monitors, a type of CGM that requires scanning to obtain glucose readings, provide similar benefits.¹¹ Despite these demonstrated advantages, research has primarily focused on insulin-dependent populations, leaving a significant gap in understanding the effect of CGMs on patients with T2DM who are not insulin-dependent.¹²

Given the high prevalence of T2DM among AI/AN populations and the potential benefits of CGMs, this study sought to evaluate the effect of CGM use on glycemic control and other health metrics in patients with non–insulin-dependent T2DM in an AI/AN population. This focus addresses a critical knowledge gap and may inform clinical practices and policies to improve diabetes management in this high-risk group.

Methods

A retrospective observational study was conducted using deidentified electronic health records (EHRs) from 2019 to 2024 at a federally operated outpatient Indian Health Service (IHS) clinic serving an AI/AN population in the IHS Portland Area (Oregon, Washington, Idaho). The study protocol was reviewed and deemed exempt by institutional review boards at Washington State University and the Portland Area IHS.

Study Population

This study included patients diagnosed with non–insulin-dependent T2DM, had used a CGM for ≥ 1 year, and had hemoglobin A_1c (HbA_1c) measurements within 4 months prior to CGM initiation (baseline) and within ± 4 months after 1 year of CGM use. For other health metrics, including blood pressure (BP), weight, low-density lipoprotein cholesterol (LDL-C), and estimated glomerular filtration rate (eGFR), this study required measurements within 6 months before CGM initiation and within 6 months after 1 year of CGM use. The baseline HbA_1c in the dataset ranged from 5.3% to > 14%.

Patients were excluded if they used insulin during the study period, had incomplete laboratory or clinical data for the required time frame, or had < 1 year of CGM use. The dataset did not include detailed information on oral DM medications; thus, we could not report or account for the type or number of oral hypoglycemic agents used by the patients. The IHS clinical applications coordinator compiled the dataset from the EHR, identifying patients who were prescribed and received a CGM at the clinic. All patients used the Abbott Freestyle Libre CGM, the only formulary CGM available at the clinic during the study period.

A 1-year follow-up endpoint was selected for several reasons: (1) to capture potential seasonal variations in diet and activity; (2) to align with the clinic’s standard practice of annual comprehensive diabetes evaluations; and (3) to allow sufficient time for patients to adapt to CGM use and reflect any meaningful changes in glycemic control.

All patients received standard DM care according to clinic protocols, which included DM self-management education and training. Patients met with the diabetes educator at least once, during which the educator emphasized making informed decisions using CGM data, such as adjusting dietary choices and physical activity levels to manage blood glucose concentrations effectively.

A total of 302 patients were initially identified. After applying exclusion criteria, 132 were excluded due to insulin use, and 77 were excluded due to incomplete HbA_1c data within the specified time frames (Figure 1). The final sample included 93 patients.

Measures

The primary outcome was the change in HbA_1c levels from baseline to 1 year after CGM initiation. Secondary outcomes included changes in weight, systolic and diastolic BP, LDL-C concentrations, and eGFR. For the primary outcome, HbA_1c values were collected within a grace period of ± 4 months from the baseline and 1-year time points. The laboratory’s upper reporting limit for HbA_1c was 14%; values reported as “> 14%” were recorded as 14.1% for data analysis, although the actual values could have been higher.

For secondary outcomes, data were included if measurements were obtained within ± 6 months of the baseline and 1-year time points. Patients who did not have measurements within these time frames for specific metrics were excluded from secondary outcome analysis but remained in the overall study if they met the criteria for HbA_1c and CGM use.

Statistical Analysis

Statistical analysis was performed using R statistical software version 4.4.2. Paired t tests were conducted to compare baseline and 1-year follow- up measurements for variables with parametric distributions. Wilcoxon signed-rank test was used for nonparametric data. A linear regression analysis was conducted to examine the relationship between baseline HbA_1c levels and the change in HbA_1c after 1 year of CGM use. Differences were considered significant at P < .05 set a priori. To guide future research, a posthoc power analysis was performed using Cohen’s d to estimate the required sample sizes for detecting significant effects, assuming a similar population.

Results

The study included 93 patients, with a mean (SD) age of 55 (13) years (range, 29-83 years). Of the participants, 56 were female (60%) and 37 were male (40%). All participants were identified as AI/AN and had non–insulin-dependent T2DM.

Primary Outcomes

A significant reduction in HbA_1c levels was observed after 1 year of CGM use. The mean (SD) baseline HbA1c was 9.5% (2.4%), which decreased to 7.6% (2.2%) at 1-year follow-up (Table 1). This difference represents a mean change of -1.86% (2.4%) (95% CI, -2.35 to -1.37; P < .001 [paired t test, -7.53]).

A linear regression model evaluated the relationship between baseline HbA_1c (predictor) and the change in HbA_1c after 1 year (outcome). The change in HbA_1c was calculated as the difference between 1-year follow-up and baseline values. The regression model revealed a significant negative association between baseline HbA_1c and the change in HbA_1c (Β = -0.576; P < .001), indicating that higher baseline HbA_1c values were associated with greater reductions in HbA_1c over the year. The regression equation was: Change in HbA_1c = 3.587 – 0.576 × Baseline HbA_1c

The regression coefficient for baseline HbA_1c was -0.576 (standard error, 0.083; t = -6.931; P < .001), indicating that for each 1% increase in baseline HbA_1c, the reduction of HbA_1c after 1 year increased by approximately 0.576% (Figure 2). The model explained 34.6% of the variance in HbA_1c change (R² = .345; adjusted R² = .338).

Secondary Outcomes

Systolic BP decreased by a mean (SD) -4.9 (17) mm Hg; 95% CI, -8.6 to -1.11; P = .01, paired t test). However, no significant change was observed for diastolic BP (P = .77, paired t test). Similarly, no significant changes were observed in weight, LDL-C concentrations, or eGFR after 1 year of CGM use. A posthoc power analysis indicated that the study was underpowered to detect smaller effect sizes in secondary outcomes. For example, sample size estimates indicated that detecting significant changes in weight and LDL-C concentrations would require sample sizes of 152 and 220 patients, respectively (Table 2).

Discussion

This study found a clinically significant reduction in HbA_1c levels after 1 year among AI/AN patients with non–insulin-dependent T2DM who used CGMs. The mean HbA_1c decreased 1.9%, from 9.5% at baseline to 7.6% after 1 year. This reduction is not only statistically significant (P < .001), it is clinically meaningful—even a 1% decrease in HbA_1c is associated with substantial reductions in the risk of microvascular complications.³ The magnitude of the HbA_1c reduction observed suggests CGM use may be associated with improved glycemic control in this high-risk population. By achieving lower HbA_1c levels, patients may experience improved long-term health outcomes and a reduced burden of DM-related complications.

Changes in oral DM medications during the study period may have contributed to the observed improvements in HbA_1c levels. While the dataset lacked detailed information on types or dosages of oral hypoglycemic agents used, adjustments in medication regimens are common in DM management and could significantly affect glycemic control. The inability to account for these changes results in an inability to attribute the improvements in HbA_1c solely to CGM use. Future studies should collect comprehensive medication data to better isolate the effects of CGM use from other treatment modifications.

Another factor that may have contributed to the improved glycemic control is the DM self-management education and training patients received as part of standard care. Patients met with diabetes educators at least once and learned how to use the CGM device and interpret the data for self-management decisions. This education may have enhanced patient engagement and empowerment, enabling them to make informed choices about diet, physical activity, and medication adherence. Studies have shown that DM self-management education can significantly improve glycemic control and patient outcomes.¹³ By combining the CGM technology with targeted education, patients may have been better equipped to manage their condition, contributing to the observed reduction in HbA_1c levels. Future studies should consider synergistic effects of CGM use and DM education when evaluating interventions for glycemic control.

The significant reduction in HbA_1c indicates CGM use is associated with improved glycemic control in non–insulin-dependent T2DM. The linear regression analysis suggests patients with poorer glycemic control at baseline experienced greater reductions in HbA_1c over the course of 1 year. This finding aligns with previous studies that have shown greater HbA_1c reductions in patients with higher initial levels when using CGMs. Yaron et al reported similar findings: higher baseline HbA_1c levels predicted more substantial improvements with CGM use in patients with T2DM on insulin therapy.¹⁴

This study contributes to existing research by examining the association between CGM use and glycemic control in patients with non– insulin-dependent T2DM within an AI/AN population, a group that has been underreported in previous studies. Most prior research has focused on insulin-dependent patients or populations with different ethnic backgrounds.12 By focusing on patients with non–insulin-dependent T2DM, this study highlights the broader applicability of CGMs beyond traditional use, showcasing their potential association with benefits in earlier stages of DM management. Targeting the AI/AN population addresses a critical knowledge gap, given the disproportionately high prevalence of T2DM and associated complications in this group. The findings of this study suggest integrating CGM technology into the standard care of AI/AN patients with non–insulin-dependent T2DM may be associated with improved glycemic control and may help reduce health disparities.

The modest decrease in systolic BP observed in this study may indicate potential cardiovascular benefits associated with CGM use, possibly due to improved glycemic control and increased patient engagement in self-management. However, given the limited sample size and exclusion criteria, the study lacked sufficient power to detect significant associations between CGM use and other secondary outcomes such as BP, weight, LDL-C, and eGFR. Therefore, the significant finding with systolic BP should be interpreted with caution.

The lack of significant changes in secondary outcomes may be attributed to the study’s limited sample size and the relatively short duration for observing changes in these parameters. Larger studies are needed to assess the full impact of CGM on these variables. The required sample sizes for achieving adequate power in future studies were calculated, highlighting the utility of our study as a pilot, providing critical data for the design of larger, adequately powered studies.

Limitations

The retrospective design of this study limits causal inferences. Moreover, potential confounding variables were not controlled, such as changes in medication regimens (other than insulin use), dietary counseling, or physical activity. Additionally, we could not account for the type or number of oral DM medications prescribed to patients. The dataset included only information on insulin use, without detailed records of other antidiabetic medications. This limitation may have influenced the observed change in glycemic control, as variations in medication regimens could affect HbA_1c levels.

Because this study lacked a comparator group, the effect of CGM use cannot be definitively isolated from other factors (eg, medication changes, dietary modifications, or physical activity). Moreover, CGM devices can be costly and are not universally covered by all insurance or IHS programs, potentially limiting widespread implementation. Policy-level restrictions and patient-specific barriers may also hinder feasibility in other settings.

The small sample size may limit the generalizability of the findings. Of the initial 302 patients, about 69% were excluded due to insulin use or incomplete laboratory data. A ± 4-month window was selected to balance data quality with real-world practices. Extending this window further (eg, ± 6 months) might have included more participants but risked diluting the 1-year endpoint consistency. The lack of statistical significance in secondary metrics may be due to insufficient power rather than the absence of an effect.

Exclusion of patients due to incomplete data may have introduced selection bias. However, patients were included in the overall analysis if they met the criteria for HbA_1c and CGM use, even if they lacked data for secondary outcomes. Additionally, the laboratory’s upper reporting limit for HbA_1c was 14%, with values above this reported as “> 14%.” For analysis, these were recorded as 14.1%, which may underestimate the true baseline HbA_1c levels and impact of the assessment of change. This occurred for 4 of the 93 patients included.

All patients used the Freestyle Libre CGM, which may limit the generalizability of the findings to other CGM brands or models. Differences in device features, accuracy, scanning frequency, and user experience may influence outcomes, and results might differ with other CGM technologies. The dataset did not include patients’ scanning frequency because this metric was not consistently included in the EHRs.

Conclusions

This study found that CGM use was significantly associated with improved glycemic control in patients with non–insulin-dependent T2DM within an AI/AN population, particularly among patients with higher baseline HbA_1c levels. The findings suggest that CGMs may be a valuable tool for managing T2DM beyond insulin-dependent populations.

Additional research with larger sample sizes, control groups, and extended follow-up periods is recommended to explore long-term benefits and impacts on other health metrics. The sample size estimates derived from this study serve as a valuable resource for researchers designing future studies aimed at addressing these gaps. Future research that expands on our findings by including larger, more diverse cohorts, accounting for medication use, and exploring different CGM technologies will enhance understanding and contribute to more effective diabetes management strategies for varied populations.

Diabetes mellitus (DM) is a national health crisis affecting > 38 million people (11.6%) in the United States.¹ American Indian and Alaska Native (AI/AN) adults are disproportionately affected, with a prevalence of 14.5%—the highest among all racial and ethnic groups.¹ Type 2 DM (T2DM) accounts for 90% to 95% of all DM cases and is a leading cause of morbidity and mortality due to its association with cardiovascular disease, kidney failure, and other complications.²

Maintaining glycemic control is important for managing T2DM and preventing microvascular and macrovascular complications.³ The cornerstone of diabetes self-management has been patient self-monitored blood glucose (SMBG) using finger-stick glucometers.⁴ However, SMBG provides measurements from a single point in time and requires frequent, painful, and inconvenient finger pricks, leading to decreased adherence.^5,6 These limitations negatively affect patient engagement and overall glycemic control.⁷

Continuous glucose monitors (CGMs) offer real-time, continuous glucose readings and trends.⁸ CGMs improve glycemic control and reduce hypoglycemic episodes in patients who are insulin-dependent.^9,10 Flash glucose monitors, a type of CGM that requires scanning to obtain glucose readings, provide similar benefits.¹¹ Despite these demonstrated advantages, research has primarily focused on insulin-dependent populations, leaving a significant gap in understanding the effect of CGMs on patients with T2DM who are not insulin-dependent.¹²

Given the high prevalence of T2DM among AI/AN populations and the potential benefits of CGMs, this study sought to evaluate the effect of CGM use on glycemic control and other health metrics in patients with non–insulin-dependent T2DM in an AI/AN population. This focus addresses a critical knowledge gap and may inform clinical practices and policies to improve diabetes management in this high-risk group.

Methods

A retrospective observational study was conducted using deidentified electronic health records (EHRs) from 2019 to 2024 at a federally operated outpatient Indian Health Service (IHS) clinic serving an AI/AN population in the IHS Portland Area (Oregon, Washington, Idaho). The study protocol was reviewed and deemed exempt by institutional review boards at Washington State University and the Portland Area IHS.

Study Population

This study included patients diagnosed with non–insulin-dependent T2DM, had used a CGM for ≥ 1 year, and had hemoglobin A_1c (HbA_1c) measurements within 4 months prior to CGM initiation (baseline) and within ± 4 months after 1 year of CGM use. For other health metrics, including blood pressure (BP), weight, low-density lipoprotein cholesterol (LDL-C), and estimated glomerular filtration rate (eGFR), this study required measurements within 6 months before CGM initiation and within 6 months after 1 year of CGM use. The baseline HbA_1c in the dataset ranged from 5.3% to > 14%.

Patients were excluded if they used insulin during the study period, had incomplete laboratory or clinical data for the required time frame, or had < 1 year of CGM use. The dataset did not include detailed information on oral DM medications; thus, we could not report or account for the type or number of oral hypoglycemic agents used by the patients. The IHS clinical applications coordinator compiled the dataset from the EHR, identifying patients who were prescribed and received a CGM at the clinic. All patients used the Abbott Freestyle Libre CGM, the only formulary CGM available at the clinic during the study period.

A 1-year follow-up endpoint was selected for several reasons: (1) to capture potential seasonal variations in diet and activity; (2) to align with the clinic’s standard practice of annual comprehensive diabetes evaluations; and (3) to allow sufficient time for patients to adapt to CGM use and reflect any meaningful changes in glycemic control.

All patients received standard DM care according to clinic protocols, which included DM self-management education and training. Patients met with the diabetes educator at least once, during which the educator emphasized making informed decisions using CGM data, such as adjusting dietary choices and physical activity levels to manage blood glucose concentrations effectively.

A total of 302 patients were initially identified. After applying exclusion criteria, 132 were excluded due to insulin use, and 77 were excluded due to incomplete HbA_1c data within the specified time frames (Figure 1). The final sample included 93 patients.

Measures

The primary outcome was the change in HbA_1c levels from baseline to 1 year after CGM initiation. Secondary outcomes included changes in weight, systolic and diastolic BP, LDL-C concentrations, and eGFR. For the primary outcome, HbA_1c values were collected within a grace period of ± 4 months from the baseline and 1-year time points. The laboratory’s upper reporting limit for HbA_1c was 14%; values reported as “> 14%” were recorded as 14.1% for data analysis, although the actual values could have been higher.

For secondary outcomes, data were included if measurements were obtained within ± 6 months of the baseline and 1-year time points. Patients who did not have measurements within these time frames for specific metrics were excluded from secondary outcome analysis but remained in the overall study if they met the criteria for HbA_1c and CGM use.

Statistical Analysis

Statistical analysis was performed using R statistical software version 4.4.2. Paired t tests were conducted to compare baseline and 1-year follow- up measurements for variables with parametric distributions. Wilcoxon signed-rank test was used for nonparametric data. A linear regression analysis was conducted to examine the relationship between baseline HbA_1c levels and the change in HbA_1c after 1 year of CGM use. Differences were considered significant at P < .05 set a priori. To guide future research, a posthoc power analysis was performed using Cohen’s d to estimate the required sample sizes for detecting significant effects, assuming a similar population.

Results

The study included 93 patients, with a mean (SD) age of 55 (13) years (range, 29-83 years). Of the participants, 56 were female (60%) and 37 were male (40%). All participants were identified as AI/AN and had non–insulin-dependent T2DM.

Primary Outcomes

A significant reduction in HbA_1c levels was observed after 1 year of CGM use. The mean (SD) baseline HbA1c was 9.5% (2.4%), which decreased to 7.6% (2.2%) at 1-year follow-up (Table 1). This difference represents a mean change of -1.86% (2.4%) (95% CI, -2.35 to -1.37; P < .001 [paired t test, -7.53]).

A linear regression model evaluated the relationship between baseline HbA_1c (predictor) and the change in HbA_1c after 1 year (outcome). The change in HbA_1c was calculated as the difference between 1-year follow-up and baseline values. The regression model revealed a significant negative association between baseline HbA_1c and the change in HbA_1c (Β = -0.576; P < .001), indicating that higher baseline HbA_1c values were associated with greater reductions in HbA_1c over the year. The regression equation was: Change in HbA_1c = 3.587 – 0.576 × Baseline HbA_1c

The regression coefficient for baseline HbA_1c was -0.576 (standard error, 0.083; t = -6.931; P < .001), indicating that for each 1% increase in baseline HbA_1c, the reduction of HbA_1c after 1 year increased by approximately 0.576% (Figure 2). The model explained 34.6% of the variance in HbA_1c change (R² = .345; adjusted R² = .338).

Secondary Outcomes

Systolic BP decreased by a mean (SD) -4.9 (17) mm Hg; 95% CI, -8.6 to -1.11; P = .01, paired t test). However, no significant change was observed for diastolic BP (P = .77, paired t test). Similarly, no significant changes were observed in weight, LDL-C concentrations, or eGFR after 1 year of CGM use. A posthoc power analysis indicated that the study was underpowered to detect smaller effect sizes in secondary outcomes. For example, sample size estimates indicated that detecting significant changes in weight and LDL-C concentrations would require sample sizes of 152 and 220 patients, respectively (Table 2).

Discussion

This study found a clinically significant reduction in HbA_1c levels after 1 year among AI/AN patients with non–insulin-dependent T2DM who used CGMs. The mean HbA_1c decreased 1.9%, from 9.5% at baseline to 7.6% after 1 year. This reduction is not only statistically significant (P < .001), it is clinically meaningful—even a 1% decrease in HbA_1c is associated with substantial reductions in the risk of microvascular complications.³ The magnitude of the HbA_1c reduction observed suggests CGM use may be associated with improved glycemic control in this high-risk population. By achieving lower HbA_1c levels, patients may experience improved long-term health outcomes and a reduced burden of DM-related complications.

Changes in oral DM medications during the study period may have contributed to the observed improvements in HbA_1c levels. While the dataset lacked detailed information on types or dosages of oral hypoglycemic agents used, adjustments in medication regimens are common in DM management and could significantly affect glycemic control. The inability to account for these changes results in an inability to attribute the improvements in HbA_1c solely to CGM use. Future studies should collect comprehensive medication data to better isolate the effects of CGM use from other treatment modifications.

Another factor that may have contributed to the improved glycemic control is the DM self-management education and training patients received as part of standard care. Patients met with diabetes educators at least once and learned how to use the CGM device and interpret the data for self-management decisions. This education may have enhanced patient engagement and empowerment, enabling them to make informed choices about diet, physical activity, and medication adherence. Studies have shown that DM self-management education can significantly improve glycemic control and patient outcomes.¹³ By combining the CGM technology with targeted education, patients may have been better equipped to manage their condition, contributing to the observed reduction in HbA_1c levels. Future studies should consider synergistic effects of CGM use and DM education when evaluating interventions for glycemic control.

The significant reduction in HbA_1c indicates CGM use is associated with improved glycemic control in non–insulin-dependent T2DM. The linear regression analysis suggests patients with poorer glycemic control at baseline experienced greater reductions in HbA_1c over the course of 1 year. This finding aligns with previous studies that have shown greater HbA_1c reductions in patients with higher initial levels when using CGMs. Yaron et al reported similar findings: higher baseline HbA_1c levels predicted more substantial improvements with CGM use in patients with T2DM on insulin therapy.¹⁴

This study contributes to existing research by examining the association between CGM use and glycemic control in patients with non– insulin-dependent T2DM within an AI/AN population, a group that has been underreported in previous studies. Most prior research has focused on insulin-dependent patients or populations with different ethnic backgrounds.12 By focusing on patients with non–insulin-dependent T2DM, this study highlights the broader applicability of CGMs beyond traditional use, showcasing their potential association with benefits in earlier stages of DM management. Targeting the AI/AN population addresses a critical knowledge gap, given the disproportionately high prevalence of T2DM and associated complications in this group. The findings of this study suggest integrating CGM technology into the standard care of AI/AN patients with non–insulin-dependent T2DM may be associated with improved glycemic control and may help reduce health disparities.

The modest decrease in systolic BP observed in this study may indicate potential cardiovascular benefits associated with CGM use, possibly due to improved glycemic control and increased patient engagement in self-management. However, given the limited sample size and exclusion criteria, the study lacked sufficient power to detect significant associations between CGM use and other secondary outcomes such as BP, weight, LDL-C, and eGFR. Therefore, the significant finding with systolic BP should be interpreted with caution.

The lack of significant changes in secondary outcomes may be attributed to the study’s limited sample size and the relatively short duration for observing changes in these parameters. Larger studies are needed to assess the full impact of CGM on these variables. The required sample sizes for achieving adequate power in future studies were calculated, highlighting the utility of our study as a pilot, providing critical data for the design of larger, adequately powered studies.

Limitations

The retrospective design of this study limits causal inferences. Moreover, potential confounding variables were not controlled, such as changes in medication regimens (other than insulin use), dietary counseling, or physical activity. Additionally, we could not account for the type or number of oral DM medications prescribed to patients. The dataset included only information on insulin use, without detailed records of other antidiabetic medications. This limitation may have influenced the observed change in glycemic control, as variations in medication regimens could affect HbA_1c levels.

Because this study lacked a comparator group, the effect of CGM use cannot be definitively isolated from other factors (eg, medication changes, dietary modifications, or physical activity). Moreover, CGM devices can be costly and are not universally covered by all insurance or IHS programs, potentially limiting widespread implementation. Policy-level restrictions and patient-specific barriers may also hinder feasibility in other settings.

The small sample size may limit the generalizability of the findings. Of the initial 302 patients, about 69% were excluded due to insulin use or incomplete laboratory data. A ± 4-month window was selected to balance data quality with real-world practices. Extending this window further (eg, ± 6 months) might have included more participants but risked diluting the 1-year endpoint consistency. The lack of statistical significance in secondary metrics may be due to insufficient power rather than the absence of an effect.

Exclusion of patients due to incomplete data may have introduced selection bias. However, patients were included in the overall analysis if they met the criteria for HbA_1c and CGM use, even if they lacked data for secondary outcomes. Additionally, the laboratory’s upper reporting limit for HbA_1c was 14%, with values above this reported as “> 14%.” For analysis, these were recorded as 14.1%, which may underestimate the true baseline HbA_1c levels and impact of the assessment of change. This occurred for 4 of the 93 patients included.

All patients used the Freestyle Libre CGM, which may limit the generalizability of the findings to other CGM brands or models. Differences in device features, accuracy, scanning frequency, and user experience may influence outcomes, and results might differ with other CGM technologies. The dataset did not include patients’ scanning frequency because this metric was not consistently included in the EHRs.

Conclusions

This study found that CGM use was significantly associated with improved glycemic control in patients with non–insulin-dependent T2DM within an AI/AN population, particularly among patients with higher baseline HbA_1c levels. The findings suggest that CGMs may be a valuable tool for managing T2DM beyond insulin-dependent populations.

Additional research with larger sample sizes, control groups, and extended follow-up periods is recommended to explore long-term benefits and impacts on other health metrics. The sample size estimates derived from this study serve as a valuable resource for researchers designing future studies aimed at addressing these gaps. Future research that expands on our findings by including larger, more diverse cohorts, accounting for medication use, and exploring different CGM technologies will enhance understanding and contribute to more effective diabetes management strategies for varied populations.

Cardiovascular disease (CVD) is the leading cause of death among women in the United States.¹ Most CVD is due to the buildup of plaque (ie, cholesterol, proteins, calcium, and inflammatory cells) in artery walls.² The plaque may lead to atherosclerotic cardiovascular disease (ASCVD), which includes coronary heart disease, cerebrovascular disease, peripheral artery disease, and aortic atherosclerotic disease.^2,3 Control and reduction of ASCVD risk factors, including high cholesterol levels, elevated blood pressure, insulin resistance, smoking, and a sedentary lifestyle, can contribute to a reduction in ASCVD morbidity and mortality.² People with type 2 diabetes mellitus (T2DM) have an increased prevalence of lipid abnormalities, contributing to their high risk of ASCVD.^4,5

The prescribing of statins (3-hydroxy-3-methyl-glutaryl-coenzmye A reductase inhibitors) is the cornerstone of lipid-lowering therapy and cardiovascular risk reduction for primary and secondary prevention of ASCVD.⁶ The American Diabetes Association (ADA) and American College of Cardiology/American Heart Association (ACC/AHA) recommend moderate- to high-intensity statins for primary prevention in patients with T2DM and high-intensity statins for secondary prevention in those with or without diabetes when not contraindicated.^4,5,7 Despite eligibility according to guideline recommendations, research predominantly shows that women are less likely to receive statin therapy; however, this trend is improving. ^[6,8-11] To explain the sex differences in statin use, Nanna et al found that there is a combination of women being offered statin therapy less frequently, declining therapy more frequently, and discontinuing treatment more frequently.¹¹ One possibility for discontinuing treatment could be statin-associated muscle symptoms (SAMS), which occur in about 10% of patients.¹² The incidence of adverse effects (AEs) may be related to the way statins are metabolized.

Pharmacogenomic testing is free for veterans through the US Department of Veterans Affairs (VA) PHASER program, which offers information and recommendations for a panel of 11 gene variants. The panel includes genes related to common medication classes such as anticoagulants, antiplatelets, proton pump inhibitors, nonsteroidal anti-inflammatory drugs, opioids, antidepressants, and statins. The VA PHASER panel includes the solute carrier organic anion transporter family member 1B1 (SLCO1B1) gene, which is predominantly expressed in the liver and facilitates the hepatic uptake of most statins.^13,14 A reduced function of SLCO1B1 can lead to higher statin levels, resulting in increased concentrations that may potentially cause SAMS.^13,14 Some alleles associated with reduced function include SLCO1B1*5, *15, *23, *31, and *46 to *49, whereas others are associated with increased function, such as SLCO1B1 *14 and *20 (Appendix).¹⁵ Supporting evidence shows the SLCO1B1*5 nucleotide polymorphism increases plasma levels of simvastatin and atorvastatin, affecting effectiveness or toxicity. ¹³ Females tend to have a lower body weight and higher percentage of body fat compared with males, which might lead to higher concentrations of lipophilic drugs, including atorvastatin and simvastatin, which may be exacerbated by decreased function of SLCO1B1*5.¹⁵ With pharmacogenomic testing, therapeutic recommendations can be made to improve the overall safety and efficacy of statins, thus improving adherence using a patient-specific approach.^14,15

Methods

Carl Vinson VA Medical Center (CVVAMC) serves about 42,000 veterans in Central and South Georgia, of which about 15% are female. Of the female veterans enrolled in care, 63% identify as Black, 27% White, and 1.5% as Asian, American Indian/Alaska Native, or Native Hawaiian/Other Pacific Islander. The 2020 Veterans Chartbook report showed that female veterans and minority racial and ethnic groups had worse access to health care and higher mortality rates than their male and non-Hispanic White counterparts.¹⁶

The Primary Care Equity Dashboard (PCED) was developed to engage the VA health care workforce in the process of identifying and addressing inequities in local patient populations.¹⁷ Using electronic quality measure data, the PCED provides Veterans Integrated Service Network-level and facility-level performance on several metrics.¹⁸ The PCED had not been previously used at the CVVAMC, and few publications or quality improvement projects regarding its use have been reported by the VA Office of Health Equity. PCED helped identify disparities when comparing female to male patients in the prescribing of statin therapy for patients with CVD and statin therapy for patients with T2DM.

VA PHASER pharmacogenomic analyses provided an opportunity to expand this quality improvement project. Sanford Health and the VA collaborated on the PHASER program to offer free genetic testing for veterans. The program launched in 2019 and expanded to various VA sites, including CVVAMC in March 2023. This program has been extended to December 31, 2025.

The primary objective of this quality improvement project was to increase statin prescribing among female veterans with T2DM and/or CVD to reduce cardiovascular risk. Secondary outcomes included increased pharmacogenomic testing and the assessment of pharmacogenomic results related to statin therapy. This project was approved by the CVVAMC Pharmacy and Therapeutics Committee. The PCED was used to identify female veterans with T2DM and/or CVD without an active prescription for a statin between July and October 2023. A review of Computerized Patient Record System patient charts was completed to screen for prespecified inclusion and exclusion criteria. Veterans were included if they were assigned female at birth, were enrolled in care at CVVAMC, and had a diagnosis of T2DM or CVD (history of myocardial infarction, coronary bypass graft, percutaneous coronary intervention, or other revascularization in any setting).

Veterans were excluded if they were currently pregnant, trying to conceive, breastfeeding, had a T1DM diagnosis, had previously documented hypersensitivity to a statin, active liver failure or decompensated cirrhosis, previously documented statin-associated rhabdomyolysis or autoimmune myopathy, an active prescription for a proprotein convertase subtilisin/kexin type 9 inhibitor, or previously documented statin intolerance (defined as the inability to tolerate ≥ 3 statins, with ≥ 1 prescribed at low intensity or alternate-day dosing). The female veterans were compared to 2 comparators: the facility's male veterans and the VA national average, identified via the PCED.

Once a veteran was screened, they were telephoned between October 2023 and February 2024 and provided education on statin use and pharmacogenomic testing using a standardized note template. An order was placed for participants who provided verbal consent for pharmacogenomic testing. Those who agreed to statin initiation were referred to a clinical pharmacist practitioner (CPP) who contacted them at a later date to prescribe a statin following the recommendations of the 2019 ACC/AHA and 2023 ADA guidelines and pharmacogenomic testing, if applicable.^4,5,7 Appropriate monitoring and follow-up occurred at the discretion of each CPP. Data collection included: age, race, diagnoses (T2DM, CVD, or both), baseline lipid panel (total cholesterol, triglycerides, high-density lipoprotein, low-density lipoprotein), hepatic function, name and dose of statin, reasons for declining statin therapy, and pharmacogenomic testing results related to SLCO1B1.

Results

At baseline in July 2023, 77.8% of female veterans with T2DM were prescribed a statin, which exceeded the national VA average (77.0%), but was below the rate for male veterans (78.7%) in the facility comparator group.17 Additionally, 82.2% of females with CVD were prescribed a statin, which was below the national VA average of 86.0% and the 84.9% of male veterans in the facility comparator group.17 The PCED identified 189 female veterans from July 2023 to October 2023 who may benefit from statin therapy. Thirty-three females met the exclusion criteria. Of the 156 included veterans, 129 (82.7%) were successfully contacted and 27 (17.3%) could not be reached by telephone after 3 attempts (Figure 1). The 129 female veterans contacted had a mean age of 59 years and the majority were Black (82.9%) (Table 1).

Primary Outcomes

Of the 129 contacted veterans, 31 (24.0%) had a non-VA statin prescription, 13 (10.1%) had an active VA statin prescription, and 85 (65.9%) did not have a statin prescription, despite being eligible. Statin adherence was confirmed with participants, and the medication list was updated accordingly.

Of the 85 veterans with no active statin therapy, 37 (43.5%) accepted a new statin prescription and 48 (56.5%) declined. There were various reasons provided for declining statin therapy: 17 participants (35.4%) declined due to concern for AEs (Table 2).

From July 2023 to March 2024, the percentage of female veterans with active statin therapy with T2DM increased from 77.8% to 79.0%. For those with active statin therapy with CVD, usage increased from 82.2% to 90.2%, which exceeded the national VA average and facility male comparator group (Figures 2 and 3).¹⁷

Secondary Outcomes

Seventy-one of 129 veterans (55.0%) gave verbal consent, and 47 (66.2%) completed the pharmacogenomic testing; 58 (45.0%) declined. Five veterans (10.6%) had a known SLCO1B1 allele variant present. One veteran required a change in statin therapy based on the results (eAppendix).

Discussion

This project aimed to increase statin prescribing among female veterans with T2DM and/or CVD to reduce cardiovascular risk and increase pharmacogenomic testing using the PCED and care managed by CPPs. The results of this quality improvement project illustrated that both metrics have improved at CVVAMC as a result of the intervention. The results in both metrics now exceed the PCED national VA average, and the CVD metric also exceeds that of the facility male comparator group. While there was only a 1.2% increase from July 2023 to March 2024 for patients with T2DM, there was an 8.0% increase for patients with CVD. Despite standardized education on statin use, more veterans declined therapy than accepted it, mostly due to concern for AEs. Recording the reasons for declining statin therapy offered valuable insight that can be used in additional discussions with veterans and clinicians.

Pharmacogenomics gives clinicians the unique opportunity to take a proactive approach to better predict drug responses, potentially allowing for less trial and error with medications, fewer AEs, greater trust in the clinician, and improved medication adherence. The CPPs incorporated pharmacogenomic testing into their practice, which led to identifying 5 SLCO1B1 gene abnormalities. The PCED served as a powerful tool for advancing equity-focused quality improvement initiatives on a local level and was crucial in prioritizing the detection of veterans potentially receiving suboptimal care.

Limitations

The nature of “cold calls” made it challenging to establish contact for inclusion in this study. An alternative to increase engagement could have been scheduled phone or face-to-face visits. While the use of the PCED was crucial, data did not account for statins listed in the non-VA medication list. All 31 patients with statins prescribed outside the VA had a start date added to provide the most accurate representation of the data moving forward.

Another limitation in this project was its small sample size and population. CVVAMC serves about 6200 female veterans, with roughly 63% identifying as Black. The preponderance of Black individuals (83%) in this project is typical for the female patient population at CVVAMC but may not reflect the demographics of other populations. Other limitations to this project consisted of scheduling conflicts. Appointments for laboratory draws at community-based outpatient clinics were subject to availability, which resulted in some delay in completion of pharmacogenomic testing.

Conclusions

CPPs can help reduce inequity in health care delivery. Increased incorporation of the PCED into regular practice within the VA is recommended to continue addressing sex disparities in statin use, diabetes control, blood pressure management, cancer screenings, and vaccination needs. CVVAMC plans to expand its use through another quality improvement project focused on reducing sex disparities in blood pressure management. Improving educational resources made available to veterans on the importance of statin therapy and potential to mitigate AEs through use of the VA PHASER program also would be helpful. This project successfully improved CVVAMC metrics for female veterans appropriately prescribed statin therapy and increased access to pharmacogenomic testing. Most importantly, it helped close the sex-based gap in CVD risk reduction care.

Cardiovascular disease (CVD) is the leading cause of death among women in the United States.¹ Most CVD is due to the buildup of plaque (ie, cholesterol, proteins, calcium, and inflammatory cells) in artery walls.² The plaque may lead to atherosclerotic cardiovascular disease (ASCVD), which includes coronary heart disease, cerebrovascular disease, peripheral artery disease, and aortic atherosclerotic disease.^2,3 Control and reduction of ASCVD risk factors, including high cholesterol levels, elevated blood pressure, insulin resistance, smoking, and a sedentary lifestyle, can contribute to a reduction in ASCVD morbidity and mortality.² People with type 2 diabetes mellitus (T2DM) have an increased prevalence of lipid abnormalities, contributing to their high risk of ASCVD.^4,5

The prescribing of statins (3-hydroxy-3-methyl-glutaryl-coenzmye A reductase inhibitors) is the cornerstone of lipid-lowering therapy and cardiovascular risk reduction for primary and secondary prevention of ASCVD.⁶ The American Diabetes Association (ADA) and American College of Cardiology/American Heart Association (ACC/AHA) recommend moderate- to high-intensity statins for primary prevention in patients with T2DM and high-intensity statins for secondary prevention in those with or without diabetes when not contraindicated.^4,5,7 Despite eligibility according to guideline recommendations, research predominantly shows that women are less likely to receive statin therapy; however, this trend is improving. ^[6,8-11] To explain the sex differences in statin use, Nanna et al found that there is a combination of women being offered statin therapy less frequently, declining therapy more frequently, and discontinuing treatment more frequently.¹¹ One possibility for discontinuing treatment could be statin-associated muscle symptoms (SAMS), which occur in about 10% of patients.¹² The incidence of adverse effects (AEs) may be related to the way statins are metabolized.

Pharmacogenomic testing is free for veterans through the US Department of Veterans Affairs (VA) PHASER program, which offers information and recommendations for a panel of 11 gene variants. The panel includes genes related to common medication classes such as anticoagulants, antiplatelets, proton pump inhibitors, nonsteroidal anti-inflammatory drugs, opioids, antidepressants, and statins. The VA PHASER panel includes the solute carrier organic anion transporter family member 1B1 (SLCO1B1) gene, which is predominantly expressed in the liver and facilitates the hepatic uptake of most statins.^13,14 A reduced function of SLCO1B1 can lead to higher statin levels, resulting in increased concentrations that may potentially cause SAMS.^13,14 Some alleles associated with reduced function include SLCO1B1*5, *15, *23, *31, and *46 to *49, whereas others are associated with increased function, such as SLCO1B1 *14 and *20 (Appendix).¹⁵ Supporting evidence shows the SLCO1B1*5 nucleotide polymorphism increases plasma levels of simvastatin and atorvastatin, affecting effectiveness or toxicity. ¹³ Females tend to have a lower body weight and higher percentage of body fat compared with males, which might lead to higher concentrations of lipophilic drugs, including atorvastatin and simvastatin, which may be exacerbated by decreased function of SLCO1B1*5.¹⁵ With pharmacogenomic testing, therapeutic recommendations can be made to improve the overall safety and efficacy of statins, thus improving adherence using a patient-specific approach.^14,15

Methods

Carl Vinson VA Medical Center (CVVAMC) serves about 42,000 veterans in Central and South Georgia, of which about 15% are female. Of the female veterans enrolled in care, 63% identify as Black, 27% White, and 1.5% as Asian, American Indian/Alaska Native, or Native Hawaiian/Other Pacific Islander. The 2020 Veterans Chartbook report showed that female veterans and minority racial and ethnic groups had worse access to health care and higher mortality rates than their male and non-Hispanic White counterparts.¹⁶

The Primary Care Equity Dashboard (PCED) was developed to engage the VA health care workforce in the process of identifying and addressing inequities in local patient populations.¹⁷ Using electronic quality measure data, the PCED provides Veterans Integrated Service Network-level and facility-level performance on several metrics.¹⁸ The PCED had not been previously used at the CVVAMC, and few publications or quality improvement projects regarding its use have been reported by the VA Office of Health Equity. PCED helped identify disparities when comparing female to male patients in the prescribing of statin therapy for patients with CVD and statin therapy for patients with T2DM.

VA PHASER pharmacogenomic analyses provided an opportunity to expand this quality improvement project. Sanford Health and the VA collaborated on the PHASER program to offer free genetic testing for veterans. The program launched in 2019 and expanded to various VA sites, including CVVAMC in March 2023. This program has been extended to December 31, 2025.

The primary objective of this quality improvement project was to increase statin prescribing among female veterans with T2DM and/or CVD to reduce cardiovascular risk. Secondary outcomes included increased pharmacogenomic testing and the assessment of pharmacogenomic results related to statin therapy. This project was approved by the CVVAMC Pharmacy and Therapeutics Committee. The PCED was used to identify female veterans with T2DM and/or CVD without an active prescription for a statin between July and October 2023. A review of Computerized Patient Record System patient charts was completed to screen for prespecified inclusion and exclusion criteria. Veterans were included if they were assigned female at birth, were enrolled in care at CVVAMC, and had a diagnosis of T2DM or CVD (history of myocardial infarction, coronary bypass graft, percutaneous coronary intervention, or other revascularization in any setting).

Veterans were excluded if they were currently pregnant, trying to conceive, breastfeeding, had a T1DM diagnosis, had previously documented hypersensitivity to a statin, active liver failure or decompensated cirrhosis, previously documented statin-associated rhabdomyolysis or autoimmune myopathy, an active prescription for a proprotein convertase subtilisin/kexin type 9 inhibitor, or previously documented statin intolerance (defined as the inability to tolerate ≥ 3 statins, with ≥ 1 prescribed at low intensity or alternate-day dosing). The female veterans were compared to 2 comparators: the facility's male veterans and the VA national average, identified via the PCED.

Once a veteran was screened, they were telephoned between October 2023 and February 2024 and provided education on statin use and pharmacogenomic testing using a standardized note template. An order was placed for participants who provided verbal consent for pharmacogenomic testing. Those who agreed to statin initiation were referred to a clinical pharmacist practitioner (CPP) who contacted them at a later date to prescribe a statin following the recommendations of the 2019 ACC/AHA and 2023 ADA guidelines and pharmacogenomic testing, if applicable.^4,5,7 Appropriate monitoring and follow-up occurred at the discretion of each CPP. Data collection included: age, race, diagnoses (T2DM, CVD, or both), baseline lipid panel (total cholesterol, triglycerides, high-density lipoprotein, low-density lipoprotein), hepatic function, name and dose of statin, reasons for declining statin therapy, and pharmacogenomic testing results related to SLCO1B1.

Results

At baseline in July 2023, 77.8% of female veterans with T2DM were prescribed a statin, which exceeded the national VA average (77.0%), but was below the rate for male veterans (78.7%) in the facility comparator group.17 Additionally, 82.2% of females with CVD were prescribed a statin, which was below the national VA average of 86.0% and the 84.9% of male veterans in the facility comparator group.17 The PCED identified 189 female veterans from July 2023 to October 2023 who may benefit from statin therapy. Thirty-three females met the exclusion criteria. Of the 156 included veterans, 129 (82.7%) were successfully contacted and 27 (17.3%) could not be reached by telephone after 3 attempts (Figure 1). The 129 female veterans contacted had a mean age of 59 years and the majority were Black (82.9%) (Table 1).

Primary Outcomes

Of the 129 contacted veterans, 31 (24.0%) had a non-VA statin prescription, 13 (10.1%) had an active VA statin prescription, and 85 (65.9%) did not have a statin prescription, despite being eligible. Statin adherence was confirmed with participants, and the medication list was updated accordingly.

Of the 85 veterans with no active statin therapy, 37 (43.5%) accepted a new statin prescription and 48 (56.5%) declined. There were various reasons provided for declining statin therapy: 17 participants (35.4%) declined due to concern for AEs (Table 2).

From July 2023 to March 2024, the percentage of female veterans with active statin therapy with T2DM increased from 77.8% to 79.0%. For those with active statin therapy with CVD, usage increased from 82.2% to 90.2%, which exceeded the national VA average and facility male comparator group (Figures 2 and 3).¹⁷

Secondary Outcomes

Seventy-one of 129 veterans (55.0%) gave verbal consent, and 47 (66.2%) completed the pharmacogenomic testing; 58 (45.0%) declined. Five veterans (10.6%) had a known SLCO1B1 allele variant present. One veteran required a change in statin therapy based on the results (eAppendix).

Discussion

This project aimed to increase statin prescribing among female veterans with T2DM and/or CVD to reduce cardiovascular risk and increase pharmacogenomic testing using the PCED and care managed by CPPs. The results of this quality improvement project illustrated that both metrics have improved at CVVAMC as a result of the intervention. The results in both metrics now exceed the PCED national VA average, and the CVD metric also exceeds that of the facility male comparator group. While there was only a 1.2% increase from July 2023 to March 2024 for patients with T2DM, there was an 8.0% increase for patients with CVD. Despite standardized education on statin use, more veterans declined therapy than accepted it, mostly due to concern for AEs. Recording the reasons for declining statin therapy offered valuable insight that can be used in additional discussions with veterans and clinicians.

Pharmacogenomics gives clinicians the unique opportunity to take a proactive approach to better predict drug responses, potentially allowing for less trial and error with medications, fewer AEs, greater trust in the clinician, and improved medication adherence. The CPPs incorporated pharmacogenomic testing into their practice, which led to identifying 5 SLCO1B1 gene abnormalities. The PCED served as a powerful tool for advancing equity-focused quality improvement initiatives on a local level and was crucial in prioritizing the detection of veterans potentially receiving suboptimal care.

Limitations

The nature of “cold calls” made it challenging to establish contact for inclusion in this study. An alternative to increase engagement could have been scheduled phone or face-to-face visits. While the use of the PCED was crucial, data did not account for statins listed in the non-VA medication list. All 31 patients with statins prescribed outside the VA had a start date added to provide the most accurate representation of the data moving forward.

Another limitation in this project was its small sample size and population. CVVAMC serves about 6200 female veterans, with roughly 63% identifying as Black. The preponderance of Black individuals (83%) in this project is typical for the female patient population at CVVAMC but may not reflect the demographics of other populations. Other limitations to this project consisted of scheduling conflicts. Appointments for laboratory draws at community-based outpatient clinics were subject to availability, which resulted in some delay in completion of pharmacogenomic testing.

Conclusions

CPPs can help reduce inequity in health care delivery. Increased incorporation of the PCED into regular practice within the VA is recommended to continue addressing sex disparities in statin use, diabetes control, blood pressure management, cancer screenings, and vaccination needs. CVVAMC plans to expand its use through another quality improvement project focused on reducing sex disparities in blood pressure management. Improving educational resources made available to veterans on the importance of statin therapy and potential to mitigate AEs through use of the VA PHASER program also would be helpful. This project successfully improved CVVAMC metrics for female veterans appropriately prescribed statin therapy and increased access to pharmacogenomic testing. Most importantly, it helped close the sex-based gap in CVD risk reduction care.

Cardiovascular disease (CVD) is the leading cause of death among women in the United States.¹ Most CVD is due to the buildup of plaque (ie, cholesterol, proteins, calcium, and inflammatory cells) in artery walls.² The plaque may lead to atherosclerotic cardiovascular disease (ASCVD), which includes coronary heart disease, cerebrovascular disease, peripheral artery disease, and aortic atherosclerotic disease.^2,3 Control and reduction of ASCVD risk factors, including high cholesterol levels, elevated blood pressure, insulin resistance, smoking, and a sedentary lifestyle, can contribute to a reduction in ASCVD morbidity and mortality.² People with type 2 diabetes mellitus (T2DM) have an increased prevalence of lipid abnormalities, contributing to their high risk of ASCVD.^4,5

The prescribing of statins (3-hydroxy-3-methyl-glutaryl-coenzmye A reductase inhibitors) is the cornerstone of lipid-lowering therapy and cardiovascular risk reduction for primary and secondary prevention of ASCVD.⁶ The American Diabetes Association (ADA) and American College of Cardiology/American Heart Association (ACC/AHA) recommend moderate- to high-intensity statins for primary prevention in patients with T2DM and high-intensity statins for secondary prevention in those with or without diabetes when not contraindicated.^4,5,7 Despite eligibility according to guideline recommendations, research predominantly shows that women are less likely to receive statin therapy; however, this trend is improving. ^[6,8-11] To explain the sex differences in statin use, Nanna et al found that there is a combination of women being offered statin therapy less frequently, declining therapy more frequently, and discontinuing treatment more frequently.¹¹ One possibility for discontinuing treatment could be statin-associated muscle symptoms (SAMS), which occur in about 10% of patients.¹² The incidence of adverse effects (AEs) may be related to the way statins are metabolized.

Pharmacogenomic testing is free for veterans through the US Department of Veterans Affairs (VA) PHASER program, which offers information and recommendations for a panel of 11 gene variants. The panel includes genes related to common medication classes such as anticoagulants, antiplatelets, proton pump inhibitors, nonsteroidal anti-inflammatory drugs, opioids, antidepressants, and statins. The VA PHASER panel includes the solute carrier organic anion transporter family member 1B1 (SLCO1B1) gene, which is predominantly expressed in the liver and facilitates the hepatic uptake of most statins.^13,14 A reduced function of SLCO1B1 can lead to higher statin levels, resulting in increased concentrations that may potentially cause SAMS.^13,14 Some alleles associated with reduced function include SLCO1B1*5, *15, *23, *31, and *46 to *49, whereas others are associated with increased function, such as SLCO1B1 *14 and *20 (Appendix).¹⁵ Supporting evidence shows the SLCO1B1*5 nucleotide polymorphism increases plasma levels of simvastatin and atorvastatin, affecting effectiveness or toxicity. ¹³ Females tend to have a lower body weight and higher percentage of body fat compared with males, which might lead to higher concentrations of lipophilic drugs, including atorvastatin and simvastatin, which may be exacerbated by decreased function of SLCO1B1*5.¹⁵ With pharmacogenomic testing, therapeutic recommendations can be made to improve the overall safety and efficacy of statins, thus improving adherence using a patient-specific approach.^14,15

Methods

Carl Vinson VA Medical Center (CVVAMC) serves about 42,000 veterans in Central and South Georgia, of which about 15% are female. Of the female veterans enrolled in care, 63% identify as Black, 27% White, and 1.5% as Asian, American Indian/Alaska Native, or Native Hawaiian/Other Pacific Islander. The 2020 Veterans Chartbook report showed that female veterans and minority racial and ethnic groups had worse access to health care and higher mortality rates than their male and non-Hispanic White counterparts.¹⁶

The Primary Care Equity Dashboard (PCED) was developed to engage the VA health care workforce in the process of identifying and addressing inequities in local patient populations.¹⁷ Using electronic quality measure data, the PCED provides Veterans Integrated Service Network-level and facility-level performance on several metrics.¹⁸ The PCED had not been previously used at the CVVAMC, and few publications or quality improvement projects regarding its use have been reported by the VA Office of Health Equity. PCED helped identify disparities when comparing female to male patients in the prescribing of statin therapy for patients with CVD and statin therapy for patients with T2DM.

VA PHASER pharmacogenomic analyses provided an opportunity to expand this quality improvement project. Sanford Health and the VA collaborated on the PHASER program to offer free genetic testing for veterans. The program launched in 2019 and expanded to various VA sites, including CVVAMC in March 2023. This program has been extended to December 31, 2025.

The primary objective of this quality improvement project was to increase statin prescribing among female veterans with T2DM and/or CVD to reduce cardiovascular risk. Secondary outcomes included increased pharmacogenomic testing and the assessment of pharmacogenomic results related to statin therapy. This project was approved by the CVVAMC Pharmacy and Therapeutics Committee. The PCED was used to identify female veterans with T2DM and/or CVD without an active prescription for a statin between July and October 2023. A review of Computerized Patient Record System patient charts was completed to screen for prespecified inclusion and exclusion criteria. Veterans were included if they were assigned female at birth, were enrolled in care at CVVAMC, and had a diagnosis of T2DM or CVD (history of myocardial infarction, coronary bypass graft, percutaneous coronary intervention, or other revascularization in any setting).

Veterans were excluded if they were currently pregnant, trying to conceive, breastfeeding, had a T1DM diagnosis, had previously documented hypersensitivity to a statin, active liver failure or decompensated cirrhosis, previously documented statin-associated rhabdomyolysis or autoimmune myopathy, an active prescription for a proprotein convertase subtilisin/kexin type 9 inhibitor, or previously documented statin intolerance (defined as the inability to tolerate ≥ 3 statins, with ≥ 1 prescribed at low intensity or alternate-day dosing). The female veterans were compared to 2 comparators: the facility's male veterans and the VA national average, identified via the PCED.

Once a veteran was screened, they were telephoned between October 2023 and February 2024 and provided education on statin use and pharmacogenomic testing using a standardized note template. An order was placed for participants who provided verbal consent for pharmacogenomic testing. Those who agreed to statin initiation were referred to a clinical pharmacist practitioner (CPP) who contacted them at a later date to prescribe a statin following the recommendations of the 2019 ACC/AHA and 2023 ADA guidelines and pharmacogenomic testing, if applicable.^4,5,7 Appropriate monitoring and follow-up occurred at the discretion of each CPP. Data collection included: age, race, diagnoses (T2DM, CVD, or both), baseline lipid panel (total cholesterol, triglycerides, high-density lipoprotein, low-density lipoprotein), hepatic function, name and dose of statin, reasons for declining statin therapy, and pharmacogenomic testing results related to SLCO1B1.

Results

At baseline in July 2023, 77.8% of female veterans with T2DM were prescribed a statin, which exceeded the national VA average (77.0%), but was below the rate for male veterans (78.7%) in the facility comparator group.17 Additionally, 82.2% of females with CVD were prescribed a statin, which was below the national VA average of 86.0% and the 84.9% of male veterans in the facility comparator group.17 The PCED identified 189 female veterans from July 2023 to October 2023 who may benefit from statin therapy. Thirty-three females met the exclusion criteria. Of the 156 included veterans, 129 (82.7%) were successfully contacted and 27 (17.3%) could not be reached by telephone after 3 attempts (Figure 1). The 129 female veterans contacted had a mean age of 59 years and the majority were Black (82.9%) (Table 1).

Primary Outcomes

Of the 129 contacted veterans, 31 (24.0%) had a non-VA statin prescription, 13 (10.1%) had an active VA statin prescription, and 85 (65.9%) did not have a statin prescription, despite being eligible. Statin adherence was confirmed with participants, and the medication list was updated accordingly.

Of the 85 veterans with no active statin therapy, 37 (43.5%) accepted a new statin prescription and 48 (56.5%) declined. There were various reasons provided for declining statin therapy: 17 participants (35.4%) declined due to concern for AEs (Table 2).

From July 2023 to March 2024, the percentage of female veterans with active statin therapy with T2DM increased from 77.8% to 79.0%. For those with active statin therapy with CVD, usage increased from 82.2% to 90.2%, which exceeded the national VA average and facility male comparator group (Figures 2 and 3).¹⁷

Secondary Outcomes

Seventy-one of 129 veterans (55.0%) gave verbal consent, and 47 (66.2%) completed the pharmacogenomic testing; 58 (45.0%) declined. Five veterans (10.6%) had a known SLCO1B1 allele variant present. One veteran required a change in statin therapy based on the results (eAppendix).

Discussion

This project aimed to increase statin prescribing among female veterans with T2DM and/or CVD to reduce cardiovascular risk and increase pharmacogenomic testing using the PCED and care managed by CPPs. The results of this quality improvement project illustrated that both metrics have improved at CVVAMC as a result of the intervention. The results in both metrics now exceed the PCED national VA average, and the CVD metric also exceeds that of the facility male comparator group. While there was only a 1.2% increase from July 2023 to March 2024 for patients with T2DM, there was an 8.0% increase for patients with CVD. Despite standardized education on statin use, more veterans declined therapy than accepted it, mostly due to concern for AEs. Recording the reasons for declining statin therapy offered valuable insight that can be used in additional discussions with veterans and clinicians.

Pharmacogenomics gives clinicians the unique opportunity to take a proactive approach to better predict drug responses, potentially allowing for less trial and error with medications, fewer AEs, greater trust in the clinician, and improved medication adherence. The CPPs incorporated pharmacogenomic testing into their practice, which led to identifying 5 SLCO1B1 gene abnormalities. The PCED served as a powerful tool for advancing equity-focused quality improvement initiatives on a local level and was crucial in prioritizing the detection of veterans potentially receiving suboptimal care.

Limitations

The nature of “cold calls” made it challenging to establish contact for inclusion in this study. An alternative to increase engagement could have been scheduled phone or face-to-face visits. While the use of the PCED was crucial, data did not account for statins listed in the non-VA medication list. All 31 patients with statins prescribed outside the VA had a start date added to provide the most accurate representation of the data moving forward.

Another limitation in this project was its small sample size and population. CVVAMC serves about 6200 female veterans, with roughly 63% identifying as Black. The preponderance of Black individuals (83%) in this project is typical for the female patient population at CVVAMC but may not reflect the demographics of other populations. Other limitations to this project consisted of scheduling conflicts. Appointments for laboratory draws at community-based outpatient clinics were subject to availability, which resulted in some delay in completion of pharmacogenomic testing.

Conclusions

CPPs can help reduce inequity in health care delivery. Increased incorporation of the PCED into regular practice within the VA is recommended to continue addressing sex disparities in statin use, diabetes control, blood pressure management, cancer screenings, and vaccination needs. CVVAMC plans to expand its use through another quality improvement project focused on reducing sex disparities in blood pressure management. Improving educational resources made available to veterans on the importance of statin therapy and potential to mitigate AEs through use of the VA PHASER program also would be helpful. This project successfully improved CVVAMC metrics for female veterans appropriately prescribed statin therapy and increased access to pharmacogenomic testing. Most importantly, it helped close the sex-based gap in CVD risk reduction care.

In 2020, the US Multi-Society Task Force (USMSTF) on Colorectal Cancer (CRC) increased the recommended colon polyp surveillance interval for 1 to 2 subcentimeter tubular adenomas from 5 to 10 years to 7 to 10 years.¹ This change was prompted by emerging research indicating that rates of CRC and advanced neoplasia among patients with a history of only 1 to 2 subcentimeter tubular adenomas are lower than initially estimated.^2,3 This extension provides an opportunity to increase endoscopy capacity and improve access to colonoscopies by retroactively applying the 2020 guidelines to surveillance interval recommendations made before their introduction. For example, based on the updated guidelines, patients previously recommended to undergo colon polyp surveillance colonoscopy 5 years after an index colonoscopy could extend their surveillance interval by 2 to 5 years. Increasing endoscopic capacity could address the growing demand for colonoscopies from new screening guidelines that reduced the age of initial CRC screening from 50 years to 45 years and the backlog of procedures due to COVID-19 restrictions.⁴

As part of a project to increase endoscopic capacity at the US Department of Veterans Affairs (VA) Pittsburgh Healthcare System (VAPHS), this study assessed the potential impact of retroactively applying the 2020 USMSTF polyp surveillance guidelines on endoscopic capacity. These results may be informative for other VA and private-sector health care systems seeking to identify strategies to improve endoscopy capacity.

Methods

VAPHS is an integrated health care system in the Veterans Health Administration (VHA) serving 85,000 patients across 8 health care institutions in Pennsylvania, Ohio, and West Virginia. VAPHS manages colorectal screening recommendations for patients receiving medical care in the health care system regardless of whether their prior colonoscopy was performed at VAPHS or external facilities. The VA maintains a national CRC screening and surveillance electronic medical record reminder that prompts health care practitioners to order colon polyp surveillance based on interval recommendations from the index colonoscopy. This study reviewed all patients from the VAPHS panel with a reminder to undergo colonoscopy for screening for CRC or surveillance of colon polyps within 12 months from September 1, 2022.

Among patients with a reminder, 3 investigators reviewed index colonoscopy and pathology reports to identify CRC risk category, colonoscopy indication, procedural quality, and recommended repeat colonoscopy interval. Per the USMSTF guidelines, patients with incomplete colonoscopy or pathology records, high-risk indications (ie, personal history of inflammatory bowel disease, personal history of CRC, or family history of CRC), or inadequate bowel preparation (Boston Bowel Preparation Score < 6) were excluded. Additionally, patients who had CRC screening or surveillance discontinued due to age or comorbidities, had completed a subsequent follow-up colonoscopy, or were deceased at the time of review were excluded.

Retroactive Interval Reclassification

Among eligible patients, this study compared the repeat colonoscopy interval recommended by the prior endoscopist with those from the 2020 USMSTF guidelines. In cases where the interval was documented as a range of years, the lower end was considered the recommendation. Similarly, the lower end of the range from the 2020 USMSTF guidelines was used for the reclassified surveillance interval. Years extended per patient were quantified relative to September 1, 2023 (ie, 1 year after the review date). For example, if the index colonoscopy was completed on September 1, 2016, the initial surveillance recommendation was 5 years, and the reclassified recommendation was 7 years, the interval extension beyond September 1, 2023, was 0 years.

Furthermore, because index surveillance recommendations are not always guideline concordant, the years extended per patient were calculated by harmonizing the index endoscopist’s recommendations with the guidelines at the time of the index colonoscopy.⁵ For example, if the index colonoscopy was completed on September 1, 2018, and the endoscopist recommended a 5-year follow-up for a patient with average risk for CRC, adequate bowel preparation, and no colorectal polyps, that patient is eligible to extend their colonoscopy to September 1, 2028, based on guideline recommendations at the time of index endoscopy recommending that the next colonoscopy occur in 10 years. In this analysis the 2012 USMSTF guidelines were applied to all index colonoscopies completed in 2021 or earlier to allow time for adoption of the 2020 guidelines. 

This project fulfilled a facility mandate to increase capacity to conduct endoscopic procedures. Institutional review board approval was not required by VAPHS policy relating to clinical operations projects. Approval for publication of clinical operations activity was obtained from the VAPHS facility director.

Results

Within 1 year of the September 1, 2022, review date, 637 patients receiving care at VAPHS had clinical reminders for an upcoming colonoscopy. Of these, 54 (8.4%) were already up to date or were deceased at the time of review. Of the 583 eligible patients, 96% were male, the median age was 74 years, the median index colonoscopy year was 2016, and 178 (30.5%) had an average-risk CRC screening indication at the index colonoscopy (Table).

Of the 583 patients due for colonoscopy, 331 (56.7%) had both colonoscopy and pathology reports available. The majority of those with incomplete records had the index colonoscopy completed outside VAPHS. Among these patients, 222 (67.0%) had adequate bowel preparation. Of those with adequate bowel preparation, 43 were not eligible for interval extension because of high-risk conditions and 13 were not eligible because there was no index surveillance interval recommendation from the index endoscopist. Of the patients due for colonoscopy, 166 (28.4%) were potentially eligible for surveillance interval extension (Figure).  

Sixty-five (39.2%) of the 166 patients had 1 to 2 subcentimeter tubular adenomas on their index colonoscopy. Sixty-two patients were eligible for interval extension to 7 years, but this only resulted in ≥ 1 year of extension beyond the review date for 36 (6% of all 583 patients due for colonoscopy). The 36 patients were extended 63 years. By harmonizing the index endoscopists’ surveillance interval recommendation with the guideline at the time of the index colonoscopy, 29 additional patients could have their colonoscopy extended by ≥ 1 year. Harmonization extended colonoscopy intervals by 93 years. Retroactively applying the 2020 USMSTF polyp surveillance guidelines and harmonizing recommendations to guidelines extended the time of index colonoscopy by 153 years.

Discussion

With retroactive application of the 2020 USMSTF polyp surveillance guidelines, 6% of patients due for an upcoming colonoscopy could extend their follow-up by ≥ 1 year by extending the surveillance interval for 1 to 2 subcentimeter tubular adenomas to 7 years. An additional 5% of patients could extend their interval by harmonizing the index endoscopist’s interval recommendation with polyp surveillance guidelines at the time of the index colonoscopy. These findings are consistent with the results of 2 studies that demonstrated that about 14% of patients due for colonoscopy could have their interval extended.^6,7 The current study enhances those insights by separating the contribution of 2020 USMSTF polyp surveillance guidelines from the contribution of harmonizing surveillance intervals with guidelines for other polyp histologies. This study found that there is an opportunity to improve endoscopic capacity by harmonizing recommendations with guidelines. This complements a 2023 study showing that even when knowledgeable about guidelines, clinicians do not necessarily follow recommendations.⁸ While this and previous research have identified that 11% to 14% of patients are eligible for extension, these individuals would also have to be willing to have their polyp surveillance intervals extended for there to be a real-world impact on endoscopic capacity. A 2024 study found that only 19% to 37% of patients with 1 to 2 small tubular adenomas were willing to have polyps surveillance interval extension.⁹ This suggests the actual effect on capacity may be even lower than reported.

Limitations

The overall impact of the 2020 USMSTF polyp surveillance guidelines on endoscopic capacity was blunted by the high prevalence of incomplete index colonoscopy records among the study population. Without data on bowel preparation quality or procedure indications, this study could not assess whether 43% of patients were eligible for surveillance interval extension. Most index colonoscopies with incomplete documentation were completed at community-care gastroenterology facilities. This high rate of incomplete documentation is likely generalizable to other VA health care systems—especially in the era of the Veterans Access, Choice, and Accountability Act of 2014, which increased veteran access to non-VA community care.¹⁰ Veterans due for colon polyp surveillance colonoscopies are more likely to have had their prior colonoscopy in community care compared with prior eras.¹¹ Furthermore, because the VHA is among the most established integrated health care systems offering primary and subspecialty care in the US, private sector health care systems may have even greater rates of care fragmentation for longitudinal CRC screening and colon polyp surveillance, as these systems have only begun to regionally integrate recently.^12,13

Another limitation is that nearly one-third of the individuals with documentation had inadequate bowel preparation for surveillance recommendations. This results in shorter surveillance follow-up colonoscopies and increases downstream demand for future colonoscopies. The low yield of extending colon polyp surveillance interval in this study emphasizes that improved efforts to obtain colonoscopy and pathology reports from community care, right-sizing the colon polyp surveillance intervals recommended by endoscopists, and improving quality of bowel preparation could have downstream health care system benefits in the future. These efforts could increase colonoscopy capacity at VA health care systems, thereby shortening colonoscopy wait times, decreasing fragmentation of care, and increasing the number of veterans who receive high-quality colonoscopies at VA health care systems.¹⁴

Conclusions

Eleven percent of patients in this study due for a colonoscopy could extend their follow-up by ≥ 1 year. About half of these extensions were directly due to the 2020 USMSTF polyp surveillance interval extension for 1 to 2 subcentimeter tubular adenomas. The rest resulted from harmonizing recommendations with guidelines at the time of the procedure. To determine whether retroactively applying polyp surveillance guidelines to follow-up interval recommendations will result in improved endoscopic capacity, health care system administrators should consider the degree of CRC screening care fragmentation in their patient population. Greater long-term gains in endoscopic capacity may be achieved by proactively supporting endoscopists in making guideline-concordant screening recommendations at the time of colonoscopy.

In 2020, the US Multi-Society Task Force (USMSTF) on Colorectal Cancer (CRC) increased the recommended colon polyp surveillance interval for 1 to 2 subcentimeter tubular adenomas from 5 to 10 years to 7 to 10 years.¹ This change was prompted by emerging research indicating that rates of CRC and advanced neoplasia among patients with a history of only 1 to 2 subcentimeter tubular adenomas are lower than initially estimated.^2,3 This extension provides an opportunity to increase endoscopy capacity and improve access to colonoscopies by retroactively applying the 2020 guidelines to surveillance interval recommendations made before their introduction. For example, based on the updated guidelines, patients previously recommended to undergo colon polyp surveillance colonoscopy 5 years after an index colonoscopy could extend their surveillance interval by 2 to 5 years. Increasing endoscopic capacity could address the growing demand for colonoscopies from new screening guidelines that reduced the age of initial CRC screening from 50 years to 45 years and the backlog of procedures due to COVID-19 restrictions.⁴

As part of a project to increase endoscopic capacity at the US Department of Veterans Affairs (VA) Pittsburgh Healthcare System (VAPHS), this study assessed the potential impact of retroactively applying the 2020 USMSTF polyp surveillance guidelines on endoscopic capacity. These results may be informative for other VA and private-sector health care systems seeking to identify strategies to improve endoscopy capacity.

Methods

VAPHS is an integrated health care system in the Veterans Health Administration (VHA) serving 85,000 patients across 8 health care institutions in Pennsylvania, Ohio, and West Virginia. VAPHS manages colorectal screening recommendations for patients receiving medical care in the health care system regardless of whether their prior colonoscopy was performed at VAPHS or external facilities. The VA maintains a national CRC screening and surveillance electronic medical record reminder that prompts health care practitioners to order colon polyp surveillance based on interval recommendations from the index colonoscopy. This study reviewed all patients from the VAPHS panel with a reminder to undergo colonoscopy for screening for CRC or surveillance of colon polyps within 12 months from September 1, 2022.

Among patients with a reminder, 3 investigators reviewed index colonoscopy and pathology reports to identify CRC risk category, colonoscopy indication, procedural quality, and recommended repeat colonoscopy interval. Per the USMSTF guidelines, patients with incomplete colonoscopy or pathology records, high-risk indications (ie, personal history of inflammatory bowel disease, personal history of CRC, or family history of CRC), or inadequate bowel preparation (Boston Bowel Preparation Score < 6) were excluded. Additionally, patients who had CRC screening or surveillance discontinued due to age or comorbidities, had completed a subsequent follow-up colonoscopy, or were deceased at the time of review were excluded.

Retroactive Interval Reclassification

Among eligible patients, this study compared the repeat colonoscopy interval recommended by the prior endoscopist with those from the 2020 USMSTF guidelines. In cases where the interval was documented as a range of years, the lower end was considered the recommendation. Similarly, the lower end of the range from the 2020 USMSTF guidelines was used for the reclassified surveillance interval. Years extended per patient were quantified relative to September 1, 2023 (ie, 1 year after the review date). For example, if the index colonoscopy was completed on September 1, 2016, the initial surveillance recommendation was 5 years, and the reclassified recommendation was 7 years, the interval extension beyond September 1, 2023, was 0 years.

Furthermore, because index surveillance recommendations are not always guideline concordant, the years extended per patient were calculated by harmonizing the index endoscopist’s recommendations with the guidelines at the time of the index colonoscopy.⁵ For example, if the index colonoscopy was completed on September 1, 2018, and the endoscopist recommended a 5-year follow-up for a patient with average risk for CRC, adequate bowel preparation, and no colorectal polyps, that patient is eligible to extend their colonoscopy to September 1, 2028, based on guideline recommendations at the time of index endoscopy recommending that the next colonoscopy occur in 10 years. In this analysis the 2012 USMSTF guidelines were applied to all index colonoscopies completed in 2021 or earlier to allow time for adoption of the 2020 guidelines. 

This project fulfilled a facility mandate to increase capacity to conduct endoscopic procedures. Institutional review board approval was not required by VAPHS policy relating to clinical operations projects. Approval for publication of clinical operations activity was obtained from the VAPHS facility director.

Results

Within 1 year of the September 1, 2022, review date, 637 patients receiving care at VAPHS had clinical reminders for an upcoming colonoscopy. Of these, 54 (8.4%) were already up to date or were deceased at the time of review. Of the 583 eligible patients, 96% were male, the median age was 74 years, the median index colonoscopy year was 2016, and 178 (30.5%) had an average-risk CRC screening indication at the index colonoscopy (Table).

Of the 583 patients due for colonoscopy, 331 (56.7%) had both colonoscopy and pathology reports available. The majority of those with incomplete records had the index colonoscopy completed outside VAPHS. Among these patients, 222 (67.0%) had adequate bowel preparation. Of those with adequate bowel preparation, 43 were not eligible for interval extension because of high-risk conditions and 13 were not eligible because there was no index surveillance interval recommendation from the index endoscopist. Of the patients due for colonoscopy, 166 (28.4%) were potentially eligible for surveillance interval extension (Figure).  

Sixty-five (39.2%) of the 166 patients had 1 to 2 subcentimeter tubular adenomas on their index colonoscopy. Sixty-two patients were eligible for interval extension to 7 years, but this only resulted in ≥ 1 year of extension beyond the review date for 36 (6% of all 583 patients due for colonoscopy). The 36 patients were extended 63 years. By harmonizing the index endoscopists’ surveillance interval recommendation with the guideline at the time of the index colonoscopy, 29 additional patients could have their colonoscopy extended by ≥ 1 year. Harmonization extended colonoscopy intervals by 93 years. Retroactively applying the 2020 USMSTF polyp surveillance guidelines and harmonizing recommendations to guidelines extended the time of index colonoscopy by 153 years.

Discussion

With retroactive application of the 2020 USMSTF polyp surveillance guidelines, 6% of patients due for an upcoming colonoscopy could extend their follow-up by ≥ 1 year by extending the surveillance interval for 1 to 2 subcentimeter tubular adenomas to 7 years. An additional 5% of patients could extend their interval by harmonizing the index endoscopist’s interval recommendation with polyp surveillance guidelines at the time of the index colonoscopy. These findings are consistent with the results of 2 studies that demonstrated that about 14% of patients due for colonoscopy could have their interval extended.^6,7 The current study enhances those insights by separating the contribution of 2020 USMSTF polyp surveillance guidelines from the contribution of harmonizing surveillance intervals with guidelines for other polyp histologies. This study found that there is an opportunity to improve endoscopic capacity by harmonizing recommendations with guidelines. This complements a 2023 study showing that even when knowledgeable about guidelines, clinicians do not necessarily follow recommendations.⁸ While this and previous research have identified that 11% to 14% of patients are eligible for extension, these individuals would also have to be willing to have their polyp surveillance intervals extended for there to be a real-world impact on endoscopic capacity. A 2024 study found that only 19% to 37% of patients with 1 to 2 small tubular adenomas were willing to have polyps surveillance interval extension.⁹ This suggests the actual effect on capacity may be even lower than reported.

Limitations

The overall impact of the 2020 USMSTF polyp surveillance guidelines on endoscopic capacity was blunted by the high prevalence of incomplete index colonoscopy records among the study population. Without data on bowel preparation quality or procedure indications, this study could not assess whether 43% of patients were eligible for surveillance interval extension. Most index colonoscopies with incomplete documentation were completed at community-care gastroenterology facilities. This high rate of incomplete documentation is likely generalizable to other VA health care systems—especially in the era of the Veterans Access, Choice, and Accountability Act of 2014, which increased veteran access to non-VA community care.¹⁰ Veterans due for colon polyp surveillance colonoscopies are more likely to have had their prior colonoscopy in community care compared with prior eras.¹¹ Furthermore, because the VHA is among the most established integrated health care systems offering primary and subspecialty care in the US, private sector health care systems may have even greater rates of care fragmentation for longitudinal CRC screening and colon polyp surveillance, as these systems have only begun to regionally integrate recently.^12,13

Another limitation is that nearly one-third of the individuals with documentation had inadequate bowel preparation for surveillance recommendations. This results in shorter surveillance follow-up colonoscopies and increases downstream demand for future colonoscopies. The low yield of extending colon polyp surveillance interval in this study emphasizes that improved efforts to obtain colonoscopy and pathology reports from community care, right-sizing the colon polyp surveillance intervals recommended by endoscopists, and improving quality of bowel preparation could have downstream health care system benefits in the future. These efforts could increase colonoscopy capacity at VA health care systems, thereby shortening colonoscopy wait times, decreasing fragmentation of care, and increasing the number of veterans who receive high-quality colonoscopies at VA health care systems.¹⁴

Conclusions

Eleven percent of patients in this study due for a colonoscopy could extend their follow-up by ≥ 1 year. About half of these extensions were directly due to the 2020 USMSTF polyp surveillance interval extension for 1 to 2 subcentimeter tubular adenomas. The rest resulted from harmonizing recommendations with guidelines at the time of the procedure. To determine whether retroactively applying polyp surveillance guidelines to follow-up interval recommendations will result in improved endoscopic capacity, health care system administrators should consider the degree of CRC screening care fragmentation in their patient population. Greater long-term gains in endoscopic capacity may be achieved by proactively supporting endoscopists in making guideline-concordant screening recommendations at the time of colonoscopy.

In 2020, the US Multi-Society Task Force (USMSTF) on Colorectal Cancer (CRC) increased the recommended colon polyp surveillance interval for 1 to 2 subcentimeter tubular adenomas from 5 to 10 years to 7 to 10 years.¹ This change was prompted by emerging research indicating that rates of CRC and advanced neoplasia among patients with a history of only 1 to 2 subcentimeter tubular adenomas are lower than initially estimated.^2,3 This extension provides an opportunity to increase endoscopy capacity and improve access to colonoscopies by retroactively applying the 2020 guidelines to surveillance interval recommendations made before their introduction. For example, based on the updated guidelines, patients previously recommended to undergo colon polyp surveillance colonoscopy 5 years after an index colonoscopy could extend their surveillance interval by 2 to 5 years. Increasing endoscopic capacity could address the growing demand for colonoscopies from new screening guidelines that reduced the age of initial CRC screening from 50 years to 45 years and the backlog of procedures due to COVID-19 restrictions.⁴

As part of a project to increase endoscopic capacity at the US Department of Veterans Affairs (VA) Pittsburgh Healthcare System (VAPHS), this study assessed the potential impact of retroactively applying the 2020 USMSTF polyp surveillance guidelines on endoscopic capacity. These results may be informative for other VA and private-sector health care systems seeking to identify strategies to improve endoscopy capacity.

Methods

VAPHS is an integrated health care system in the Veterans Health Administration (VHA) serving 85,000 patients across 8 health care institutions in Pennsylvania, Ohio, and West Virginia. VAPHS manages colorectal screening recommendations for patients receiving medical care in the health care system regardless of whether their prior colonoscopy was performed at VAPHS or external facilities. The VA maintains a national CRC screening and surveillance electronic medical record reminder that prompts health care practitioners to order colon polyp surveillance based on interval recommendations from the index colonoscopy. This study reviewed all patients from the VAPHS panel with a reminder to undergo colonoscopy for screening for CRC or surveillance of colon polyps within 12 months from September 1, 2022.

Among patients with a reminder, 3 investigators reviewed index colonoscopy and pathology reports to identify CRC risk category, colonoscopy indication, procedural quality, and recommended repeat colonoscopy interval. Per the USMSTF guidelines, patients with incomplete colonoscopy or pathology records, high-risk indications (ie, personal history of inflammatory bowel disease, personal history of CRC, or family history of CRC), or inadequate bowel preparation (Boston Bowel Preparation Score < 6) were excluded. Additionally, patients who had CRC screening or surveillance discontinued due to age or comorbidities, had completed a subsequent follow-up colonoscopy, or were deceased at the time of review were excluded.

Retroactive Interval Reclassification

Among eligible patients, this study compared the repeat colonoscopy interval recommended by the prior endoscopist with those from the 2020 USMSTF guidelines. In cases where the interval was documented as a range of years, the lower end was considered the recommendation. Similarly, the lower end of the range from the 2020 USMSTF guidelines was used for the reclassified surveillance interval. Years extended per patient were quantified relative to September 1, 2023 (ie, 1 year after the review date). For example, if the index colonoscopy was completed on September 1, 2016, the initial surveillance recommendation was 5 years, and the reclassified recommendation was 7 years, the interval extension beyond September 1, 2023, was 0 years.

Furthermore, because index surveillance recommendations are not always guideline concordant, the years extended per patient were calculated by harmonizing the index endoscopist’s recommendations with the guidelines at the time of the index colonoscopy.⁵ For example, if the index colonoscopy was completed on September 1, 2018, and the endoscopist recommended a 5-year follow-up for a patient with average risk for CRC, adequate bowel preparation, and no colorectal polyps, that patient is eligible to extend their colonoscopy to September 1, 2028, based on guideline recommendations at the time of index endoscopy recommending that the next colonoscopy occur in 10 years. In this analysis the 2012 USMSTF guidelines were applied to all index colonoscopies completed in 2021 or earlier to allow time for adoption of the 2020 guidelines. 

This project fulfilled a facility mandate to increase capacity to conduct endoscopic procedures. Institutional review board approval was not required by VAPHS policy relating to clinical operations projects. Approval for publication of clinical operations activity was obtained from the VAPHS facility director.

Results

Within 1 year of the September 1, 2022, review date, 637 patients receiving care at VAPHS had clinical reminders for an upcoming colonoscopy. Of these, 54 (8.4%) were already up to date or were deceased at the time of review. Of the 583 eligible patients, 96% were male, the median age was 74 years, the median index colonoscopy year was 2016, and 178 (30.5%) had an average-risk CRC screening indication at the index colonoscopy (Table).

Of the 583 patients due for colonoscopy, 331 (56.7%) had both colonoscopy and pathology reports available. The majority of those with incomplete records had the index colonoscopy completed outside VAPHS. Among these patients, 222 (67.0%) had adequate bowel preparation. Of those with adequate bowel preparation, 43 were not eligible for interval extension because of high-risk conditions and 13 were not eligible because there was no index surveillance interval recommendation from the index endoscopist. Of the patients due for colonoscopy, 166 (28.4%) were potentially eligible for surveillance interval extension (Figure).  

Sixty-five (39.2%) of the 166 patients had 1 to 2 subcentimeter tubular adenomas on their index colonoscopy. Sixty-two patients were eligible for interval extension to 7 years, but this only resulted in ≥ 1 year of extension beyond the review date for 36 (6% of all 583 patients due for colonoscopy). The 36 patients were extended 63 years. By harmonizing the index endoscopists’ surveillance interval recommendation with the guideline at the time of the index colonoscopy, 29 additional patients could have their colonoscopy extended by ≥ 1 year. Harmonization extended colonoscopy intervals by 93 years. Retroactively applying the 2020 USMSTF polyp surveillance guidelines and harmonizing recommendations to guidelines extended the time of index colonoscopy by 153 years.

Discussion

With retroactive application of the 2020 USMSTF polyp surveillance guidelines, 6% of patients due for an upcoming colonoscopy could extend their follow-up by ≥ 1 year by extending the surveillance interval for 1 to 2 subcentimeter tubular adenomas to 7 years. An additional 5% of patients could extend their interval by harmonizing the index endoscopist’s interval recommendation with polyp surveillance guidelines at the time of the index colonoscopy. These findings are consistent with the results of 2 studies that demonstrated that about 14% of patients due for colonoscopy could have their interval extended.^6,7 The current study enhances those insights by separating the contribution of 2020 USMSTF polyp surveillance guidelines from the contribution of harmonizing surveillance intervals with guidelines for other polyp histologies. This study found that there is an opportunity to improve endoscopic capacity by harmonizing recommendations with guidelines. This complements a 2023 study showing that even when knowledgeable about guidelines, clinicians do not necessarily follow recommendations.⁸ While this and previous research have identified that 11% to 14% of patients are eligible for extension, these individuals would also have to be willing to have their polyp surveillance intervals extended for there to be a real-world impact on endoscopic capacity. A 2024 study found that only 19% to 37% of patients with 1 to 2 small tubular adenomas were willing to have polyps surveillance interval extension.⁹ This suggests the actual effect on capacity may be even lower than reported.

Limitations

The overall impact of the 2020 USMSTF polyp surveillance guidelines on endoscopic capacity was blunted by the high prevalence of incomplete index colonoscopy records among the study population. Without data on bowel preparation quality or procedure indications, this study could not assess whether 43% of patients were eligible for surveillance interval extension. Most index colonoscopies with incomplete documentation were completed at community-care gastroenterology facilities. This high rate of incomplete documentation is likely generalizable to other VA health care systems—especially in the era of the Veterans Access, Choice, and Accountability Act of 2014, which increased veteran access to non-VA community care.¹⁰ Veterans due for colon polyp surveillance colonoscopies are more likely to have had their prior colonoscopy in community care compared with prior eras.¹¹ Furthermore, because the VHA is among the most established integrated health care systems offering primary and subspecialty care in the US, private sector health care systems may have even greater rates of care fragmentation for longitudinal CRC screening and colon polyp surveillance, as these systems have only begun to regionally integrate recently.^12,13

Another limitation is that nearly one-third of the individuals with documentation had inadequate bowel preparation for surveillance recommendations. This results in shorter surveillance follow-up colonoscopies and increases downstream demand for future colonoscopies. The low yield of extending colon polyp surveillance interval in this study emphasizes that improved efforts to obtain colonoscopy and pathology reports from community care, right-sizing the colon polyp surveillance intervals recommended by endoscopists, and improving quality of bowel preparation could have downstream health care system benefits in the future. These efforts could increase colonoscopy capacity at VA health care systems, thereby shortening colonoscopy wait times, decreasing fragmentation of care, and increasing the number of veterans who receive high-quality colonoscopies at VA health care systems.¹⁴

Conclusions

Eleven percent of patients in this study due for a colonoscopy could extend their follow-up by ≥ 1 year. About half of these extensions were directly due to the 2020 USMSTF polyp surveillance interval extension for 1 to 2 subcentimeter tubular adenomas. The rest resulted from harmonizing recommendations with guidelines at the time of the procedure. To determine whether retroactively applying polyp surveillance guidelines to follow-up interval recommendations will result in improved endoscopic capacity, health care system administrators should consider the degree of CRC screening care fragmentation in their patient population. Greater long-term gains in endoscopic capacity may be achieved by proactively supporting endoscopists in making guideline-concordant screening recommendations at the time of colonoscopy.

Intravascular large B-cell lymphoma (IVBCL) is an exceedingly rare aggressive form of non-Hodgkin lymphoma with tumor cells growing selectively in vascular lumina.¹ The annual incidence of IVBCL is fewer than 0.5 cases per 1,000,000 individuals worldwide.² Only about 500 known cases of IVBCL have been recorded in the literature,³ and it accounts for less than 1% of all lymphomas. It generally affects middle-aged to elderly individuals, with an average age at diagnosis of 70 years.² It has a predilection for men and commonly develops in individuals who are immunosuppressed.^3,4

Multiple variants of IVBCL have been described in the literature, with central nervous system and cutaneous involvement being the most classic findings.⁵ Bone marrow involvement with hepatosplenomegaly also has been noted in the literature.^6,7 Diagnosis of IVBCL and its variants requires a high index of suspicion, as the clinical manifestations and tissues involved typically are nonspecific and highly variable. Even in the classic variant of IVBCL, skin involvement is only reported in approximately half of cases.³ When present, cutaneous manifestations can range from nodules and violaceous plaques to induration and telangiectasias.³ Lymphadenopathy and lymphoma (leukemic) cells are not seen on a peripheral blood smear.^2,8,9

The lack of lymphadenopathy or identifiable leukemic cells in the peripheral blood presents a diagnostic dilemma, as sufficient information for accurate diagnosis must be obtained while minimizing invasive procedures and resource expenditure. Because IVBCL cells can reside in the vascular lumina of various organs, numerous biopsy sites have been proposed for diagnosis of lymphoma, including the bone marrow, skin, prostate, adrenal gland, brain, liver, and kidneys.¹⁰ While some studies have reported that the optimal diagnostic site is the bone marrow, skin biopsies are more routinely carried out, as they represent a convenient and cost-effective alternative to other more invasive techniques.^6,7,10 Studies have shown biopsy sensitivity values ranging from 77.8% to 83.3% for detection of IVBCL in normal-appearing skin, which is comparable to the sensitivities of a bone marrow biopsy.^7,8 Although skin biopsy of random sites has shown diagnostic efficacy, some studies have proposed that biopsies taken from hemangiomas and other hypervascular lesions can further improve diagnostic yield, as lymphoma cells often are present in capillaries of subcutaneous adipose tissue.^6,10,11 However, no obvious clinicopathologic differences were observed between IVBCL with and without involvement of a cutaneous hemangioma.¹¹

The purpose of this study was to determine the diagnostic efficacy of skin biopsies for detecting IVBCL at various body sites and to establish whether biopsies from hemangiomas yield higher diagnostic value.

Methods

A 66-year-old man recently died at our institution secondary to IVBCL. His disease course was characterized by multiple hospital admissions in a 6-month period for fever of unknown origin and tachycardia unresponsive to broad-spectrum antibiotics and systemic steroids. The patient declined over the course of 3 to 4 weeks with findings suggestive of lymphoma and tumor lysis syndrome, and he eventually developed shock, hypoxic respiratory failure, and acute renal failure. As imaging studies and examinations had not shown lymphadenopathy, bone marrow biopsy was performed, and dermatology was consulted to perform skin biopsies to evaluate for IVBCL. Both bone marrow biopsies and random skin biopsies from the abdomen showed large and atypical CD20+ B cells within select vascular lumina (Figure). No extravascular lymphoma cells were seen. Based on the bone marrow and skin biopsies, a diagnosis of IVBCL was made. Unfortunately, no progress was made clinically, and the patient was transitioned to comfort measures. Upon the patient’s death, his family expressed interest in participating in IVBCL research and agreed to a limited autopsy consisting of numerous skin biopsies to evaluate different body sites and biopsy types (normal skin vs hemangiomas) to ascertain whether diagnostic yield could be increased by performing selective biopsies of hemangiomas if IVBCL was suspected.

Twenty-four postmortem 4-mm punch biopsies containing subcutaneous adipose tissue were taken within 24 hours of the death of the patient before embalming. The biopsies were taken from all regions of the body except the head and neck for cosmetic preservation of the decedent. Eighteen of the biopsies were taken from random sites of normal-appearing skin; the remaining 6 were taken from clinically identifiable cherry hemangiomas (5 on the trunk and 1 on the thigh). There was a variable degree of livor mortis in the dependent areas of the body, which was included in the random biopsies from the back to ensure any pooling of dependent blood would not alter the findings.

A histopathologic examination by a board-certified dermatopathologist (M.P.) on a single hematoxylin-eosin–stained level was performed to evaluate each biopsy for superficial involvement and deep involvement by IVBCL. Superficial involvement was defined as dermal involvement at or above the level of the eccrine sweat glands; deep involvement was defined as dermal involvement beneath the eccrine sweat glands and all subcutaneous fat present. Skin and bone marrow biopsies used to make the original diagnosis prior to the patient’s death were reviewed, including CD20 immunohistochemistry for morphologic comparison to the study slides. Involvement was graded as 0 to 3+ (eTable).

Results

Results from all 24 biopsies are shown in the eTable. Twenty-two (91.7%) biopsies showed at least focal involvement by IVBCL. Nine (37.5%) biopsies showed more deep vs superficial involvement of the same site. On average, the 6 biopsies from clinically detected hemangiomas showed more involvement by IVBCL than the random biopsies (eFigures 1 and 2A). The superficial involvement of skin with a hemangioma showed an average score of 2.33 v 0.78 when compared with the superficial aspect of the random biopsies; the deep involvement of skin with a hemangioma showed an average score of 2.67 vs 1.16 when compared with the deep aspect of the random biopsies (eFigure 2B).

Comment

Intravascular large B-cell lymphoma is an aggressive malignancy that traditionally is difficult to diagnose. Many efforts have been made to improve detection and early diagnosis. As cutaneous involvement is common and sometimes the only sign of disease, dermatologists may be called upon to evaluate and biopsy patients with this suspected diagnosis. The purpose of our study was to improve diagnostic efficacy by methodically performing numerous biopsies and assessing the level of involvement of the superficial and deep skin as well as involvement of hemangiomas. The goal of this meticulous approach was to identify the highest-yield areas for biopsy with minimal impact on the patient. Our results showed that random skin biopsies are an effective way to identify IVBCL. Twenty-two (91.7%) biopsies contained at least focal lymphoma cells. Although the 2 biopsies that showed no tumor cells at all happened to both be from the left arm, this is believed to be coincidental. No discernable pattern was identified regarding involvement and anatomic region. Even though 20 (83.3%) biopsies showed superficial involvement, deep biopsy is essential, as 9 (37.5%) biopsies showed increased deep involvement compared to superficial involvement. Therefore, a deep punch biopsy is essential for maximum sensitivity.

Hemangiomas provide a potential target that could increase the sensitivity of a biopsy in the absence of clinical findings, when the disease in question is exclusively intravascular. The data gathered in this study support this idea, as biopsies from hemangiomas showed increased involvement compared to random biopsies, both superficially and deep (2.33 vs 0.78 and 2.67 vs 1.16, respectively). Interestingly, the hemangioma biopsy sites showed increased deep and superficial involvement, despite these typical cherry hemangiomas only involving the superficial dermis. One possible explanation for this is that the hemangiomas have larger-caliber feeder vessels with increased blood flow beneath them. It would then follow that this increased vasculature would increase the chances of identifying intravascular lymphoma cells. This finding further accentuates the need for a deep punch biopsy containing subcutaneous fat. 

Completing the study in the setting of an autopsy provided the advantage of being able to take numerous biopsies without increased harm to the patient. This extensive set of biopsies would not be reasonable to complete on a living patient. This study also has limitations. Although this patient did fall within the typical demographics for patients with IVBCL, the data were still limited to 1 patient. This autopsy format (on a patient whose cause of death was indeed IVBCL) also implies terminal disease, which may mean the patient had a larger disease burden than a living patient who would typically be biopsied. Although this increased disease burden may have increased the sensitivity of finding IVBCL in the biopsies of this study, this further emphasizes the importance of trying to determine any factors that could increase sensitivity in a living patient with a lower disease burden.

Conclusion

Skin biopsies can provide a sensitive, low-cost, and low-morbidity method to assess a patient for IVBCL. Though random skin biopsies can yield valuable information, deep, 4-mm punch biopsies of clinically identifiable hemangiomas may provide the highest sensitivity for IVBCL.

Intravascular large B-cell lymphoma (IVBCL) is an exceedingly rare aggressive form of non-Hodgkin lymphoma with tumor cells growing selectively in vascular lumina.¹ The annual incidence of IVBCL is fewer than 0.5 cases per 1,000,000 individuals worldwide.² Only about 500 known cases of IVBCL have been recorded in the literature,³ and it accounts for less than 1% of all lymphomas. It generally affects middle-aged to elderly individuals, with an average age at diagnosis of 70 years.² It has a predilection for men and commonly develops in individuals who are immunosuppressed.^3,4

Multiple variants of IVBCL have been described in the literature, with central nervous system and cutaneous involvement being the most classic findings.⁵ Bone marrow involvement with hepatosplenomegaly also has been noted in the literature.^6,7 Diagnosis of IVBCL and its variants requires a high index of suspicion, as the clinical manifestations and tissues involved typically are nonspecific and highly variable. Even in the classic variant of IVBCL, skin involvement is only reported in approximately half of cases.³ When present, cutaneous manifestations can range from nodules and violaceous plaques to induration and telangiectasias.³ Lymphadenopathy and lymphoma (leukemic) cells are not seen on a peripheral blood smear.^2,8,9

The lack of lymphadenopathy or identifiable leukemic cells in the peripheral blood presents a diagnostic dilemma, as sufficient information for accurate diagnosis must be obtained while minimizing invasive procedures and resource expenditure. Because IVBCL cells can reside in the vascular lumina of various organs, numerous biopsy sites have been proposed for diagnosis of lymphoma, including the bone marrow, skin, prostate, adrenal gland, brain, liver, and kidneys.¹⁰ While some studies have reported that the optimal diagnostic site is the bone marrow, skin biopsies are more routinely carried out, as they represent a convenient and cost-effective alternative to other more invasive techniques.^6,7,10 Studies have shown biopsy sensitivity values ranging from 77.8% to 83.3% for detection of IVBCL in normal-appearing skin, which is comparable to the sensitivities of a bone marrow biopsy.^7,8 Although skin biopsy of random sites has shown diagnostic efficacy, some studies have proposed that biopsies taken from hemangiomas and other hypervascular lesions can further improve diagnostic yield, as lymphoma cells often are present in capillaries of subcutaneous adipose tissue.^6,10,11 However, no obvious clinicopathologic differences were observed between IVBCL with and without involvement of a cutaneous hemangioma.¹¹

The purpose of this study was to determine the diagnostic efficacy of skin biopsies for detecting IVBCL at various body sites and to establish whether biopsies from hemangiomas yield higher diagnostic value.

Methods

A 66-year-old man recently died at our institution secondary to IVBCL. His disease course was characterized by multiple hospital admissions in a 6-month period for fever of unknown origin and tachycardia unresponsive to broad-spectrum antibiotics and systemic steroids. The patient declined over the course of 3 to 4 weeks with findings suggestive of lymphoma and tumor lysis syndrome, and he eventually developed shock, hypoxic respiratory failure, and acute renal failure. As imaging studies and examinations had not shown lymphadenopathy, bone marrow biopsy was performed, and dermatology was consulted to perform skin biopsies to evaluate for IVBCL. Both bone marrow biopsies and random skin biopsies from the abdomen showed large and atypical CD20+ B cells within select vascular lumina (Figure). No extravascular lymphoma cells were seen. Based on the bone marrow and skin biopsies, a diagnosis of IVBCL was made. Unfortunately, no progress was made clinically, and the patient was transitioned to comfort measures. Upon the patient’s death, his family expressed interest in participating in IVBCL research and agreed to a limited autopsy consisting of numerous skin biopsies to evaluate different body sites and biopsy types (normal skin vs hemangiomas) to ascertain whether diagnostic yield could be increased by performing selective biopsies of hemangiomas if IVBCL was suspected.

Twenty-four postmortem 4-mm punch biopsies containing subcutaneous adipose tissue were taken within 24 hours of the death of the patient before embalming. The biopsies were taken from all regions of the body except the head and neck for cosmetic preservation of the decedent. Eighteen of the biopsies were taken from random sites of normal-appearing skin; the remaining 6 were taken from clinically identifiable cherry hemangiomas (5 on the trunk and 1 on the thigh). There was a variable degree of livor mortis in the dependent areas of the body, which was included in the random biopsies from the back to ensure any pooling of dependent blood would not alter the findings.

A histopathologic examination by a board-certified dermatopathologist (M.P.) on a single hematoxylin-eosin–stained level was performed to evaluate each biopsy for superficial involvement and deep involvement by IVBCL. Superficial involvement was defined as dermal involvement at or above the level of the eccrine sweat glands; deep involvement was defined as dermal involvement beneath the eccrine sweat glands and all subcutaneous fat present. Skin and bone marrow biopsies used to make the original diagnosis prior to the patient’s death were reviewed, including CD20 immunohistochemistry for morphologic comparison to the study slides. Involvement was graded as 0 to 3+ (eTable).

Results

Results from all 24 biopsies are shown in the eTable. Twenty-two (91.7%) biopsies showed at least focal involvement by IVBCL. Nine (37.5%) biopsies showed more deep vs superficial involvement of the same site. On average, the 6 biopsies from clinically detected hemangiomas showed more involvement by IVBCL than the random biopsies (eFigures 1 and 2A). The superficial involvement of skin with a hemangioma showed an average score of 2.33 v 0.78 when compared with the superficial aspect of the random biopsies; the deep involvement of skin with a hemangioma showed an average score of 2.67 vs 1.16 when compared with the deep aspect of the random biopsies (eFigure 2B).

Comment

Intravascular large B-cell lymphoma is an aggressive malignancy that traditionally is difficult to diagnose. Many efforts have been made to improve detection and early diagnosis. As cutaneous involvement is common and sometimes the only sign of disease, dermatologists may be called upon to evaluate and biopsy patients with this suspected diagnosis. The purpose of our study was to improve diagnostic efficacy by methodically performing numerous biopsies and assessing the level of involvement of the superficial and deep skin as well as involvement of hemangiomas. The goal of this meticulous approach was to identify the highest-yield areas for biopsy with minimal impact on the patient. Our results showed that random skin biopsies are an effective way to identify IVBCL. Twenty-two (91.7%) biopsies contained at least focal lymphoma cells. Although the 2 biopsies that showed no tumor cells at all happened to both be from the left arm, this is believed to be coincidental. No discernable pattern was identified regarding involvement and anatomic region. Even though 20 (83.3%) biopsies showed superficial involvement, deep biopsy is essential, as 9 (37.5%) biopsies showed increased deep involvement compared to superficial involvement. Therefore, a deep punch biopsy is essential for maximum sensitivity.

Hemangiomas provide a potential target that could increase the sensitivity of a biopsy in the absence of clinical findings, when the disease in question is exclusively intravascular. The data gathered in this study support this idea, as biopsies from hemangiomas showed increased involvement compared to random biopsies, both superficially and deep (2.33 vs 0.78 and 2.67 vs 1.16, respectively). Interestingly, the hemangioma biopsy sites showed increased deep and superficial involvement, despite these typical cherry hemangiomas only involving the superficial dermis. One possible explanation for this is that the hemangiomas have larger-caliber feeder vessels with increased blood flow beneath them. It would then follow that this increased vasculature would increase the chances of identifying intravascular lymphoma cells. This finding further accentuates the need for a deep punch biopsy containing subcutaneous fat. 

Completing the study in the setting of an autopsy provided the advantage of being able to take numerous biopsies without increased harm to the patient. This extensive set of biopsies would not be reasonable to complete on a living patient. This study also has limitations. Although this patient did fall within the typical demographics for patients with IVBCL, the data were still limited to 1 patient. This autopsy format (on a patient whose cause of death was indeed IVBCL) also implies terminal disease, which may mean the patient had a larger disease burden than a living patient who would typically be biopsied. Although this increased disease burden may have increased the sensitivity of finding IVBCL in the biopsies of this study, this further emphasizes the importance of trying to determine any factors that could increase sensitivity in a living patient with a lower disease burden.

Conclusion

Skin biopsies can provide a sensitive, low-cost, and low-morbidity method to assess a patient for IVBCL. Though random skin biopsies can yield valuable information, deep, 4-mm punch biopsies of clinically identifiable hemangiomas may provide the highest sensitivity for IVBCL.

Intravascular large B-cell lymphoma (IVBCL) is an exceedingly rare aggressive form of non-Hodgkin lymphoma with tumor cells growing selectively in vascular lumina.¹ The annual incidence of IVBCL is fewer than 0.5 cases per 1,000,000 individuals worldwide.² Only about 500 known cases of IVBCL have been recorded in the literature,³ and it accounts for less than 1% of all lymphomas. It generally affects middle-aged to elderly individuals, with an average age at diagnosis of 70 years.² It has a predilection for men and commonly develops in individuals who are immunosuppressed.^3,4

Multiple variants of IVBCL have been described in the literature, with central nervous system and cutaneous involvement being the most classic findings.⁵ Bone marrow involvement with hepatosplenomegaly also has been noted in the literature.^6,7 Diagnosis of IVBCL and its variants requires a high index of suspicion, as the clinical manifestations and tissues involved typically are nonspecific and highly variable. Even in the classic variant of IVBCL, skin involvement is only reported in approximately half of cases.³ When present, cutaneous manifestations can range from nodules and violaceous plaques to induration and telangiectasias.³ Lymphadenopathy and lymphoma (leukemic) cells are not seen on a peripheral blood smear.^2,8,9

The lack of lymphadenopathy or identifiable leukemic cells in the peripheral blood presents a diagnostic dilemma, as sufficient information for accurate diagnosis must be obtained while minimizing invasive procedures and resource expenditure. Because IVBCL cells can reside in the vascular lumina of various organs, numerous biopsy sites have been proposed for diagnosis of lymphoma, including the bone marrow, skin, prostate, adrenal gland, brain, liver, and kidneys.¹⁰ While some studies have reported that the optimal diagnostic site is the bone marrow, skin biopsies are more routinely carried out, as they represent a convenient and cost-effective alternative to other more invasive techniques.^6,7,10 Studies have shown biopsy sensitivity values ranging from 77.8% to 83.3% for detection of IVBCL in normal-appearing skin, which is comparable to the sensitivities of a bone marrow biopsy.^7,8 Although skin biopsy of random sites has shown diagnostic efficacy, some studies have proposed that biopsies taken from hemangiomas and other hypervascular lesions can further improve diagnostic yield, as lymphoma cells often are present in capillaries of subcutaneous adipose tissue.^6,10,11 However, no obvious clinicopathologic differences were observed between IVBCL with and without involvement of a cutaneous hemangioma.¹¹

The purpose of this study was to determine the diagnostic efficacy of skin biopsies for detecting IVBCL at various body sites and to establish whether biopsies from hemangiomas yield higher diagnostic value.

Methods

A 66-year-old man recently died at our institution secondary to IVBCL. His disease course was characterized by multiple hospital admissions in a 6-month period for fever of unknown origin and tachycardia unresponsive to broad-spectrum antibiotics and systemic steroids. The patient declined over the course of 3 to 4 weeks with findings suggestive of lymphoma and tumor lysis syndrome, and he eventually developed shock, hypoxic respiratory failure, and acute renal failure. As imaging studies and examinations had not shown lymphadenopathy, bone marrow biopsy was performed, and dermatology was consulted to perform skin biopsies to evaluate for IVBCL. Both bone marrow biopsies and random skin biopsies from the abdomen showed large and atypical CD20+ B cells within select vascular lumina (Figure). No extravascular lymphoma cells were seen. Based on the bone marrow and skin biopsies, a diagnosis of IVBCL was made. Unfortunately, no progress was made clinically, and the patient was transitioned to comfort measures. Upon the patient’s death, his family expressed interest in participating in IVBCL research and agreed to a limited autopsy consisting of numerous skin biopsies to evaluate different body sites and biopsy types (normal skin vs hemangiomas) to ascertain whether diagnostic yield could be increased by performing selective biopsies of hemangiomas if IVBCL was suspected.

Twenty-four postmortem 4-mm punch biopsies containing subcutaneous adipose tissue were taken within 24 hours of the death of the patient before embalming. The biopsies were taken from all regions of the body except the head and neck for cosmetic preservation of the decedent. Eighteen of the biopsies were taken from random sites of normal-appearing skin; the remaining 6 were taken from clinically identifiable cherry hemangiomas (5 on the trunk and 1 on the thigh). There was a variable degree of livor mortis in the dependent areas of the body, which was included in the random biopsies from the back to ensure any pooling of dependent blood would not alter the findings.

A histopathologic examination by a board-certified dermatopathologist (M.P.) on a single hematoxylin-eosin–stained level was performed to evaluate each biopsy for superficial involvement and deep involvement by IVBCL. Superficial involvement was defined as dermal involvement at or above the level of the eccrine sweat glands; deep involvement was defined as dermal involvement beneath the eccrine sweat glands and all subcutaneous fat present. Skin and bone marrow biopsies used to make the original diagnosis prior to the patient’s death were reviewed, including CD20 immunohistochemistry for morphologic comparison to the study slides. Involvement was graded as 0 to 3+ (eTable).

Results

Results from all 24 biopsies are shown in the eTable. Twenty-two (91.7%) biopsies showed at least focal involvement by IVBCL. Nine (37.5%) biopsies showed more deep vs superficial involvement of the same site. On average, the 6 biopsies from clinically detected hemangiomas showed more involvement by IVBCL than the random biopsies (eFigures 1 and 2A). The superficial involvement of skin with a hemangioma showed an average score of 2.33 v 0.78 when compared with the superficial aspect of the random biopsies; the deep involvement of skin with a hemangioma showed an average score of 2.67 vs 1.16 when compared with the deep aspect of the random biopsies (eFigure 2B).

Comment

Intravascular large B-cell lymphoma is an aggressive malignancy that traditionally is difficult to diagnose. Many efforts have been made to improve detection and early diagnosis. As cutaneous involvement is common and sometimes the only sign of disease, dermatologists may be called upon to evaluate and biopsy patients with this suspected diagnosis. The purpose of our study was to improve diagnostic efficacy by methodically performing numerous biopsies and assessing the level of involvement of the superficial and deep skin as well as involvement of hemangiomas. The goal of this meticulous approach was to identify the highest-yield areas for biopsy with minimal impact on the patient. Our results showed that random skin biopsies are an effective way to identify IVBCL. Twenty-two (91.7%) biopsies contained at least focal lymphoma cells. Although the 2 biopsies that showed no tumor cells at all happened to both be from the left arm, this is believed to be coincidental. No discernable pattern was identified regarding involvement and anatomic region. Even though 20 (83.3%) biopsies showed superficial involvement, deep biopsy is essential, as 9 (37.5%) biopsies showed increased deep involvement compared to superficial involvement. Therefore, a deep punch biopsy is essential for maximum sensitivity.

Hemangiomas provide a potential target that could increase the sensitivity of a biopsy in the absence of clinical findings, when the disease in question is exclusively intravascular. The data gathered in this study support this idea, as biopsies from hemangiomas showed increased involvement compared to random biopsies, both superficially and deep (2.33 vs 0.78 and 2.67 vs 1.16, respectively). Interestingly, the hemangioma biopsy sites showed increased deep and superficial involvement, despite these typical cherry hemangiomas only involving the superficial dermis. One possible explanation for this is that the hemangiomas have larger-caliber feeder vessels with increased blood flow beneath them. It would then follow that this increased vasculature would increase the chances of identifying intravascular lymphoma cells. This finding further accentuates the need for a deep punch biopsy containing subcutaneous fat. 

Completing the study in the setting of an autopsy provided the advantage of being able to take numerous biopsies without increased harm to the patient. This extensive set of biopsies would not be reasonable to complete on a living patient. This study also has limitations. Although this patient did fall within the typical demographics for patients with IVBCL, the data were still limited to 1 patient. This autopsy format (on a patient whose cause of death was indeed IVBCL) also implies terminal disease, which may mean the patient had a larger disease burden than a living patient who would typically be biopsied. Although this increased disease burden may have increased the sensitivity of finding IVBCL in the biopsies of this study, this further emphasizes the importance of trying to determine any factors that could increase sensitivity in a living patient with a lower disease burden.

Conclusion

Skin biopsies can provide a sensitive, low-cost, and low-morbidity method to assess a patient for IVBCL. Though random skin biopsies can yield valuable information, deep, 4-mm punch biopsies of clinically identifiable hemangiomas may provide the highest sensitivity for IVBCL.