Hospital Medicine

A previously healthy 4-year-old boy presented to his pediatrician for nasal congestion, left ear pain, and intermittent fevers, which he’d been experiencing for 2 days. His exam was consistent with acute otitis media. Cefdinir was prescribed given a rash allergy to amoxicillin. His fever, congestion, and otalgia improved the next day.

Three days later he developed abdominal pain, fever, and labored breathing; his mother brought him to the emergency department (ED). His temperature was 38.0 °C, heart rate 141 beats per minute, blood pressure 117/71 mm Hg, respiratory rate 22 breaths per minute; he had oxygen saturation of 96% on ambient air. Despite mild accessory muscle use, he appeared comfortable and interactive. His left tympanic membrane was bulging without erythema. His neck was supple and mucous membranes moist. He had neither cervical lymphadenopathy nor conjunctival pallor. The cardiopulmonary exam was normal except for tachycardia. His abdomen was soft and not distended without organomegaly or tenderness.

Upper respiratory tract symptoms are commonly encountered in pediatrics and most often result from self-limited viral processes. Evaluation of a child with upper respiratory tract symptoms aims to identify serious causes like meningitis, as well as assessing the need for antimicrobial therapy. Supportive management is often appropriate in otitis media. His new, more concerning symptoms portend either a progression of the original process causing his upper respiratory tract symptoms or a separate etiology. It is key to determine which signs and symptoms are associated with the primary process and which are compensatory or secondary. If he were to be more ill appearing, for example, it is possible that his respiratory distress may be related to an underlying systemic illness rather than a primary lung process. Respiratory distress, abdominal pain, and fever could be a result of sepsis from an intrabdominal process such as ruptured appendicitis, intussusception, or malrotation with volvulus. Other causes of sepsis, such as meningitis or severe mastoiditis, both rare complications of otitis media, should be considered, although he does not appear severely ill. Acute myelogenous leukemia or other malignancies and illnesses associated with immunodeficiency can present with sepsis and chloromas in the middle ear that can be misconstrued as otitis media.

A chest radiograph demonstrated left lower lobe patchy opacities concerning for pneumonia. Rapid respiratory syncytial virus and influenza antigen test results were negative. Laboratory testing for general bloodwork was not obtained. He was administered a single dose of intramuscular ceftriaxone, prescribed a 5-day course of azithromycin, and discharged home. The child’s breathing gradually improved, but he continued to have subjective fevers. Two days later, he developed dark red urine. His mother brought him back to the outpatient clinic.

At the time of the ED visit, a diagnosis of community-acquired pneumonia was plausible given fever, mildly increased work of breathing, and an opacification on chest radiography. Most community-acquired pneumonia is caused by viruses; common bacterial causes for his age include Streptococcus pneumoniae and Moraxella catarrhalis. The first-line treatment for uncomplicated community-acquired pneumonia in children is amoxicillin, but this was appropriately avoided given his allergy.

The persistent fevers are surprising. The improvement in breathing corresponds to the treatment (and resolution) of community-acquired pneumonia. However, the development of dark urine does not. Red urine—in the absence of ingested pigments (such as those found in beets)—usually results from hematuria, hemoglobinuria, or myoglobinuria. Gross hematuria can originate from the kidneys to the urethral meatus. Abdominal masses, kidney trauma, or underlying kidney disease may all present with gross hematuria (or microscopic hematuria, seen only on urinalysis). The urine should be examined for the presence of heme, protein, and for evidence of infection; microscopy should be performed to examine for cellular casts and dysmorphic red cells. Tests of renal function, a comprehensive metabolic panel, evaluation of hematologic indexes, and assessments of inflammatory markers should be performed.

The child lived with his parents and had no siblings. He experienced no physical trauma, and there was no family history of kidney disease or hematuria. His father had a persistent cough and fever for 1 month, but recovered around the time the patient began to experience his initial symptoms. This was the patient’s third diagnosis of pneumonia. He had not traveled and was up to date with immunizations. He attended day care.

The fact that this is not the first episode of “pneumonia” raises important possibilities. The most likely one is that the child has had multiple viral infections; however, he could have an underlying primary immunodeficiency (PI) that predisposes him to recurrent infections. More severe PIs often present with recurrent sepsis, bacteremia, and failure to thrive, none of which were present in this case. Less severe PIs (such as selective IgA deficiency) could be possible. Another possibility is that these recurrent episodes of pneumonia are a relapsing and remitting noninfectious process, such as an antineutrophil cytoplasmic antibodies–associated vasculitis or anti–glomerular basement membrane disease. The patient’s father’s recent prolonged respiratory symptoms may be suggestive of pertussis or a “walking pneumonia” potentially caused by Mycoplasma or another atypical bacterium.

His temperature was 36.9 °C, heart rate 107 beats per minute, blood pressure was 106/67 mm Hg, and respiratory rate was 24 breaths per minute with oxygen saturation of 100% on ambient air. He was well appearing. His mucous membranes were moist, and oropharynx was clear. He had scleral icterus. The cardiopulmonary exam was normal. He had no significant lymphadenopathy, hepatosplenomegaly, or rashes.

The finding of jaundice is an important diagnostic pivot point, especially when combined with hematuria. The next step is determining if the jaundice is resulting from unconjugated or conjugated hyperbilirubinemia; the former most often stems from hemolysis or impairment in conjugation, while the latter results from intrahepatic or extrahepatic biliary defects. Tests for hepatobiliary injury including evaluations of alanine and aspartate aminotransferases and alkaline phosphatase, as well as for hepatic function such as tests of coagulation, should be performed.

The patient was referred to the ED and admitted for further evaluation. A complete blood count revealed a white blood cell (WBC) count of 10,700/µL (61% polymorphonuclear neutrophils, 30% lymphocytes, 5% monocytes, 3% eosinophils, 1% basophils), hemoglobin count was 10.3 g/dL (reticulocyte 2% with absolute reticulocyte count 58,400/μL), and platelet count was 265,000/µL. Components of the basic metabolic panel were within reference ranges except for a mildly elevated blood urea nitrogen level of 14 mg/dL with normal creatinine level of 0.3 mg/dL. Total protein was 6.7 g/dL (reference range, 6.4-8.3) and albumin 3.9 g/dL (reference range, 3.4-4.8). Alkaline phosphatase level was 188 U/L (reference range, 44-147), aspartate aminotransferase level 76 U/L (reference range, 0-40), and alanine aminotransferase level 12 U/L (reference range, 7-40). Total bilirubin level was 2.4 mg/dL (reference range, less than 1.5) with direct bilirubin level of 0.4 mg/dL. His C-reactive protein level was 1.5 mg/mL (reference range, 0-0.75). Creatinine kinase (CK) level was 2,550 U/L (reference range, 2-198). International Normalized Ratio (INR) was 1.0. Urinalysis was notable for 2+ proteinuria, large hemoglobin pigment, and 6 red blood cells per high power field (reference range, 0-4).

His blood urea nitrogen is elevated, reflecting either prerenal azotemia or increased absorption of nitrogenous products. Unconjugated hyperbilirubinemia may result from impaired hepatic bilirubin uptake (such as in heart failure or portosystemic shunts), impaired bilirubin conjugation (resulting from genetic conditions or drugs), or excess bilirubin production (such as in hemolysis); his anemia and lack of other evidence of hepatic dysfunction point to hemolysis as the etiology. The reticulocyte production index is approximately 1%, which suggests that an increase in erythrocyte generation is present but inadequate. This, however, does not mean that an erythrocyte production abnormality is present since reticulocytosis can be delayed in many cases of acute hemolytic anemia. It is also possible that the same hemolytic process is affecting mature and immature erythrocytes. A peripheral blood smear should be reviewed for evidence of intravascular hemolysis and testing for autoimmune hemolysis should be performed. Notably, his white blood cell and platelet counts are preserved, which makes a bone marrow–involved malignancy or infiltrative process less likely. The alkaline phosphatase elevation may result from either intrahepatic or extrahepatic biliopathy; bone damage is also possible. The elevation of aspartate aminotransferase, CK, and potassium, along with marked urinary heme pigment, may indicate muscle damage; the most common myositis in children is benign acute childhood myositis resulting from viral infection. However, the moderate level of CK elevation seen in this case is nonspecific and can result from many different etiologies. A metabolic myopathy, such as carnitine palmitoyltransferase II deficiency, can be made worse by metabolic stress and result in rhabdomyolysis; the presentations of inborn errors of metabolism are varied and a planned-out, stepwise approach in evaluation is fundamental.

Lactic acid dehydrogenase (LDH) level was 1,457 U/L (reference range, 140-280), and haptoglobin level was less than 6 mg/dL (reference range, 30-200). Peripheral blood smear demonstrated occasional atypical, reactive-appearing lymphocytes with red cell clumping and agglutination, as well as rare target, burr, and fragmented red cells. Test results for urine myoglobin were negative. Results of urine culture were negative. No blood culture was collected.

The elevated LDH, decreased haptoglobin, and findings on the peripheral blood smear confirm hemolysis. The clumping of erythrocytes can be artifactual in the preparation of peripheral smears, but when considered in the context of hemolysis, may be clinically important. Clumping of erythrocytes on the peripheral smear indicates the binding of a protein to antigens on the erythrocyte membrane; when this occurs below body temperature, this is consistent with the presence of a “cold agglutinin,” usually an IgM antibody directed at erythrocyte surface antigens that causes agglutination and destruction, especially in cooler areas of the body. This is a well-known complication of Mycoplasma pneumoniae infections as well as Epstein-Barr virus (EBV) infections; it may also occur with lymphoid malignancies or autoimmune disease.

Direct Coombs IgG test findings were negative, direct Coombs C3 test was positive, and direct Coombs polyspecific test was positive. M pneumoniae IgG antibody level was 1.4 mg/dL (reference ranges: <0.9, negative; 0.91-1.09, equivocal; >1.1, positive); M pneumoniae IgM level was 529 U/mL (reference range: <770, negative). EBV capsid IgM and IgG levels were undetectable. EBV nuclear antigen IgG level was also undetectable. EBV viral load was fewer than 10 copies/mL. Antinuclear antibodies (ANA) level was negative. General IgE and IgM levels were normal, at 11 and 81 mg/dL, respectively. Repeat complete blood count showed WBC of 7,800/µL, hemoglobin of 8.7 g/dL, and platelet count of 341,000/µL. The patient’s hemoglobin remained stable during hospitalization.

This directed testing is helpful in further classifying the patient’s hemolytic anemia. Autoimmune hemolytic anemias are classified into warm antibody–mediated, cold antibody–mediated, and mixed-type forms; drug-induced and alloimmune hemolytic anemias also occur. In addition, both systemic lupus erythematosus and antiphospholipid antibody syndrome can have hemolytic anemia with variable Coombs testing results; neither fit well in this case. The absence of red blood cell–directed IgG antibodies substantially decreases the likelihood of warm antibody–mediated hemolytic anemia. In cold antibody–mediated hemolytic anemia, antibodies bind to the erythrocyte membrane and then adhere to complement C3, which leads to both intravascular and extravascular hemolysis. Important types of cold antibody–mediated hemolytic anemia in children are primary and secondary cold agglutinin disease, along with paroxysmal cold hemoglobinuria. The Donath-Landsteiner test can be helpful in differentiating these conditions. Antibodies to Mycoplasma may be delayed in response to acute infection, and a child who is reinfected may only produce IgG antibodies. Given the patient’s clinical stability and previous health, the most likely diagnosis is Mycoplasma-induced cold antibody–mediated hemolytic anemia. It may be helpful to check convalescent titers to Mycoplasma in 2 to 4 weeks.

Donath-Landsteiner (D-L) antibody test results were positive. Medication-derived hemolytic anemia testing was conducted, but the presence of positive D-L antibody makes the test results inconclusive. This ultimately led to a diagnosis of paroxysmal cold hemoglobinuria (PCH), presumably triggered by a viral syndrome. Convalescent titers to Mycoplasma were not checked given clinical improvement. Because the patient’s hemoglobin was stable during hospitalization, he was not treated with steroids. His parents were counseled on avoiding cold temperatures for several days. Within 1 month, his hemoglobin had recovered without further evidence of hemolysis.

Hemolytic anemia refers to the accelerated destruction of red blood cells and can be further classified as acquired or hereditary.¹ Hereditary conditions causing hemolytic anemia include enzymopathies (eg, glucose-6-phosphate dehydrogenase deficiency), hemoglobinopathies (eg, sickle cell disease), and membrane abnormalities (eg, hereditary spherocytosis). Acquired pathologies include microangiopathic hemolytic anemia (MAHA), anemias directly caused by certain infections such as malaria, and immune-mediated (Coombs-positive) hemolytic anemias.

MAHA can sometimes be life-threatening and is therefore important to identify quickly. In the right clinical context, such processes may be rapidly recognized by the presence of schistocytes on blood smear in addition to an elevated serum LDH level. Schistocytes suggest mechanical destruction of erythrocytes in the vasculature, the hallmark of MAHA. Important MAHAs include thrombocytopenic purpura, hemolytic-uremic syndrome, and disseminated intravascular coagulation. Though this patient did have a mildly elevated LDH, MAHA was less likely because there were no schistocytes on the blood smear.

Autoimmune hemolytic anemias (AIHAs) are another important subset of acquired hemolytic anemias. AIHAs occur when there is antibody-mediated destruction of erythrocytes. The direct Coombs test evaluates for antibody- or complement-­coated erythrocytes. After administration of anti-IgG and anti-­C3 serum, the test evaluates for agglutination of the red cells caused by attached antibodies or complement. Coombs-­positive AIHA can also be categorized by the temperature of agglutination. “Warm” hemolysis often involves IgG autoantibodies (ie, warm agglutinins), while “cold” antibodies, usually IgM autoantibodies, bind at colder temperatures (0-4 °C) and activate complements, including C3. In this patient, the Coombs C3 was positive while the Coombs IgG was negative, which is more suggestive of a cold complement–mediated pathway.

Cold AIHA can be further categorized into primary cold agglutinin disease, secondary cold agglutinin disease, and PCH. Primary cold agglutinin disease is an autoimmune disorder that mostly occurs in adults. Secondary cold AIHA can often be triggered by bacterial infection (commonly M pneumoniae) or viruses including EBV, measles, and mumps.² Medications, including penicillin and cephalosporins, can also be implicated. Secondary cold AIHA is also linked with autoimmune diseases, such as systemic lupus erythematosus and lymphoproliferative disorders. PCH can be identified with the unique presence of a specific autoantibody (ie, D-L autoantibody) that agglutinates at cold temperatures but dissociates on subsequent rewarming.³ Complement remains affixed and activates hemolysis.

The D-L antibody responsible for PCH is an IgG antibody to the P-antigen present on the erythrocyte surface. Since the Coombs test is conducted at normal temperature, it will be positive for the affixed complement but not for IgG. The underlying mechanism for PCH was proposed by Julius Donath, MD, and Karl Landsteiner, MD, in 1904 and is considered to be the first description of autoimmune disease being precipitated by antibodies.⁴ The D-L antibody test itself is uncommonly performed and somewhat difficult to interpret, particularly in adults, and may lead to false-negative results.⁵

PCH is an acquired, cold AIHA more common to children^6,7 and may account for up to 33% of pediatric AIHA cases.⁸ Typical presentation is after an upper respiratory tract illness; however, the trigger is often not identified. Implicated triggers include a number of viruses.⁹ Clinical presentation includes findings of intravascular hemolysis similar to those in our patient. The pathogenic IgG autoantibody is polyclonal and is likely formed because of immune stimulation, which is consistent with the predominance of nonmalignant triggers of this disease process.¹⁰ Hemolysis and associated symptoms are often exacerbated with cold exposure; both typically resolve within 2 weeks. In recurrent cases, which are a minority, immunosuppression may be considered.¹⁰

PCH remains an often-understated cause of hemolytic anemia particularly in children. Lacking obvious pathognomonic clinical symptoms, it may be overlooked for other forms of AIHA or MAHA. However, with a structured approach to evaluation, as with this patient who had hematuria and jaundice, early diagnosis can prevent an unnecessarily extensive workup and can provide reassurance to patient and parents. By understanding the basic categories of hemolytic anemia, the relevant blood testing available, and interpretation of Coombs test results, clinicians can ensure that PCH is a diagnosis that is not left out in the cold.

• Examination for schistocytes on a blood smear can help identify life-threatening causes of hemolytic anemia.
• Characterization of cold AIHA includes defining the underlying etiology as primary cold agglutinin disease, secondary cold agglutinin disease, or PCH.
• PCH is a cold AIHA that is an underrecognized cause of hemolytic anemia in children. The diagnosis of PCH is made by testing for the presence of the D-L antibody.

A previously healthy 4-year-old boy presented to his pediatrician for nasal congestion, left ear pain, and intermittent fevers, which he’d been experiencing for 2 days. His exam was consistent with acute otitis media. Cefdinir was prescribed given a rash allergy to amoxicillin. His fever, congestion, and otalgia improved the next day.

Three days later he developed abdominal pain, fever, and labored breathing; his mother brought him to the emergency department (ED). His temperature was 38.0 °C, heart rate 141 beats per minute, blood pressure 117/71 mm Hg, respiratory rate 22 breaths per minute; he had oxygen saturation of 96% on ambient air. Despite mild accessory muscle use, he appeared comfortable and interactive. His left tympanic membrane was bulging without erythema. His neck was supple and mucous membranes moist. He had neither cervical lymphadenopathy nor conjunctival pallor. The cardiopulmonary exam was normal except for tachycardia. His abdomen was soft and not distended without organomegaly or tenderness.

Upper respiratory tract symptoms are commonly encountered in pediatrics and most often result from self-limited viral processes. Evaluation of a child with upper respiratory tract symptoms aims to identify serious causes like meningitis, as well as assessing the need for antimicrobial therapy. Supportive management is often appropriate in otitis media. His new, more concerning symptoms portend either a progression of the original process causing his upper respiratory tract symptoms or a separate etiology. It is key to determine which signs and symptoms are associated with the primary process and which are compensatory or secondary. If he were to be more ill appearing, for example, it is possible that his respiratory distress may be related to an underlying systemic illness rather than a primary lung process. Respiratory distress, abdominal pain, and fever could be a result of sepsis from an intrabdominal process such as ruptured appendicitis, intussusception, or malrotation with volvulus. Other causes of sepsis, such as meningitis or severe mastoiditis, both rare complications of otitis media, should be considered, although he does not appear severely ill. Acute myelogenous leukemia or other malignancies and illnesses associated with immunodeficiency can present with sepsis and chloromas in the middle ear that can be misconstrued as otitis media.

A chest radiograph demonstrated left lower lobe patchy opacities concerning for pneumonia. Rapid respiratory syncytial virus and influenza antigen test results were negative. Laboratory testing for general bloodwork was not obtained. He was administered a single dose of intramuscular ceftriaxone, prescribed a 5-day course of azithromycin, and discharged home. The child’s breathing gradually improved, but he continued to have subjective fevers. Two days later, he developed dark red urine. His mother brought him back to the outpatient clinic.

At the time of the ED visit, a diagnosis of community-acquired pneumonia was plausible given fever, mildly increased work of breathing, and an opacification on chest radiography. Most community-acquired pneumonia is caused by viruses; common bacterial causes for his age include Streptococcus pneumoniae and Moraxella catarrhalis. The first-line treatment for uncomplicated community-acquired pneumonia in children is amoxicillin, but this was appropriately avoided given his allergy.

The persistent fevers are surprising. The improvement in breathing corresponds to the treatment (and resolution) of community-acquired pneumonia. However, the development of dark urine does not. Red urine—in the absence of ingested pigments (such as those found in beets)—usually results from hematuria, hemoglobinuria, or myoglobinuria. Gross hematuria can originate from the kidneys to the urethral meatus. Abdominal masses, kidney trauma, or underlying kidney disease may all present with gross hematuria (or microscopic hematuria, seen only on urinalysis). The urine should be examined for the presence of heme, protein, and for evidence of infection; microscopy should be performed to examine for cellular casts and dysmorphic red cells. Tests of renal function, a comprehensive metabolic panel, evaluation of hematologic indexes, and assessments of inflammatory markers should be performed.

The child lived with his parents and had no siblings. He experienced no physical trauma, and there was no family history of kidney disease or hematuria. His father had a persistent cough and fever for 1 month, but recovered around the time the patient began to experience his initial symptoms. This was the patient’s third diagnosis of pneumonia. He had not traveled and was up to date with immunizations. He attended day care.

The fact that this is not the first episode of “pneumonia” raises important possibilities. The most likely one is that the child has had multiple viral infections; however, he could have an underlying primary immunodeficiency (PI) that predisposes him to recurrent infections. More severe PIs often present with recurrent sepsis, bacteremia, and failure to thrive, none of which were present in this case. Less severe PIs (such as selective IgA deficiency) could be possible. Another possibility is that these recurrent episodes of pneumonia are a relapsing and remitting noninfectious process, such as an antineutrophil cytoplasmic antibodies–associated vasculitis or anti–glomerular basement membrane disease. The patient’s father’s recent prolonged respiratory symptoms may be suggestive of pertussis or a “walking pneumonia” potentially caused by Mycoplasma or another atypical bacterium.

His temperature was 36.9 °C, heart rate 107 beats per minute, blood pressure was 106/67 mm Hg, and respiratory rate was 24 breaths per minute with oxygen saturation of 100% on ambient air. He was well appearing. His mucous membranes were moist, and oropharynx was clear. He had scleral icterus. The cardiopulmonary exam was normal. He had no significant lymphadenopathy, hepatosplenomegaly, or rashes.

The finding of jaundice is an important diagnostic pivot point, especially when combined with hematuria. The next step is determining if the jaundice is resulting from unconjugated or conjugated hyperbilirubinemia; the former most often stems from hemolysis or impairment in conjugation, while the latter results from intrahepatic or extrahepatic biliary defects. Tests for hepatobiliary injury including evaluations of alanine and aspartate aminotransferases and alkaline phosphatase, as well as for hepatic function such as tests of coagulation, should be performed.

The patient was referred to the ED and admitted for further evaluation. A complete blood count revealed a white blood cell (WBC) count of 10,700/µL (61% polymorphonuclear neutrophils, 30% lymphocytes, 5% monocytes, 3% eosinophils, 1% basophils), hemoglobin count was 10.3 g/dL (reticulocyte 2% with absolute reticulocyte count 58,400/μL), and platelet count was 265,000/µL. Components of the basic metabolic panel were within reference ranges except for a mildly elevated blood urea nitrogen level of 14 mg/dL with normal creatinine level of 0.3 mg/dL. Total protein was 6.7 g/dL (reference range, 6.4-8.3) and albumin 3.9 g/dL (reference range, 3.4-4.8). Alkaline phosphatase level was 188 U/L (reference range, 44-147), aspartate aminotransferase level 76 U/L (reference range, 0-40), and alanine aminotransferase level 12 U/L (reference range, 7-40). Total bilirubin level was 2.4 mg/dL (reference range, less than 1.5) with direct bilirubin level of 0.4 mg/dL. His C-reactive protein level was 1.5 mg/mL (reference range, 0-0.75). Creatinine kinase (CK) level was 2,550 U/L (reference range, 2-198). International Normalized Ratio (INR) was 1.0. Urinalysis was notable for 2+ proteinuria, large hemoglobin pigment, and 6 red blood cells per high power field (reference range, 0-4).

His blood urea nitrogen is elevated, reflecting either prerenal azotemia or increased absorption of nitrogenous products. Unconjugated hyperbilirubinemia may result from impaired hepatic bilirubin uptake (such as in heart failure or portosystemic shunts), impaired bilirubin conjugation (resulting from genetic conditions or drugs), or excess bilirubin production (such as in hemolysis); his anemia and lack of other evidence of hepatic dysfunction point to hemolysis as the etiology. The reticulocyte production index is approximately 1%, which suggests that an increase in erythrocyte generation is present but inadequate. This, however, does not mean that an erythrocyte production abnormality is present since reticulocytosis can be delayed in many cases of acute hemolytic anemia. It is also possible that the same hemolytic process is affecting mature and immature erythrocytes. A peripheral blood smear should be reviewed for evidence of intravascular hemolysis and testing for autoimmune hemolysis should be performed. Notably, his white blood cell and platelet counts are preserved, which makes a bone marrow–involved malignancy or infiltrative process less likely. The alkaline phosphatase elevation may result from either intrahepatic or extrahepatic biliopathy; bone damage is also possible. The elevation of aspartate aminotransferase, CK, and potassium, along with marked urinary heme pigment, may indicate muscle damage; the most common myositis in children is benign acute childhood myositis resulting from viral infection. However, the moderate level of CK elevation seen in this case is nonspecific and can result from many different etiologies. A metabolic myopathy, such as carnitine palmitoyltransferase II deficiency, can be made worse by metabolic stress and result in rhabdomyolysis; the presentations of inborn errors of metabolism are varied and a planned-out, stepwise approach in evaluation is fundamental.

Lactic acid dehydrogenase (LDH) level was 1,457 U/L (reference range, 140-280), and haptoglobin level was less than 6 mg/dL (reference range, 30-200). Peripheral blood smear demonstrated occasional atypical, reactive-appearing lymphocytes with red cell clumping and agglutination, as well as rare target, burr, and fragmented red cells. Test results for urine myoglobin were negative. Results of urine culture were negative. No blood culture was collected.

The elevated LDH, decreased haptoglobin, and findings on the peripheral blood smear confirm hemolysis. The clumping of erythrocytes can be artifactual in the preparation of peripheral smears, but when considered in the context of hemolysis, may be clinically important. Clumping of erythrocytes on the peripheral smear indicates the binding of a protein to antigens on the erythrocyte membrane; when this occurs below body temperature, this is consistent with the presence of a “cold agglutinin,” usually an IgM antibody directed at erythrocyte surface antigens that causes agglutination and destruction, especially in cooler areas of the body. This is a well-known complication of Mycoplasma pneumoniae infections as well as Epstein-Barr virus (EBV) infections; it may also occur with lymphoid malignancies or autoimmune disease.

Direct Coombs IgG test findings were negative, direct Coombs C3 test was positive, and direct Coombs polyspecific test was positive. M pneumoniae IgG antibody level was 1.4 mg/dL (reference ranges: <0.9, negative; 0.91-1.09, equivocal; >1.1, positive); M pneumoniae IgM level was 529 U/mL (reference range: <770, negative). EBV capsid IgM and IgG levels were undetectable. EBV nuclear antigen IgG level was also undetectable. EBV viral load was fewer than 10 copies/mL. Antinuclear antibodies (ANA) level was negative. General IgE and IgM levels were normal, at 11 and 81 mg/dL, respectively. Repeat complete blood count showed WBC of 7,800/µL, hemoglobin of 8.7 g/dL, and platelet count of 341,000/µL. The patient’s hemoglobin remained stable during hospitalization.

This directed testing is helpful in further classifying the patient’s hemolytic anemia. Autoimmune hemolytic anemias are classified into warm antibody–mediated, cold antibody–mediated, and mixed-type forms; drug-induced and alloimmune hemolytic anemias also occur. In addition, both systemic lupus erythematosus and antiphospholipid antibody syndrome can have hemolytic anemia with variable Coombs testing results; neither fit well in this case. The absence of red blood cell–directed IgG antibodies substantially decreases the likelihood of warm antibody–mediated hemolytic anemia. In cold antibody–mediated hemolytic anemia, antibodies bind to the erythrocyte membrane and then adhere to complement C3, which leads to both intravascular and extravascular hemolysis. Important types of cold antibody–mediated hemolytic anemia in children are primary and secondary cold agglutinin disease, along with paroxysmal cold hemoglobinuria. The Donath-Landsteiner test can be helpful in differentiating these conditions. Antibodies to Mycoplasma may be delayed in response to acute infection, and a child who is reinfected may only produce IgG antibodies. Given the patient’s clinical stability and previous health, the most likely diagnosis is Mycoplasma-induced cold antibody–mediated hemolytic anemia. It may be helpful to check convalescent titers to Mycoplasma in 2 to 4 weeks.

Donath-Landsteiner (D-L) antibody test results were positive. Medication-derived hemolytic anemia testing was conducted, but the presence of positive D-L antibody makes the test results inconclusive. This ultimately led to a diagnosis of paroxysmal cold hemoglobinuria (PCH), presumably triggered by a viral syndrome. Convalescent titers to Mycoplasma were not checked given clinical improvement. Because the patient’s hemoglobin was stable during hospitalization, he was not treated with steroids. His parents were counseled on avoiding cold temperatures for several days. Within 1 month, his hemoglobin had recovered without further evidence of hemolysis.

Hemolytic anemia refers to the accelerated destruction of red blood cells and can be further classified as acquired or hereditary.¹ Hereditary conditions causing hemolytic anemia include enzymopathies (eg, glucose-6-phosphate dehydrogenase deficiency), hemoglobinopathies (eg, sickle cell disease), and membrane abnormalities (eg, hereditary spherocytosis). Acquired pathologies include microangiopathic hemolytic anemia (MAHA), anemias directly caused by certain infections such as malaria, and immune-mediated (Coombs-positive) hemolytic anemias.

MAHA can sometimes be life-threatening and is therefore important to identify quickly. In the right clinical context, such processes may be rapidly recognized by the presence of schistocytes on blood smear in addition to an elevated serum LDH level. Schistocytes suggest mechanical destruction of erythrocytes in the vasculature, the hallmark of MAHA. Important MAHAs include thrombocytopenic purpura, hemolytic-uremic syndrome, and disseminated intravascular coagulation. Though this patient did have a mildly elevated LDH, MAHA was less likely because there were no schistocytes on the blood smear.

Autoimmune hemolytic anemias (AIHAs) are another important subset of acquired hemolytic anemias. AIHAs occur when there is antibody-mediated destruction of erythrocytes. The direct Coombs test evaluates for antibody- or complement-­coated erythrocytes. After administration of anti-IgG and anti-­C3 serum, the test evaluates for agglutination of the red cells caused by attached antibodies or complement. Coombs-­positive AIHA can also be categorized by the temperature of agglutination. “Warm” hemolysis often involves IgG autoantibodies (ie, warm agglutinins), while “cold” antibodies, usually IgM autoantibodies, bind at colder temperatures (0-4 °C) and activate complements, including C3. In this patient, the Coombs C3 was positive while the Coombs IgG was negative, which is more suggestive of a cold complement–mediated pathway.

Cold AIHA can be further categorized into primary cold agglutinin disease, secondary cold agglutinin disease, and PCH. Primary cold agglutinin disease is an autoimmune disorder that mostly occurs in adults. Secondary cold AIHA can often be triggered by bacterial infection (commonly M pneumoniae) or viruses including EBV, measles, and mumps.² Medications, including penicillin and cephalosporins, can also be implicated. Secondary cold AIHA is also linked with autoimmune diseases, such as systemic lupus erythematosus and lymphoproliferative disorders. PCH can be identified with the unique presence of a specific autoantibody (ie, D-L autoantibody) that agglutinates at cold temperatures but dissociates on subsequent rewarming.³ Complement remains affixed and activates hemolysis.

The D-L antibody responsible for PCH is an IgG antibody to the P-antigen present on the erythrocyte surface. Since the Coombs test is conducted at normal temperature, it will be positive for the affixed complement but not for IgG. The underlying mechanism for PCH was proposed by Julius Donath, MD, and Karl Landsteiner, MD, in 1904 and is considered to be the first description of autoimmune disease being precipitated by antibodies.⁴ The D-L antibody test itself is uncommonly performed and somewhat difficult to interpret, particularly in adults, and may lead to false-negative results.⁵

PCH is an acquired, cold AIHA more common to children^6,7 and may account for up to 33% of pediatric AIHA cases.⁸ Typical presentation is after an upper respiratory tract illness; however, the trigger is often not identified. Implicated triggers include a number of viruses.⁹ Clinical presentation includes findings of intravascular hemolysis similar to those in our patient. The pathogenic IgG autoantibody is polyclonal and is likely formed because of immune stimulation, which is consistent with the predominance of nonmalignant triggers of this disease process.¹⁰ Hemolysis and associated symptoms are often exacerbated with cold exposure; both typically resolve within 2 weeks. In recurrent cases, which are a minority, immunosuppression may be considered.¹⁰

PCH remains an often-understated cause of hemolytic anemia particularly in children. Lacking obvious pathognomonic clinical symptoms, it may be overlooked for other forms of AIHA or MAHA. However, with a structured approach to evaluation, as with this patient who had hematuria and jaundice, early diagnosis can prevent an unnecessarily extensive workup and can provide reassurance to patient and parents. By understanding the basic categories of hemolytic anemia, the relevant blood testing available, and interpretation of Coombs test results, clinicians can ensure that PCH is a diagnosis that is not left out in the cold.

• Examination for schistocytes on a blood smear can help identify life-threatening causes of hemolytic anemia.
• Characterization of cold AIHA includes defining the underlying etiology as primary cold agglutinin disease, secondary cold agglutinin disease, or PCH.
• PCH is a cold AIHA that is an underrecognized cause of hemolytic anemia in children. The diagnosis of PCH is made by testing for the presence of the D-L antibody.

A previously healthy 4-year-old boy presented to his pediatrician for nasal congestion, left ear pain, and intermittent fevers, which he’d been experiencing for 2 days. His exam was consistent with acute otitis media. Cefdinir was prescribed given a rash allergy to amoxicillin. His fever, congestion, and otalgia improved the next day.

Three days later he developed abdominal pain, fever, and labored breathing; his mother brought him to the emergency department (ED). His temperature was 38.0 °C, heart rate 141 beats per minute, blood pressure 117/71 mm Hg, respiratory rate 22 breaths per minute; he had oxygen saturation of 96% on ambient air. Despite mild accessory muscle use, he appeared comfortable and interactive. His left tympanic membrane was bulging without erythema. His neck was supple and mucous membranes moist. He had neither cervical lymphadenopathy nor conjunctival pallor. The cardiopulmonary exam was normal except for tachycardia. His abdomen was soft and not distended without organomegaly or tenderness.

Upper respiratory tract symptoms are commonly encountered in pediatrics and most often result from self-limited viral processes. Evaluation of a child with upper respiratory tract symptoms aims to identify serious causes like meningitis, as well as assessing the need for antimicrobial therapy. Supportive management is often appropriate in otitis media. His new, more concerning symptoms portend either a progression of the original process causing his upper respiratory tract symptoms or a separate etiology. It is key to determine which signs and symptoms are associated with the primary process and which are compensatory or secondary. If he were to be more ill appearing, for example, it is possible that his respiratory distress may be related to an underlying systemic illness rather than a primary lung process. Respiratory distress, abdominal pain, and fever could be a result of sepsis from an intrabdominal process such as ruptured appendicitis, intussusception, or malrotation with volvulus. Other causes of sepsis, such as meningitis or severe mastoiditis, both rare complications of otitis media, should be considered, although he does not appear severely ill. Acute myelogenous leukemia or other malignancies and illnesses associated with immunodeficiency can present with sepsis and chloromas in the middle ear that can be misconstrued as otitis media.

A chest radiograph demonstrated left lower lobe patchy opacities concerning for pneumonia. Rapid respiratory syncytial virus and influenza antigen test results were negative. Laboratory testing for general bloodwork was not obtained. He was administered a single dose of intramuscular ceftriaxone, prescribed a 5-day course of azithromycin, and discharged home. The child’s breathing gradually improved, but he continued to have subjective fevers. Two days later, he developed dark red urine. His mother brought him back to the outpatient clinic.

At the time of the ED visit, a diagnosis of community-acquired pneumonia was plausible given fever, mildly increased work of breathing, and an opacification on chest radiography. Most community-acquired pneumonia is caused by viruses; common bacterial causes for his age include Streptococcus pneumoniae and Moraxella catarrhalis. The first-line treatment for uncomplicated community-acquired pneumonia in children is amoxicillin, but this was appropriately avoided given his allergy.

The persistent fevers are surprising. The improvement in breathing corresponds to the treatment (and resolution) of community-acquired pneumonia. However, the development of dark urine does not. Red urine—in the absence of ingested pigments (such as those found in beets)—usually results from hematuria, hemoglobinuria, or myoglobinuria. Gross hematuria can originate from the kidneys to the urethral meatus. Abdominal masses, kidney trauma, or underlying kidney disease may all present with gross hematuria (or microscopic hematuria, seen only on urinalysis). The urine should be examined for the presence of heme, protein, and for evidence of infection; microscopy should be performed to examine for cellular casts and dysmorphic red cells. Tests of renal function, a comprehensive metabolic panel, evaluation of hematologic indexes, and assessments of inflammatory markers should be performed.

The child lived with his parents and had no siblings. He experienced no physical trauma, and there was no family history of kidney disease or hematuria. His father had a persistent cough and fever for 1 month, but recovered around the time the patient began to experience his initial symptoms. This was the patient’s third diagnosis of pneumonia. He had not traveled and was up to date with immunizations. He attended day care.

The fact that this is not the first episode of “pneumonia” raises important possibilities. The most likely one is that the child has had multiple viral infections; however, he could have an underlying primary immunodeficiency (PI) that predisposes him to recurrent infections. More severe PIs often present with recurrent sepsis, bacteremia, and failure to thrive, none of which were present in this case. Less severe PIs (such as selective IgA deficiency) could be possible. Another possibility is that these recurrent episodes of pneumonia are a relapsing and remitting noninfectious process, such as an antineutrophil cytoplasmic antibodies–associated vasculitis or anti–glomerular basement membrane disease. The patient’s father’s recent prolonged respiratory symptoms may be suggestive of pertussis or a “walking pneumonia” potentially caused by Mycoplasma or another atypical bacterium.

His temperature was 36.9 °C, heart rate 107 beats per minute, blood pressure was 106/67 mm Hg, and respiratory rate was 24 breaths per minute with oxygen saturation of 100% on ambient air. He was well appearing. His mucous membranes were moist, and oropharynx was clear. He had scleral icterus. The cardiopulmonary exam was normal. He had no significant lymphadenopathy, hepatosplenomegaly, or rashes.

The finding of jaundice is an important diagnostic pivot point, especially when combined with hematuria. The next step is determining if the jaundice is resulting from unconjugated or conjugated hyperbilirubinemia; the former most often stems from hemolysis or impairment in conjugation, while the latter results from intrahepatic or extrahepatic biliary defects. Tests for hepatobiliary injury including evaluations of alanine and aspartate aminotransferases and alkaline phosphatase, as well as for hepatic function such as tests of coagulation, should be performed.

The patient was referred to the ED and admitted for further evaluation. A complete blood count revealed a white blood cell (WBC) count of 10,700/µL (61% polymorphonuclear neutrophils, 30% lymphocytes, 5% monocytes, 3% eosinophils, 1% basophils), hemoglobin count was 10.3 g/dL (reticulocyte 2% with absolute reticulocyte count 58,400/μL), and platelet count was 265,000/µL. Components of the basic metabolic panel were within reference ranges except for a mildly elevated blood urea nitrogen level of 14 mg/dL with normal creatinine level of 0.3 mg/dL. Total protein was 6.7 g/dL (reference range, 6.4-8.3) and albumin 3.9 g/dL (reference range, 3.4-4.8). Alkaline phosphatase level was 188 U/L (reference range, 44-147), aspartate aminotransferase level 76 U/L (reference range, 0-40), and alanine aminotransferase level 12 U/L (reference range, 7-40). Total bilirubin level was 2.4 mg/dL (reference range, less than 1.5) with direct bilirubin level of 0.4 mg/dL. His C-reactive protein level was 1.5 mg/mL (reference range, 0-0.75). Creatinine kinase (CK) level was 2,550 U/L (reference range, 2-198). International Normalized Ratio (INR) was 1.0. Urinalysis was notable for 2+ proteinuria, large hemoglobin pigment, and 6 red blood cells per high power field (reference range, 0-4).

His blood urea nitrogen is elevated, reflecting either prerenal azotemia or increased absorption of nitrogenous products. Unconjugated hyperbilirubinemia may result from impaired hepatic bilirubin uptake (such as in heart failure or portosystemic shunts), impaired bilirubin conjugation (resulting from genetic conditions or drugs), or excess bilirubin production (such as in hemolysis); his anemia and lack of other evidence of hepatic dysfunction point to hemolysis as the etiology. The reticulocyte production index is approximately 1%, which suggests that an increase in erythrocyte generation is present but inadequate. This, however, does not mean that an erythrocyte production abnormality is present since reticulocytosis can be delayed in many cases of acute hemolytic anemia. It is also possible that the same hemolytic process is affecting mature and immature erythrocytes. A peripheral blood smear should be reviewed for evidence of intravascular hemolysis and testing for autoimmune hemolysis should be performed. Notably, his white blood cell and platelet counts are preserved, which makes a bone marrow–involved malignancy or infiltrative process less likely. The alkaline phosphatase elevation may result from either intrahepatic or extrahepatic biliopathy; bone damage is also possible. The elevation of aspartate aminotransferase, CK, and potassium, along with marked urinary heme pigment, may indicate muscle damage; the most common myositis in children is benign acute childhood myositis resulting from viral infection. However, the moderate level of CK elevation seen in this case is nonspecific and can result from many different etiologies. A metabolic myopathy, such as carnitine palmitoyltransferase II deficiency, can be made worse by metabolic stress and result in rhabdomyolysis; the presentations of inborn errors of metabolism are varied and a planned-out, stepwise approach in evaluation is fundamental.

Lactic acid dehydrogenase (LDH) level was 1,457 U/L (reference range, 140-280), and haptoglobin level was less than 6 mg/dL (reference range, 30-200). Peripheral blood smear demonstrated occasional atypical, reactive-appearing lymphocytes with red cell clumping and agglutination, as well as rare target, burr, and fragmented red cells. Test results for urine myoglobin were negative. Results of urine culture were negative. No blood culture was collected.

The elevated LDH, decreased haptoglobin, and findings on the peripheral blood smear confirm hemolysis. The clumping of erythrocytes can be artifactual in the preparation of peripheral smears, but when considered in the context of hemolysis, may be clinically important. Clumping of erythrocytes on the peripheral smear indicates the binding of a protein to antigens on the erythrocyte membrane; when this occurs below body temperature, this is consistent with the presence of a “cold agglutinin,” usually an IgM antibody directed at erythrocyte surface antigens that causes agglutination and destruction, especially in cooler areas of the body. This is a well-known complication of Mycoplasma pneumoniae infections as well as Epstein-Barr virus (EBV) infections; it may also occur with lymphoid malignancies or autoimmune disease.

Direct Coombs IgG test findings were negative, direct Coombs C3 test was positive, and direct Coombs polyspecific test was positive. M pneumoniae IgG antibody level was 1.4 mg/dL (reference ranges: <0.9, negative; 0.91-1.09, equivocal; >1.1, positive); M pneumoniae IgM level was 529 U/mL (reference range: <770, negative). EBV capsid IgM and IgG levels were undetectable. EBV nuclear antigen IgG level was also undetectable. EBV viral load was fewer than 10 copies/mL. Antinuclear antibodies (ANA) level was negative. General IgE and IgM levels were normal, at 11 and 81 mg/dL, respectively. Repeat complete blood count showed WBC of 7,800/µL, hemoglobin of 8.7 g/dL, and platelet count of 341,000/µL. The patient’s hemoglobin remained stable during hospitalization.

This directed testing is helpful in further classifying the patient’s hemolytic anemia. Autoimmune hemolytic anemias are classified into warm antibody–mediated, cold antibody–mediated, and mixed-type forms; drug-induced and alloimmune hemolytic anemias also occur. In addition, both systemic lupus erythematosus and antiphospholipid antibody syndrome can have hemolytic anemia with variable Coombs testing results; neither fit well in this case. The absence of red blood cell–directed IgG antibodies substantially decreases the likelihood of warm antibody–mediated hemolytic anemia. In cold antibody–mediated hemolytic anemia, antibodies bind to the erythrocyte membrane and then adhere to complement C3, which leads to both intravascular and extravascular hemolysis. Important types of cold antibody–mediated hemolytic anemia in children are primary and secondary cold agglutinin disease, along with paroxysmal cold hemoglobinuria. The Donath-Landsteiner test can be helpful in differentiating these conditions. Antibodies to Mycoplasma may be delayed in response to acute infection, and a child who is reinfected may only produce IgG antibodies. Given the patient’s clinical stability and previous health, the most likely diagnosis is Mycoplasma-induced cold antibody–mediated hemolytic anemia. It may be helpful to check convalescent titers to Mycoplasma in 2 to 4 weeks.

Donath-Landsteiner (D-L) antibody test results were positive. Medication-derived hemolytic anemia testing was conducted, but the presence of positive D-L antibody makes the test results inconclusive. This ultimately led to a diagnosis of paroxysmal cold hemoglobinuria (PCH), presumably triggered by a viral syndrome. Convalescent titers to Mycoplasma were not checked given clinical improvement. Because the patient’s hemoglobin was stable during hospitalization, he was not treated with steroids. His parents were counseled on avoiding cold temperatures for several days. Within 1 month, his hemoglobin had recovered without further evidence of hemolysis.

Hemolytic anemia refers to the accelerated destruction of red blood cells and can be further classified as acquired or hereditary.¹ Hereditary conditions causing hemolytic anemia include enzymopathies (eg, glucose-6-phosphate dehydrogenase deficiency), hemoglobinopathies (eg, sickle cell disease), and membrane abnormalities (eg, hereditary spherocytosis). Acquired pathologies include microangiopathic hemolytic anemia (MAHA), anemias directly caused by certain infections such as malaria, and immune-mediated (Coombs-positive) hemolytic anemias.

MAHA can sometimes be life-threatening and is therefore important to identify quickly. In the right clinical context, such processes may be rapidly recognized by the presence of schistocytes on blood smear in addition to an elevated serum LDH level. Schistocytes suggest mechanical destruction of erythrocytes in the vasculature, the hallmark of MAHA. Important MAHAs include thrombocytopenic purpura, hemolytic-uremic syndrome, and disseminated intravascular coagulation. Though this patient did have a mildly elevated LDH, MAHA was less likely because there were no schistocytes on the blood smear.

Autoimmune hemolytic anemias (AIHAs) are another important subset of acquired hemolytic anemias. AIHAs occur when there is antibody-mediated destruction of erythrocytes. The direct Coombs test evaluates for antibody- or complement-­coated erythrocytes. After administration of anti-IgG and anti-­C3 serum, the test evaluates for agglutination of the red cells caused by attached antibodies or complement. Coombs-­positive AIHA can also be categorized by the temperature of agglutination. “Warm” hemolysis often involves IgG autoantibodies (ie, warm agglutinins), while “cold” antibodies, usually IgM autoantibodies, bind at colder temperatures (0-4 °C) and activate complements, including C3. In this patient, the Coombs C3 was positive while the Coombs IgG was negative, which is more suggestive of a cold complement–mediated pathway.

Cold AIHA can be further categorized into primary cold agglutinin disease, secondary cold agglutinin disease, and PCH. Primary cold agglutinin disease is an autoimmune disorder that mostly occurs in adults. Secondary cold AIHA can often be triggered by bacterial infection (commonly M pneumoniae) or viruses including EBV, measles, and mumps.² Medications, including penicillin and cephalosporins, can also be implicated. Secondary cold AIHA is also linked with autoimmune diseases, such as systemic lupus erythematosus and lymphoproliferative disorders. PCH can be identified with the unique presence of a specific autoantibody (ie, D-L autoantibody) that agglutinates at cold temperatures but dissociates on subsequent rewarming.³ Complement remains affixed and activates hemolysis.

The D-L antibody responsible for PCH is an IgG antibody to the P-antigen present on the erythrocyte surface. Since the Coombs test is conducted at normal temperature, it will be positive for the affixed complement but not for IgG. The underlying mechanism for PCH was proposed by Julius Donath, MD, and Karl Landsteiner, MD, in 1904 and is considered to be the first description of autoimmune disease being precipitated by antibodies.⁴ The D-L antibody test itself is uncommonly performed and somewhat difficult to interpret, particularly in adults, and may lead to false-negative results.⁵

PCH is an acquired, cold AIHA more common to children^6,7 and may account for up to 33% of pediatric AIHA cases.⁸ Typical presentation is after an upper respiratory tract illness; however, the trigger is often not identified. Implicated triggers include a number of viruses.⁹ Clinical presentation includes findings of intravascular hemolysis similar to those in our patient. The pathogenic IgG autoantibody is polyclonal and is likely formed because of immune stimulation, which is consistent with the predominance of nonmalignant triggers of this disease process.¹⁰ Hemolysis and associated symptoms are often exacerbated with cold exposure; both typically resolve within 2 weeks. In recurrent cases, which are a minority, immunosuppression may be considered.¹⁰

PCH remains an often-understated cause of hemolytic anemia particularly in children. Lacking obvious pathognomonic clinical symptoms, it may be overlooked for other forms of AIHA or MAHA. However, with a structured approach to evaluation, as with this patient who had hematuria and jaundice, early diagnosis can prevent an unnecessarily extensive workup and can provide reassurance to patient and parents. By understanding the basic categories of hemolytic anemia, the relevant blood testing available, and interpretation of Coombs test results, clinicians can ensure that PCH is a diagnosis that is not left out in the cold.

• Examination for schistocytes on a blood smear can help identify life-threatening causes of hemolytic anemia.
• Characterization of cold AIHA includes defining the underlying etiology as primary cold agglutinin disease, secondary cold agglutinin disease, or PCH.
• PCH is a cold AIHA that is an underrecognized cause of hemolytic anemia in children. The diagnosis of PCH is made by testing for the presence of the D-L antibody.

The story of the coronavirus disease 2019 (COVID-19) pandemic in the United States has been defined, in part, by a persistent shortage of medical supplies that has made it difficult and dangerous for healthcare workers to care for infected patients. States, health systems, and even individual hospitals are currently competing against one another—sometimes at auction—to obtain personal protective equipment (PPE). This “Wild West” scenario has resulted in bizarre stories involving attempts to obtain PPE. One health system recently described a James Bond–like pursuit of essential PPE, complete with a covert trip to an industrial warehouse, trucks filled with masks but labeled as food delivery vehicles, and an intervention by a United States congressman.¹ Many states have experienced analogous, but still atypical, stories: masks flown in from China using the private jet of a professional sports team owner,² widespread use of novel sterilization modalities to allow PPE reuse,³ and one attempt to purchase price-gouged PPE from the host of the show “Shark Tank.”⁴ In some cases, hospitals and healthcare workers have pleaded for PPE on fundraising and social media sites.⁵

These profound deviations from operations of contemporary health system supply chains would have seemed beyond belief just a few months ago. Instead, they now echo the collective experiences of healthcare stakeholders trying to obtain PPE to protect their frontline healthcare workers during the COVID-19 pandemic.

How did we get into this situation? The manufacture of medical supplies like gowns and masks is a highly competitive business with very slim margins, and as a result, medical equipment manufacturers aim to match their supply with the market’s demand, with hospitals and health systems using just-in-time ordering to limit excess inventory.⁶ While this approach adds efficiency and reduces costs, it also renders manufacturers and customers vulnerable to supply disruptions and shortages when need surges. The COVID-19 pandemic represents perhaps the most extreme example of a massive, widespread surge in demand that occurred multifocally and in a highly compressed time frame. Unlike other industries (eg, consumer paper products), however, in which demand exceeding supply causes inconvenience, the lack of PPE has led to critical public health consequences, with lives of both healthcare workers and vulnerable patients lost because of these shortages of medical equipment.

There are many reasons for the PPE crisis. As noted above, manufacturers have prioritized efficiency over the ability to quickly increase production. They adhere to just-in-time ordering rather than planning for a surge in demand with extra production capacity, all to avoid having warehouses filled with unsold products if surges never occur. This strategy, compounded by the fact that most PPE in the United States is imported from areas in Asia that were profoundly affected early on by COVID-19, led to the observed widespread shortages. When PPE became unavailable from usual suppliers, hospitals were unable to locate other sources of existing PPE because of a lack of transparency about where PPE could be found and how it could be obtained. The Food and Drug Administration and other federal regulatory agencies maintained strict regulations around PPE production and, despite the crisis, made few exceptions.⁷ The FDA did grant a few Emergency Use Authorizations (EUAs) for certain improvised, decontaminated, or alternative respirators (eg, the Chinese-made KN95), but it has only very infrequently issued EUAs to allow domestic manufacturers to ramp up production.⁸ These failures were accompanied by an serious increase in PPE use, leading to spikes in price, price gouging, and hoarding,⁹ problems that were further magnified as health systems and hospitals were forced to compete with nonhealthcare businesses for PPE.

The Defense Production Act (DPA) gives the federal government the power to increase production of goods needed during a crisis⁸ to offer purchasing guarantees, coordinate federal agencies, and regulate distribution and pricing. However, the current administration’s failure to mount a coordinated federal response has contributed to the observed market instability, medical supply shortages, and public health crisis we face. We have previously recommended that the federal government use the power of the DPA to reduce manufacturers’ risk of being uncompensated for excess supply, support temporary reductions in regulatory barriers, and create mandatory centralized reporting of PPE supply, including completed PPE and its components.¹⁰ We stand by these recommendations but also acknowledge that hospitals and health systems may be simultaneously considering how to best prepare for future crises and even surges in demand over the next 18 months as the COVID-19 pandemic continues.

1. Encourage mandates at the hospital, health system, and state level regarding minimum inventory levels for essential equipment. Stockpiles are essential for emergency preparedness. In the long term, these sorts of stockpiles are economically infeasible without government help to maintain them. In the near term, however, it is sensible that hospitals and health systems would maintain a minimum of 2 weeks’ worth of PPE to prepare for expected regional spikes in COVID-19 cases. However, a soon-to-published study suggests that over 40% of hospitals had a PPE stockpile of less than 2 weeks.¹¹ Although this survey was conducted at the height of the shortage, it suggests that there is opportunity for improvement.

2. Coordinate efforts among states and health systems to collect and report inventory, regionalize resources, and coordinate their distribution. The best example of this is the seven-state purchasing consortium announced by New York Governor Andrew Cuomo in early May.¹² Unfortunately, since the announcement, there have been few details about whether the states were successful in their effort to reduce prices or to obtain PPE in bulk. Still, hospitals and health systems could join or emulate purchasing collaboratives to allow resources to be better allocated according to need. There are barriers to such collaboratives because the market is currently set up to encourage competition among health systems and hospitals. During the pandemic, however, cooperation has increasingly been favored over competition in science and healthcare delivery. There are also existing hospital purchasing collaboratives (eg, Premier, Inc¹³), which have taken steps to vet suppliers and improve access to PPE, but it is not clear how successful these efforts have been to date.

3. Advocate for strong federal leadership, including support for increased domestic manufacturing; replenishment and maintenance of state and health system stockpiles of PPE, ventilators, and medications; and development of a centrally coordinated PPE allocation and distribution process. While hospitals and health systems may favor remaining as apolitical as possible, the need for a federal response to stabilize the PPE market may be too urgent and necessary to ignore.

As hospitals and health systems prepare for continued surges in COVID-19 cases, they face challenges in providing PPE for frontline clinicians and staff. A federal plan to enhance nimbleness in responding to multifocal, geographic outbreaks and ensure awareness regarding inventory would improve our chances to successfully navigate the next pandemic and optimize the protection of our health workers, patients, and public health. In the absence of such a plan, hospitals should maintain a minimum of 2 weeks’ worth of PPE to prepare for expected regional spikes in COVID-19 cases and should continue to attempt to coordinate efforts among states and health systems to collect and report inventory, regionalize resources, and coordinate their distribution.

The story of the coronavirus disease 2019 (COVID-19) pandemic in the United States has been defined, in part, by a persistent shortage of medical supplies that has made it difficult and dangerous for healthcare workers to care for infected patients. States, health systems, and even individual hospitals are currently competing against one another—sometimes at auction—to obtain personal protective equipment (PPE). This “Wild West” scenario has resulted in bizarre stories involving attempts to obtain PPE. One health system recently described a James Bond–like pursuit of essential PPE, complete with a covert trip to an industrial warehouse, trucks filled with masks but labeled as food delivery vehicles, and an intervention by a United States congressman.¹ Many states have experienced analogous, but still atypical, stories: masks flown in from China using the private jet of a professional sports team owner,² widespread use of novel sterilization modalities to allow PPE reuse,³ and one attempt to purchase price-gouged PPE from the host of the show “Shark Tank.”⁴ In some cases, hospitals and healthcare workers have pleaded for PPE on fundraising and social media sites.⁵

These profound deviations from operations of contemporary health system supply chains would have seemed beyond belief just a few months ago. Instead, they now echo the collective experiences of healthcare stakeholders trying to obtain PPE to protect their frontline healthcare workers during the COVID-19 pandemic.

How did we get into this situation? The manufacture of medical supplies like gowns and masks is a highly competitive business with very slim margins, and as a result, medical equipment manufacturers aim to match their supply with the market’s demand, with hospitals and health systems using just-in-time ordering to limit excess inventory.⁶ While this approach adds efficiency and reduces costs, it also renders manufacturers and customers vulnerable to supply disruptions and shortages when need surges. The COVID-19 pandemic represents perhaps the most extreme example of a massive, widespread surge in demand that occurred multifocally and in a highly compressed time frame. Unlike other industries (eg, consumer paper products), however, in which demand exceeding supply causes inconvenience, the lack of PPE has led to critical public health consequences, with lives of both healthcare workers and vulnerable patients lost because of these shortages of medical equipment.

There are many reasons for the PPE crisis. As noted above, manufacturers have prioritized efficiency over the ability to quickly increase production. They adhere to just-in-time ordering rather than planning for a surge in demand with extra production capacity, all to avoid having warehouses filled with unsold products if surges never occur. This strategy, compounded by the fact that most PPE in the United States is imported from areas in Asia that were profoundly affected early on by COVID-19, led to the observed widespread shortages. When PPE became unavailable from usual suppliers, hospitals were unable to locate other sources of existing PPE because of a lack of transparency about where PPE could be found and how it could be obtained. The Food and Drug Administration and other federal regulatory agencies maintained strict regulations around PPE production and, despite the crisis, made few exceptions.⁷ The FDA did grant a few Emergency Use Authorizations (EUAs) for certain improvised, decontaminated, or alternative respirators (eg, the Chinese-made KN95), but it has only very infrequently issued EUAs to allow domestic manufacturers to ramp up production.⁸ These failures were accompanied by an serious increase in PPE use, leading to spikes in price, price gouging, and hoarding,⁹ problems that were further magnified as health systems and hospitals were forced to compete with nonhealthcare businesses for PPE.

The Defense Production Act (DPA) gives the federal government the power to increase production of goods needed during a crisis⁸ to offer purchasing guarantees, coordinate federal agencies, and regulate distribution and pricing. However, the current administration’s failure to mount a coordinated federal response has contributed to the observed market instability, medical supply shortages, and public health crisis we face. We have previously recommended that the federal government use the power of the DPA to reduce manufacturers’ risk of being uncompensated for excess supply, support temporary reductions in regulatory barriers, and create mandatory centralized reporting of PPE supply, including completed PPE and its components.¹⁰ We stand by these recommendations but also acknowledge that hospitals and health systems may be simultaneously considering how to best prepare for future crises and even surges in demand over the next 18 months as the COVID-19 pandemic continues.

1. Encourage mandates at the hospital, health system, and state level regarding minimum inventory levels for essential equipment. Stockpiles are essential for emergency preparedness. In the long term, these sorts of stockpiles are economically infeasible without government help to maintain them. In the near term, however, it is sensible that hospitals and health systems would maintain a minimum of 2 weeks’ worth of PPE to prepare for expected regional spikes in COVID-19 cases. However, a soon-to-published study suggests that over 40% of hospitals had a PPE stockpile of less than 2 weeks.¹¹ Although this survey was conducted at the height of the shortage, it suggests that there is opportunity for improvement.

2. Coordinate efforts among states and health systems to collect and report inventory, regionalize resources, and coordinate their distribution. The best example of this is the seven-state purchasing consortium announced by New York Governor Andrew Cuomo in early May.¹² Unfortunately, since the announcement, there have been few details about whether the states were successful in their effort to reduce prices or to obtain PPE in bulk. Still, hospitals and health systems could join or emulate purchasing collaboratives to allow resources to be better allocated according to need. There are barriers to such collaboratives because the market is currently set up to encourage competition among health systems and hospitals. During the pandemic, however, cooperation has increasingly been favored over competition in science and healthcare delivery. There are also existing hospital purchasing collaboratives (eg, Premier, Inc¹³), which have taken steps to vet suppliers and improve access to PPE, but it is not clear how successful these efforts have been to date.

3. Advocate for strong federal leadership, including support for increased domestic manufacturing; replenishment and maintenance of state and health system stockpiles of PPE, ventilators, and medications; and development of a centrally coordinated PPE allocation and distribution process. While hospitals and health systems may favor remaining as apolitical as possible, the need for a federal response to stabilize the PPE market may be too urgent and necessary to ignore.

As hospitals and health systems prepare for continued surges in COVID-19 cases, they face challenges in providing PPE for frontline clinicians and staff. A federal plan to enhance nimbleness in responding to multifocal, geographic outbreaks and ensure awareness regarding inventory would improve our chances to successfully navigate the next pandemic and optimize the protection of our health workers, patients, and public health. In the absence of such a plan, hospitals should maintain a minimum of 2 weeks’ worth of PPE to prepare for expected regional spikes in COVID-19 cases and should continue to attempt to coordinate efforts among states and health systems to collect and report inventory, regionalize resources, and coordinate their distribution.

The story of the coronavirus disease 2019 (COVID-19) pandemic in the United States has been defined, in part, by a persistent shortage of medical supplies that has made it difficult and dangerous for healthcare workers to care for infected patients. States, health systems, and even individual hospitals are currently competing against one another—sometimes at auction—to obtain personal protective equipment (PPE). This “Wild West” scenario has resulted in bizarre stories involving attempts to obtain PPE. One health system recently described a James Bond–like pursuit of essential PPE, complete with a covert trip to an industrial warehouse, trucks filled with masks but labeled as food delivery vehicles, and an intervention by a United States congressman.¹ Many states have experienced analogous, but still atypical, stories: masks flown in from China using the private jet of a professional sports team owner,² widespread use of novel sterilization modalities to allow PPE reuse,³ and one attempt to purchase price-gouged PPE from the host of the show “Shark Tank.”⁴ In some cases, hospitals and healthcare workers have pleaded for PPE on fundraising and social media sites.⁵

These profound deviations from operations of contemporary health system supply chains would have seemed beyond belief just a few months ago. Instead, they now echo the collective experiences of healthcare stakeholders trying to obtain PPE to protect their frontline healthcare workers during the COVID-19 pandemic.

How did we get into this situation? The manufacture of medical supplies like gowns and masks is a highly competitive business with very slim margins, and as a result, medical equipment manufacturers aim to match their supply with the market’s demand, with hospitals and health systems using just-in-time ordering to limit excess inventory.⁶ While this approach adds efficiency and reduces costs, it also renders manufacturers and customers vulnerable to supply disruptions and shortages when need surges. The COVID-19 pandemic represents perhaps the most extreme example of a massive, widespread surge in demand that occurred multifocally and in a highly compressed time frame. Unlike other industries (eg, consumer paper products), however, in which demand exceeding supply causes inconvenience, the lack of PPE has led to critical public health consequences, with lives of both healthcare workers and vulnerable patients lost because of these shortages of medical equipment.

There are many reasons for the PPE crisis. As noted above, manufacturers have prioritized efficiency over the ability to quickly increase production. They adhere to just-in-time ordering rather than planning for a surge in demand with extra production capacity, all to avoid having warehouses filled with unsold products if surges never occur. This strategy, compounded by the fact that most PPE in the United States is imported from areas in Asia that were profoundly affected early on by COVID-19, led to the observed widespread shortages. When PPE became unavailable from usual suppliers, hospitals were unable to locate other sources of existing PPE because of a lack of transparency about where PPE could be found and how it could be obtained. The Food and Drug Administration and other federal regulatory agencies maintained strict regulations around PPE production and, despite the crisis, made few exceptions.⁷ The FDA did grant a few Emergency Use Authorizations (EUAs) for certain improvised, decontaminated, or alternative respirators (eg, the Chinese-made KN95), but it has only very infrequently issued EUAs to allow domestic manufacturers to ramp up production.⁸ These failures were accompanied by an serious increase in PPE use, leading to spikes in price, price gouging, and hoarding,⁹ problems that were further magnified as health systems and hospitals were forced to compete with nonhealthcare businesses for PPE.

The Defense Production Act (DPA) gives the federal government the power to increase production of goods needed during a crisis⁸ to offer purchasing guarantees, coordinate federal agencies, and regulate distribution and pricing. However, the current administration’s failure to mount a coordinated federal response has contributed to the observed market instability, medical supply shortages, and public health crisis we face. We have previously recommended that the federal government use the power of the DPA to reduce manufacturers’ risk of being uncompensated for excess supply, support temporary reductions in regulatory barriers, and create mandatory centralized reporting of PPE supply, including completed PPE and its components.¹⁰ We stand by these recommendations but also acknowledge that hospitals and health systems may be simultaneously considering how to best prepare for future crises and even surges in demand over the next 18 months as the COVID-19 pandemic continues.

1. Encourage mandates at the hospital, health system, and state level regarding minimum inventory levels for essential equipment. Stockpiles are essential for emergency preparedness. In the long term, these sorts of stockpiles are economically infeasible without government help to maintain them. In the near term, however, it is sensible that hospitals and health systems would maintain a minimum of 2 weeks’ worth of PPE to prepare for expected regional spikes in COVID-19 cases. However, a soon-to-published study suggests that over 40% of hospitals had a PPE stockpile of less than 2 weeks.¹¹ Although this survey was conducted at the height of the shortage, it suggests that there is opportunity for improvement.

2. Coordinate efforts among states and health systems to collect and report inventory, regionalize resources, and coordinate their distribution. The best example of this is the seven-state purchasing consortium announced by New York Governor Andrew Cuomo in early May.¹² Unfortunately, since the announcement, there have been few details about whether the states were successful in their effort to reduce prices or to obtain PPE in bulk. Still, hospitals and health systems could join or emulate purchasing collaboratives to allow resources to be better allocated according to need. There are barriers to such collaboratives because the market is currently set up to encourage competition among health systems and hospitals. During the pandemic, however, cooperation has increasingly been favored over competition in science and healthcare delivery. There are also existing hospital purchasing collaboratives (eg, Premier, Inc¹³), which have taken steps to vet suppliers and improve access to PPE, but it is not clear how successful these efforts have been to date.

3. Advocate for strong federal leadership, including support for increased domestic manufacturing; replenishment and maintenance of state and health system stockpiles of PPE, ventilators, and medications; and development of a centrally coordinated PPE allocation and distribution process. While hospitals and health systems may favor remaining as apolitical as possible, the need for a federal response to stabilize the PPE market may be too urgent and necessary to ignore.

As hospitals and health systems prepare for continued surges in COVID-19 cases, they face challenges in providing PPE for frontline clinicians and staff. A federal plan to enhance nimbleness in responding to multifocal, geographic outbreaks and ensure awareness regarding inventory would improve our chances to successfully navigate the next pandemic and optimize the protection of our health workers, patients, and public health. In the absence of such a plan, hospitals should maintain a minimum of 2 weeks’ worth of PPE to prepare for expected regional spikes in COVID-19 cases and should continue to attempt to coordinate efforts among states and health systems to collect and report inventory, regionalize resources, and coordinate their distribution.

When my grandfather, who speaks limited English, was admitted to a hospital following a stroke amid the coronavirus disease 2019 (COVID-19) pandemic, my family was understandably worried. Sure enough, within just hours of his admission, we were told our normally very calm and beloved Nana was experiencing significant agitation and delirium. He did not understand nurses’ efforts to calm him down, became even more confused, and was eventually sedated and placed in physical restraints. Even though my family’s presence might have prevented some or all of this terrible series of events, the hospital’s visiting policies during the wave of COVID-19 admissions meant that we were forced to wait in the parking lot as they transpired. The hospital’s policy at the time only allowed visitors for pediatrics, end-of-life care, or labor, not for patients with delirium or altered mental status. We were given the option to make a video call, but my grandfather’s stroke had almost completely taken away his vision. Instead of sitting by his side, comforting him, providing explanations in voices he knew and a language he understood, we were left imagining how difficult it must be to suddenly wake up in an unfamiliar environment, with strangers speaking a different language, limited vision, and your arms and legs tied. Intellectually, I understood the hospital’s goals to minimize transmission, but spiritually and emotionally, it felt very cruel and very wrong.

The next day, we successfully petitioned administration to make an exception for one visitor. We argued that our presence would allow for removal of the sedation and restraints. The clinical team agreed that video calls were insufficient in his situation; he was allowed a visitor. We decided that it should be my mother. As soon my grandfather heard her familiar voice, there was a dramatic improvement. He immediately became calmer and restraints were no longer necessary. The team was grateful for a better physical exam and my grandfather was more cooperative with physical therapy. A few days later, unfortunately, the hospital let us know that they had reevaluated their position on my mother’s visits and that she posed an unnecessary COVID-19 risk to medical staff and other patients. And as soon as she left, my grandfather was again agitated and confused for the remaining 3 days of his hospitalization. Although we are grateful that his delirium resolved once he returned home, delirium also has the potential to lead to long-term cognitive impairment.¹

The COVID-19 pandemic has required hospitals around the world to make difficult decisions about how to balance minimizing disease transmission with continuing to provide compassionate and high-quality patient care. Of these many dilemmas, developing flexible visitor policies is particularly difficult. Currently, the Centers for Disease Control and Prevention and many state health departments encourage limiting visitation in general but recognize the need for exceptions in special circumstances such as in end-of-life settings or altered mental status.^2-4

At the hospital level, there is substantial variation in visitation policies among hospitals. Near our family home in San Jose, California, one hospital currently allows visitation for pediatric patients, pregnant patients, end-of-life patients, surgical patients, and patients in the emergency department, as well as those with mental disabilities or safety needs.⁵ A mere 10 minutes away, another hospital has implemented a very different policy that allows only one visitor for pregnant patients and in end-of-life settings; there are no exceptions for patients with cognitive or physical disabilities.⁶ Other hospitals in the United States have gone even further, not permitting visitors even for those at the end of life.⁷ These patients are forced to spend their last few moments alone.

From an infection control perspective, there are certainly valid reasons to limit visitation. Even with temperature screenings, any movement into and out of a hospital poses a risk of transmitting disease. Infected but asymptomatic persons are known to transmit the disease. Additionally, hospitals still treat non–COVID-19 patients who are most susceptible to severe illness should they develop COVID-19 infection. Early in the COVID-19 pandemic, limitations in testing capacity, personal protective equipment (PPE), and staffing made it challenging to ensure safe visitation. In many cases, it was almost impossible to mitigate the transmission risk that visitors posed. Because many hospitals did not have the capacity to test all symptomatic patients, they could not reliably limit visits to COVID-19–positive patients. Additionally, without enough PPE for healthcare workers, hospitals could not afford for visitors to use additional PPE.

Now that testing is more readily available and some aspects of the PPE shortages have been addressed, we should not forget that visitation has significant benefits for both patients’ psychological well-being and their overall outcomes.⁸ Putting aside the emotional support that the physical presence of loved ones can offer, a large body of research indicates that allowing visitors can also meaningfully improve other important patient outcomes. Specifically, the presence of visitors is associated with less fear,⁹ reduced delirium,¹⁰ and even faster recovery.⁸ In many cases, family members can also help improve hospital safety surveillance and catch medical errors.¹¹

I saw firsthand how these benefits are particularly true for visitors who are also a patient’s primary caretaker. When my mother visited my grandfather, she was not simply a visitor but instead served as an active member of the care team. In addition to providing emotional comfort, my mother oriented him to his surroundings, successfully encouraged oral intake, and even caught some medication errors. Particularly for patients with cognitive impairment, caretakers know the patient better than anyone on the clinical team, and their absence can negatively affect the quality of care.

As a family member who also has familiarity with the healthcare system, I share hospitals’ concerns about wanting to minimize disease transmission. I recognize that, even with PPE and screenings, there is still a chance that visitors unknowingly spread COVID-19 to others in the hospital. On a personal level, however, it feels inhumane to maintain this policy even when it affects particularly vulnerable patients like my grandfather. As some hospitals are already doing,¹² we can take steps to allow visitors for such patients while minimizing the likelihood of COVID-19 disease transmission from visitors. Arriving visitors can be screened and required to wear PPE. While these measures may not eliminate the risk of COVID-19 transmission from visitors, they will likely reduce it significantly when implemented properly and make possible a more humane experience for all.¹³

Fortunately, my grandfather is now recovering comfortably at home, surrounded by his loved ones. To this day, however, he has not forgotten what it was like to be confused and alone in the hospital after his stroke. Even with loved ones around, a stroke is a profoundly distressing experience. To go through such an experience alone is even worse. Because of our petitioning, my grandfather was at least allowed a visitor for part of his stay. Other patients are not even allowed that. As we plan for the pandemic’s next waves, hospitals should reevaluate their visitor policies to ensure that their most vulnerable patients do not have to suffer alone.

The author sincerely thanks Dr Allan Goroll (Massachusetts General Hospital/Harvard Medical School) for his mentorship and critical review of this manuscript.

The author has nothing to disclose.

When my grandfather, who speaks limited English, was admitted to a hospital following a stroke amid the coronavirus disease 2019 (COVID-19) pandemic, my family was understandably worried. Sure enough, within just hours of his admission, we were told our normally very calm and beloved Nana was experiencing significant agitation and delirium. He did not understand nurses’ efforts to calm him down, became even more confused, and was eventually sedated and placed in physical restraints. Even though my family’s presence might have prevented some or all of this terrible series of events, the hospital’s visiting policies during the wave of COVID-19 admissions meant that we were forced to wait in the parking lot as they transpired. The hospital’s policy at the time only allowed visitors for pediatrics, end-of-life care, or labor, not for patients with delirium or altered mental status. We were given the option to make a video call, but my grandfather’s stroke had almost completely taken away his vision. Instead of sitting by his side, comforting him, providing explanations in voices he knew and a language he understood, we were left imagining how difficult it must be to suddenly wake up in an unfamiliar environment, with strangers speaking a different language, limited vision, and your arms and legs tied. Intellectually, I understood the hospital’s goals to minimize transmission, but spiritually and emotionally, it felt very cruel and very wrong.

The next day, we successfully petitioned administration to make an exception for one visitor. We argued that our presence would allow for removal of the sedation and restraints. The clinical team agreed that video calls were insufficient in his situation; he was allowed a visitor. We decided that it should be my mother. As soon my grandfather heard her familiar voice, there was a dramatic improvement. He immediately became calmer and restraints were no longer necessary. The team was grateful for a better physical exam and my grandfather was more cooperative with physical therapy. A few days later, unfortunately, the hospital let us know that they had reevaluated their position on my mother’s visits and that she posed an unnecessary COVID-19 risk to medical staff and other patients. And as soon as she left, my grandfather was again agitated and confused for the remaining 3 days of his hospitalization. Although we are grateful that his delirium resolved once he returned home, delirium also has the potential to lead to long-term cognitive impairment.¹

The COVID-19 pandemic has required hospitals around the world to make difficult decisions about how to balance minimizing disease transmission with continuing to provide compassionate and high-quality patient care. Of these many dilemmas, developing flexible visitor policies is particularly difficult. Currently, the Centers for Disease Control and Prevention and many state health departments encourage limiting visitation in general but recognize the need for exceptions in special circumstances such as in end-of-life settings or altered mental status.^2-4

At the hospital level, there is substantial variation in visitation policies among hospitals. Near our family home in San Jose, California, one hospital currently allows visitation for pediatric patients, pregnant patients, end-of-life patients, surgical patients, and patients in the emergency department, as well as those with mental disabilities or safety needs.⁵ A mere 10 minutes away, another hospital has implemented a very different policy that allows only one visitor for pregnant patients and in end-of-life settings; there are no exceptions for patients with cognitive or physical disabilities.⁶ Other hospitals in the United States have gone even further, not permitting visitors even for those at the end of life.⁷ These patients are forced to spend their last few moments alone.

From an infection control perspective, there are certainly valid reasons to limit visitation. Even with temperature screenings, any movement into and out of a hospital poses a risk of transmitting disease. Infected but asymptomatic persons are known to transmit the disease. Additionally, hospitals still treat non–COVID-19 patients who are most susceptible to severe illness should they develop COVID-19 infection. Early in the COVID-19 pandemic, limitations in testing capacity, personal protective equipment (PPE), and staffing made it challenging to ensure safe visitation. In many cases, it was almost impossible to mitigate the transmission risk that visitors posed. Because many hospitals did not have the capacity to test all symptomatic patients, they could not reliably limit visits to COVID-19–positive patients. Additionally, without enough PPE for healthcare workers, hospitals could not afford for visitors to use additional PPE.

Now that testing is more readily available and some aspects of the PPE shortages have been addressed, we should not forget that visitation has significant benefits for both patients’ psychological well-being and their overall outcomes.⁸ Putting aside the emotional support that the physical presence of loved ones can offer, a large body of research indicates that allowing visitors can also meaningfully improve other important patient outcomes. Specifically, the presence of visitors is associated with less fear,⁹ reduced delirium,¹⁰ and even faster recovery.⁸ In many cases, family members can also help improve hospital safety surveillance and catch medical errors.¹¹

I saw firsthand how these benefits are particularly true for visitors who are also a patient’s primary caretaker. When my mother visited my grandfather, she was not simply a visitor but instead served as an active member of the care team. In addition to providing emotional comfort, my mother oriented him to his surroundings, successfully encouraged oral intake, and even caught some medication errors. Particularly for patients with cognitive impairment, caretakers know the patient better than anyone on the clinical team, and their absence can negatively affect the quality of care.

As a family member who also has familiarity with the healthcare system, I share hospitals’ concerns about wanting to minimize disease transmission. I recognize that, even with PPE and screenings, there is still a chance that visitors unknowingly spread COVID-19 to others in the hospital. On a personal level, however, it feels inhumane to maintain this policy even when it affects particularly vulnerable patients like my grandfather. As some hospitals are already doing,¹² we can take steps to allow visitors for such patients while minimizing the likelihood of COVID-19 disease transmission from visitors. Arriving visitors can be screened and required to wear PPE. While these measures may not eliminate the risk of COVID-19 transmission from visitors, they will likely reduce it significantly when implemented properly and make possible a more humane experience for all.¹³

Fortunately, my grandfather is now recovering comfortably at home, surrounded by his loved ones. To this day, however, he has not forgotten what it was like to be confused and alone in the hospital after his stroke. Even with loved ones around, a stroke is a profoundly distressing experience. To go through such an experience alone is even worse. Because of our petitioning, my grandfather was at least allowed a visitor for part of his stay. Other patients are not even allowed that. As we plan for the pandemic’s next waves, hospitals should reevaluate their visitor policies to ensure that their most vulnerable patients do not have to suffer alone.

The author sincerely thanks Dr Allan Goroll (Massachusetts General Hospital/Harvard Medical School) for his mentorship and critical review of this manuscript.

The author has nothing to disclose.

When my grandfather, who speaks limited English, was admitted to a hospital following a stroke amid the coronavirus disease 2019 (COVID-19) pandemic, my family was understandably worried. Sure enough, within just hours of his admission, we were told our normally very calm and beloved Nana was experiencing significant agitation and delirium. He did not understand nurses’ efforts to calm him down, became even more confused, and was eventually sedated and placed in physical restraints. Even though my family’s presence might have prevented some or all of this terrible series of events, the hospital’s visiting policies during the wave of COVID-19 admissions meant that we were forced to wait in the parking lot as they transpired. The hospital’s policy at the time only allowed visitors for pediatrics, end-of-life care, or labor, not for patients with delirium or altered mental status. We were given the option to make a video call, but my grandfather’s stroke had almost completely taken away his vision. Instead of sitting by his side, comforting him, providing explanations in voices he knew and a language he understood, we were left imagining how difficult it must be to suddenly wake up in an unfamiliar environment, with strangers speaking a different language, limited vision, and your arms and legs tied. Intellectually, I understood the hospital’s goals to minimize transmission, but spiritually and emotionally, it felt very cruel and very wrong.

The next day, we successfully petitioned administration to make an exception for one visitor. We argued that our presence would allow for removal of the sedation and restraints. The clinical team agreed that video calls were insufficient in his situation; he was allowed a visitor. We decided that it should be my mother. As soon my grandfather heard her familiar voice, there was a dramatic improvement. He immediately became calmer and restraints were no longer necessary. The team was grateful for a better physical exam and my grandfather was more cooperative with physical therapy. A few days later, unfortunately, the hospital let us know that they had reevaluated their position on my mother’s visits and that she posed an unnecessary COVID-19 risk to medical staff and other patients. And as soon as she left, my grandfather was again agitated and confused for the remaining 3 days of his hospitalization. Although we are grateful that his delirium resolved once he returned home, delirium also has the potential to lead to long-term cognitive impairment.¹

The COVID-19 pandemic has required hospitals around the world to make difficult decisions about how to balance minimizing disease transmission with continuing to provide compassionate and high-quality patient care. Of these many dilemmas, developing flexible visitor policies is particularly difficult. Currently, the Centers for Disease Control and Prevention and many state health departments encourage limiting visitation in general but recognize the need for exceptions in special circumstances such as in end-of-life settings or altered mental status.^2-4

At the hospital level, there is substantial variation in visitation policies among hospitals. Near our family home in San Jose, California, one hospital currently allows visitation for pediatric patients, pregnant patients, end-of-life patients, surgical patients, and patients in the emergency department, as well as those with mental disabilities or safety needs.⁵ A mere 10 minutes away, another hospital has implemented a very different policy that allows only one visitor for pregnant patients and in end-of-life settings; there are no exceptions for patients with cognitive or physical disabilities.⁶ Other hospitals in the United States have gone even further, not permitting visitors even for those at the end of life.⁷ These patients are forced to spend their last few moments alone.

From an infection control perspective, there are certainly valid reasons to limit visitation. Even with temperature screenings, any movement into and out of a hospital poses a risk of transmitting disease. Infected but asymptomatic persons are known to transmit the disease. Additionally, hospitals still treat non–COVID-19 patients who are most susceptible to severe illness should they develop COVID-19 infection. Early in the COVID-19 pandemic, limitations in testing capacity, personal protective equipment (PPE), and staffing made it challenging to ensure safe visitation. In many cases, it was almost impossible to mitigate the transmission risk that visitors posed. Because many hospitals did not have the capacity to test all symptomatic patients, they could not reliably limit visits to COVID-19–positive patients. Additionally, without enough PPE for healthcare workers, hospitals could not afford for visitors to use additional PPE.

Now that testing is more readily available and some aspects of the PPE shortages have been addressed, we should not forget that visitation has significant benefits for both patients’ psychological well-being and their overall outcomes.⁸ Putting aside the emotional support that the physical presence of loved ones can offer, a large body of research indicates that allowing visitors can also meaningfully improve other important patient outcomes. Specifically, the presence of visitors is associated with less fear,⁹ reduced delirium,¹⁰ and even faster recovery.⁸ In many cases, family members can also help improve hospital safety surveillance and catch medical errors.¹¹

I saw firsthand how these benefits are particularly true for visitors who are also a patient’s primary caretaker. When my mother visited my grandfather, she was not simply a visitor but instead served as an active member of the care team. In addition to providing emotional comfort, my mother oriented him to his surroundings, successfully encouraged oral intake, and even caught some medication errors. Particularly for patients with cognitive impairment, caretakers know the patient better than anyone on the clinical team, and their absence can negatively affect the quality of care.

As a family member who also has familiarity with the healthcare system, I share hospitals’ concerns about wanting to minimize disease transmission. I recognize that, even with PPE and screenings, there is still a chance that visitors unknowingly spread COVID-19 to others in the hospital. On a personal level, however, it feels inhumane to maintain this policy even when it affects particularly vulnerable patients like my grandfather. As some hospitals are already doing,¹² we can take steps to allow visitors for such patients while minimizing the likelihood of COVID-19 disease transmission from visitors. Arriving visitors can be screened and required to wear PPE. While these measures may not eliminate the risk of COVID-19 transmission from visitors, they will likely reduce it significantly when implemented properly and make possible a more humane experience for all.¹³

Fortunately, my grandfather is now recovering comfortably at home, surrounded by his loved ones. To this day, however, he has not forgotten what it was like to be confused and alone in the hospital after his stroke. Even with loved ones around, a stroke is a profoundly distressing experience. To go through such an experience alone is even worse. Because of our petitioning, my grandfather was at least allowed a visitor for part of his stay. Other patients are not even allowed that. As we plan for the pandemic’s next waves, hospitals should reevaluate their visitor policies to ensure that their most vulnerable patients do not have to suffer alone.

The author sincerely thanks Dr Allan Goroll (Massachusetts General Hospital/Harvard Medical School) for his mentorship and critical review of this manuscript.

The author has nothing to disclose.

The coronavirus disease of 2019 (COVID-19) pandemic has created innumerable challenges on scales both global and personal while straining health systems and their personnel. Hospitalists and hospital medicine groups are experiencing unique burdens as they confront the pandemic on the frontlines. Hospital medicine groups are being challenged by the rapid operational changes necessary in preparing for and caring for patients with COVID-19. These challenges include drafting new diagnostic and management algorithms, establishing and enacting policies on personal protective equipment (PPE) and patient and provider testing, modifying staffing protocols including deploying staff to new roles or integrating non-hospitalists into hospital medicine roles, and developing capacity for patient surges¹—all in the setting of uncertainty about how the pandemic may affect individual hospitals or health systems and how long these repercussions may last. In this perspective, we describe key lessons we have learned in leading our hospital medicine group during the COVID-19 pandemic: how to apply emotional intelligence to proactively address the emotional effects of the crisis.

In the early days of the COVID-19 pandemic, the evolving knowledge of the disease process, changing national and local public health guidelines, and instability of the PPE supply chain necessitated rapid change. This pace no longer allowed for our typical time frame of weeks to months for implementation of large-scale operational changes; instead, it demanded adaptation in hours to days. We operated under a strategy of developing new workflows and policies that were logical and reflected the best available information at the time.

For instance, our hospital medicine service cared for some of the earliest-identified COVID-19 patients in the United States in early February 2020. Our initial operational plan for caring for patients with COVID-19 involved grouping these patients on a limited number of direct-care hospitalist teams. The advantages of this approach, which benefitted from low numbers of initial patients, were clear: consolidation of clinical and operational knowledge (including optimal PPE practices) in a few individuals, streamlining communication with infectious diseases specialists and public health departments, and requiring change on only a couple of teams while allowing others to continue their usual workflow. However, we soon learned that providers caring for COVID-19 patients were experiencing an onslaught of negative emotions: fear of contracting the virus themselves or carrying it home to infect loved ones, anxiety of not understanding the clinical disease or having treatments to offer, resentment of having been randomly assigned to the team that would care for these patients, and loneliness of being a sole provider experiencing these emotions. We found ourselves in the position of managing these emotional responses reactively.

To reduce the distress experienced by our hospitalists and to lead more effectively, we realized the need to proactively address the emotional effects that the pandemic was having. Several authors who have written about valuable leadership lessons during this time have noted the importance of acknowledging the emotional tolls of such a crisis and creating venues for hospitalists to share their experiences.^1-4 However, solely adding “wellness” as a checklist item for leaders to address fails to capture the nuances of the complex human emotions that hospitalists may endure at this time and how these emotions influence frontline hospitalists’ responses to operational changes. It is critically important for hospital medicine leaders to employ emotional intelligence, defined as “the ability to monitor one’s own and others’ feelings and emotions, to discriminate among them and to use this information to guide one’s thinking and actions.”^5-7 Integrating emotional intelligence allows hospital medicine leaders to anticipate, identify, articulate, and manage the emotional responses to necessary changes and stresses that occur during a crisis such as the COVID-19 pandemic.

As we applied principles of emotional intelligence to our leadership response to the COVID crisis, we found the following seven techniques effective (Appendix Table):

1. ASK. Leaders should ask individual hospitalists “How are you feeling?” instead of “How are you doing?” or “How can I help?” This question may feel too intimate for some, or leaders may worry that the question feels patronizing; however, in our experience, hospitalists respond positively to this prompt, welcome the opportunity to communicate their feelings, and value being heard. Moreover, when hospitalists feel overwhelmed, they may not be able to determine what help they do or do not need. By understanding the emotions of frontline hospitalists, leaders may be better able to address those emotions directly, find solutions to problems, and anticipate reactions to future policies.⁴

2. SHARE. Leaders should model what they ask of frontline hospitalists and share their own feelings, even if they are experiencing mixed or negative emotions. For instance, a leader who is feeling saddened about the death of a patient can begin a meeting by sharing this sentiment. By allowing themselves to display vulnerability, leaders demonstrate courage and promote a culture of openness, honesty, and mutual trust.

3. INITIATE. Leaders should embrace difficult conversations and be transparent about uncertainty, although they may not have the answers and may need to take local responsibility for consequences of decisions made externally, such as those made by the health system or government. Confronting difficult discussions and being transparent about “unknowns” provides acknowledgement, reassurance, and shared experience that expresses to the hospitalist group that, while the future may be unsettled, they will face it together.

4. ANTICIPATE. Leaders should anticipate the emotional responses to operational changes while designing them and rolling them out. While negative emotions may heavily outweigh positive emotions in times of crisis, we have also found that harnessing positive emotions when designing operational initiatives can assist with successful implementation. For example, by surveying our hospitalists, we found that many felt enthusiastic about caring for patients with COVID-19, curious about new skill sets, and passionate about helping in a time of crisis. By generating a list of these hospitalists up front, we were able to preferentially staff COVID-19 teams with providers who were eager to care for those patients and, thereby, minimize anxiety among those who were more apprehensive.

5. ENCOURAGE. Leaders should provide time and space (including virtually) for hospitalists to discuss their emotions.⁸ We found that creating multiple layers of opportunity for expression allows for engagement with a wider range of hospitalists, some of whom may be reluctant to share feelings openly or to a group, whereas others may enjoy the opportunity to reveal their feelings publicly. These varied venues for emotional expression may range from brief individual check-ins to open “office hours” to dedicated meetings such as “Hospitalist Town Halls.” For instance, spending the first few minutes of a meeting with a smaller group by encouraging each participant to share something personal can build community and mutual understanding, as well as cue leaders in to where participants may be on the emotional landscape.

6. NURTURE. Beyond inviting the expression of emotions, leaders should ensure that hospitalists have access to more formal systems of support, especially for hospitalists who may be experiencing more intense negative emotions. Support may be provided through unit- or team-based debriefing sessions, health-system sponsored support programs, or individual counseling sessions.^4,8

7. APPRECIATE. Leaders should deliberately foster gratitude by sincerely and frequently expressing their appreciation. Because expressing gratitude builds resiliency,⁹ cultivating a culture of gratitude may bolster resilience in the entire hospital medicine group. Opportunities for thankfulness abound as hospitalists volunteer for extra shifts, cover for ill colleagues, participate in new working groups and task forces, and sacrifice their personal safety on the front lines. We often incorporate statements of appreciation into one-on-one conversations with hospitalists, during operational and divisional meetings, and in email. We also built gratitude expressions into the daily work on the Respiratory Isolation Unit at our hospital via daily interdisciplinary huddles for frontline providers to share their experiences and emotions. During huddles, providers are asked to pair negative emotions with suggestions for improvement and to share a moment of gratitude. This helps to engender a spirit of camaraderie, shared mission, and collective optimism.

Hospitalists are experiencing a wide range of emotions related to the COVID-19 pandemic. Hospital medicine leaders must have strategies to understand the emotions providers are experiencing. Being aware of and acknowledging these emotions up front can help leaders plan and implement the operational changes necessary to manage the crisis. Because our health system and city have fortunately been spared the worst of the pandemic so far without large volumes of patients with COVID-19, we recognize that the strategies above may be challenging for leaders in overwhelmed health systems. However, we hope that leaders at all levels can apply the lessons we have learned: to ask hospitalists how they are feeling, share their own feelings, initiate difficult conversations when needed, anticipate the emotional effects of operational changes, encourage expressions of emotion in multiple venues, nurture hospitalists who need more formal support, and appreciate frontline hospitalists. While the emotional needs of hospitalists will undoubtedly change over time as the pandemic evolves, we suspect that these strategies will continue to be important over the coming weeks, months, and longer as we settle into the postpandemic world.

The coronavirus disease of 2019 (COVID-19) pandemic has created innumerable challenges on scales both global and personal while straining health systems and their personnel. Hospitalists and hospital medicine groups are experiencing unique burdens as they confront the pandemic on the frontlines. Hospital medicine groups are being challenged by the rapid operational changes necessary in preparing for and caring for patients with COVID-19. These challenges include drafting new diagnostic and management algorithms, establishing and enacting policies on personal protective equipment (PPE) and patient and provider testing, modifying staffing protocols including deploying staff to new roles or integrating non-hospitalists into hospital medicine roles, and developing capacity for patient surges¹—all in the setting of uncertainty about how the pandemic may affect individual hospitals or health systems and how long these repercussions may last. In this perspective, we describe key lessons we have learned in leading our hospital medicine group during the COVID-19 pandemic: how to apply emotional intelligence to proactively address the emotional effects of the crisis.

In the early days of the COVID-19 pandemic, the evolving knowledge of the disease process, changing national and local public health guidelines, and instability of the PPE supply chain necessitated rapid change. This pace no longer allowed for our typical time frame of weeks to months for implementation of large-scale operational changes; instead, it demanded adaptation in hours to days. We operated under a strategy of developing new workflows and policies that were logical and reflected the best available information at the time.

For instance, our hospital medicine service cared for some of the earliest-identified COVID-19 patients in the United States in early February 2020. Our initial operational plan for caring for patients with COVID-19 involved grouping these patients on a limited number of direct-care hospitalist teams. The advantages of this approach, which benefitted from low numbers of initial patients, were clear: consolidation of clinical and operational knowledge (including optimal PPE practices) in a few individuals, streamlining communication with infectious diseases specialists and public health departments, and requiring change on only a couple of teams while allowing others to continue their usual workflow. However, we soon learned that providers caring for COVID-19 patients were experiencing an onslaught of negative emotions: fear of contracting the virus themselves or carrying it home to infect loved ones, anxiety of not understanding the clinical disease or having treatments to offer, resentment of having been randomly assigned to the team that would care for these patients, and loneliness of being a sole provider experiencing these emotions. We found ourselves in the position of managing these emotional responses reactively.

To reduce the distress experienced by our hospitalists and to lead more effectively, we realized the need to proactively address the emotional effects that the pandemic was having. Several authors who have written about valuable leadership lessons during this time have noted the importance of acknowledging the emotional tolls of such a crisis and creating venues for hospitalists to share their experiences.^1-4 However, solely adding “wellness” as a checklist item for leaders to address fails to capture the nuances of the complex human emotions that hospitalists may endure at this time and how these emotions influence frontline hospitalists’ responses to operational changes. It is critically important for hospital medicine leaders to employ emotional intelligence, defined as “the ability to monitor one’s own and others’ feelings and emotions, to discriminate among them and to use this information to guide one’s thinking and actions.”^5-7 Integrating emotional intelligence allows hospital medicine leaders to anticipate, identify, articulate, and manage the emotional responses to necessary changes and stresses that occur during a crisis such as the COVID-19 pandemic.

As we applied principles of emotional intelligence to our leadership response to the COVID crisis, we found the following seven techniques effective (Appendix Table):

1. ASK. Leaders should ask individual hospitalists “How are you feeling?” instead of “How are you doing?” or “How can I help?” This question may feel too intimate for some, or leaders may worry that the question feels patronizing; however, in our experience, hospitalists respond positively to this prompt, welcome the opportunity to communicate their feelings, and value being heard. Moreover, when hospitalists feel overwhelmed, they may not be able to determine what help they do or do not need. By understanding the emotions of frontline hospitalists, leaders may be better able to address those emotions directly, find solutions to problems, and anticipate reactions to future policies.⁴

2. SHARE. Leaders should model what they ask of frontline hospitalists and share their own feelings, even if they are experiencing mixed or negative emotions. For instance, a leader who is feeling saddened about the death of a patient can begin a meeting by sharing this sentiment. By allowing themselves to display vulnerability, leaders demonstrate courage and promote a culture of openness, honesty, and mutual trust.

3. INITIATE. Leaders should embrace difficult conversations and be transparent about uncertainty, although they may not have the answers and may need to take local responsibility for consequences of decisions made externally, such as those made by the health system or government. Confronting difficult discussions and being transparent about “unknowns” provides acknowledgement, reassurance, and shared experience that expresses to the hospitalist group that, while the future may be unsettled, they will face it together.

4. ANTICIPATE. Leaders should anticipate the emotional responses to operational changes while designing them and rolling them out. While negative emotions may heavily outweigh positive emotions in times of crisis, we have also found that harnessing positive emotions when designing operational initiatives can assist with successful implementation. For example, by surveying our hospitalists, we found that many felt enthusiastic about caring for patients with COVID-19, curious about new skill sets, and passionate about helping in a time of crisis. By generating a list of these hospitalists up front, we were able to preferentially staff COVID-19 teams with providers who were eager to care for those patients and, thereby, minimize anxiety among those who were more apprehensive.

5. ENCOURAGE. Leaders should provide time and space (including virtually) for hospitalists to discuss their emotions.⁸ We found that creating multiple layers of opportunity for expression allows for engagement with a wider range of hospitalists, some of whom may be reluctant to share feelings openly or to a group, whereas others may enjoy the opportunity to reveal their feelings publicly. These varied venues for emotional expression may range from brief individual check-ins to open “office hours” to dedicated meetings such as “Hospitalist Town Halls.” For instance, spending the first few minutes of a meeting with a smaller group by encouraging each participant to share something personal can build community and mutual understanding, as well as cue leaders in to where participants may be on the emotional landscape.

6. NURTURE. Beyond inviting the expression of emotions, leaders should ensure that hospitalists have access to more formal systems of support, especially for hospitalists who may be experiencing more intense negative emotions. Support may be provided through unit- or team-based debriefing sessions, health-system sponsored support programs, or individual counseling sessions.^4,8

7. APPRECIATE. Leaders should deliberately foster gratitude by sincerely and frequently expressing their appreciation. Because expressing gratitude builds resiliency,⁹ cultivating a culture of gratitude may bolster resilience in the entire hospital medicine group. Opportunities for thankfulness abound as hospitalists volunteer for extra shifts, cover for ill colleagues, participate in new working groups and task forces, and sacrifice their personal safety on the front lines. We often incorporate statements of appreciation into one-on-one conversations with hospitalists, during operational and divisional meetings, and in email. We also built gratitude expressions into the daily work on the Respiratory Isolation Unit at our hospital via daily interdisciplinary huddles for frontline providers to share their experiences and emotions. During huddles, providers are asked to pair negative emotions with suggestions for improvement and to share a moment of gratitude. This helps to engender a spirit of camaraderie, shared mission, and collective optimism.

Hospitalists are experiencing a wide range of emotions related to the COVID-19 pandemic. Hospital medicine leaders must have strategies to understand the emotions providers are experiencing. Being aware of and acknowledging these emotions up front can help leaders plan and implement the operational changes necessary to manage the crisis. Because our health system and city have fortunately been spared the worst of the pandemic so far without large volumes of patients with COVID-19, we recognize that the strategies above may be challenging for leaders in overwhelmed health systems. However, we hope that leaders at all levels can apply the lessons we have learned: to ask hospitalists how they are feeling, share their own feelings, initiate difficult conversations when needed, anticipate the emotional effects of operational changes, encourage expressions of emotion in multiple venues, nurture hospitalists who need more formal support, and appreciate frontline hospitalists. While the emotional needs of hospitalists will undoubtedly change over time as the pandemic evolves, we suspect that these strategies will continue to be important over the coming weeks, months, and longer as we settle into the postpandemic world.

The coronavirus disease of 2019 (COVID-19) pandemic has created innumerable challenges on scales both global and personal while straining health systems and their personnel. Hospitalists and hospital medicine groups are experiencing unique burdens as they confront the pandemic on the frontlines. Hospital medicine groups are being challenged by the rapid operational changes necessary in preparing for and caring for patients with COVID-19. These challenges include drafting new diagnostic and management algorithms, establishing and enacting policies on personal protective equipment (PPE) and patient and provider testing, modifying staffing protocols including deploying staff to new roles or integrating non-hospitalists into hospital medicine roles, and developing capacity for patient surges¹—all in the setting of uncertainty about how the pandemic may affect individual hospitals or health systems and how long these repercussions may last. In this perspective, we describe key lessons we have learned in leading our hospital medicine group during the COVID-19 pandemic: how to apply emotional intelligence to proactively address the emotional effects of the crisis.

In the early days of the COVID-19 pandemic, the evolving knowledge of the disease process, changing national and local public health guidelines, and instability of the PPE supply chain necessitated rapid change. This pace no longer allowed for our typical time frame of weeks to months for implementation of large-scale operational changes; instead, it demanded adaptation in hours to days. We operated under a strategy of developing new workflows and policies that were logical and reflected the best available information at the time.

For instance, our hospital medicine service cared for some of the earliest-identified COVID-19 patients in the United States in early February 2020. Our initial operational plan for caring for patients with COVID-19 involved grouping these patients on a limited number of direct-care hospitalist teams. The advantages of this approach, which benefitted from low numbers of initial patients, were clear: consolidation of clinical and operational knowledge (including optimal PPE practices) in a few individuals, streamlining communication with infectious diseases specialists and public health departments, and requiring change on only a couple of teams while allowing others to continue their usual workflow. However, we soon learned that providers caring for COVID-19 patients were experiencing an onslaught of negative emotions: fear of contracting the virus themselves or carrying it home to infect loved ones, anxiety of not understanding the clinical disease or having treatments to offer, resentment of having been randomly assigned to the team that would care for these patients, and loneliness of being a sole provider experiencing these emotions. We found ourselves in the position of managing these emotional responses reactively.

To reduce the distress experienced by our hospitalists and to lead more effectively, we realized the need to proactively address the emotional effects that the pandemic was having. Several authors who have written about valuable leadership lessons during this time have noted the importance of acknowledging the emotional tolls of such a crisis and creating venues for hospitalists to share their experiences.^1-4 However, solely adding “wellness” as a checklist item for leaders to address fails to capture the nuances of the complex human emotions that hospitalists may endure at this time and how these emotions influence frontline hospitalists’ responses to operational changes. It is critically important for hospital medicine leaders to employ emotional intelligence, defined as “the ability to monitor one’s own and others’ feelings and emotions, to discriminate among them and to use this information to guide one’s thinking and actions.”^5-7 Integrating emotional intelligence allows hospital medicine leaders to anticipate, identify, articulate, and manage the emotional responses to necessary changes and stresses that occur during a crisis such as the COVID-19 pandemic.

As we applied principles of emotional intelligence to our leadership response to the COVID crisis, we found the following seven techniques effective (Appendix Table):

1. ASK. Leaders should ask individual hospitalists “How are you feeling?” instead of “How are you doing?” or “How can I help?” This question may feel too intimate for some, or leaders may worry that the question feels patronizing; however, in our experience, hospitalists respond positively to this prompt, welcome the opportunity to communicate their feelings, and value being heard. Moreover, when hospitalists feel overwhelmed, they may not be able to determine what help they do or do not need. By understanding the emotions of frontline hospitalists, leaders may be better able to address those emotions directly, find solutions to problems, and anticipate reactions to future policies.⁴

2. SHARE. Leaders should model what they ask of frontline hospitalists and share their own feelings, even if they are experiencing mixed or negative emotions. For instance, a leader who is feeling saddened about the death of a patient can begin a meeting by sharing this sentiment. By allowing themselves to display vulnerability, leaders demonstrate courage and promote a culture of openness, honesty, and mutual trust.

3. INITIATE. Leaders should embrace difficult conversations and be transparent about uncertainty, although they may not have the answers and may need to take local responsibility for consequences of decisions made externally, such as those made by the health system or government. Confronting difficult discussions and being transparent about “unknowns” provides acknowledgement, reassurance, and shared experience that expresses to the hospitalist group that, while the future may be unsettled, they will face it together.

4. ANTICIPATE. Leaders should anticipate the emotional responses to operational changes while designing them and rolling them out. While negative emotions may heavily outweigh positive emotions in times of crisis, we have also found that harnessing positive emotions when designing operational initiatives can assist with successful implementation. For example, by surveying our hospitalists, we found that many felt enthusiastic about caring for patients with COVID-19, curious about new skill sets, and passionate about helping in a time of crisis. By generating a list of these hospitalists up front, we were able to preferentially staff COVID-19 teams with providers who were eager to care for those patients and, thereby, minimize anxiety among those who were more apprehensive.

5. ENCOURAGE. Leaders should provide time and space (including virtually) for hospitalists to discuss their emotions.⁸ We found that creating multiple layers of opportunity for expression allows for engagement with a wider range of hospitalists, some of whom may be reluctant to share feelings openly or to a group, whereas others may enjoy the opportunity to reveal their feelings publicly. These varied venues for emotional expression may range from brief individual check-ins to open “office hours” to dedicated meetings such as “Hospitalist Town Halls.” For instance, spending the first few minutes of a meeting with a smaller group by encouraging each participant to share something personal can build community and mutual understanding, as well as cue leaders in to where participants may be on the emotional landscape.

6. NURTURE. Beyond inviting the expression of emotions, leaders should ensure that hospitalists have access to more formal systems of support, especially for hospitalists who may be experiencing more intense negative emotions. Support may be provided through unit- or team-based debriefing sessions, health-system sponsored support programs, or individual counseling sessions.^4,8

7. APPRECIATE. Leaders should deliberately foster gratitude by sincerely and frequently expressing their appreciation. Because expressing gratitude builds resiliency,⁹ cultivating a culture of gratitude may bolster resilience in the entire hospital medicine group. Opportunities for thankfulness abound as hospitalists volunteer for extra shifts, cover for ill colleagues, participate in new working groups and task forces, and sacrifice their personal safety on the front lines. We often incorporate statements of appreciation into one-on-one conversations with hospitalists, during operational and divisional meetings, and in email. We also built gratitude expressions into the daily work on the Respiratory Isolation Unit at our hospital via daily interdisciplinary huddles for frontline providers to share their experiences and emotions. During huddles, providers are asked to pair negative emotions with suggestions for improvement and to share a moment of gratitude. This helps to engender a spirit of camaraderie, shared mission, and collective optimism.

Hospitalists are experiencing a wide range of emotions related to the COVID-19 pandemic. Hospital medicine leaders must have strategies to understand the emotions providers are experiencing. Being aware of and acknowledging these emotions up front can help leaders plan and implement the operational changes necessary to manage the crisis. Because our health system and city have fortunately been spared the worst of the pandemic so far without large volumes of patients with COVID-19, we recognize that the strategies above may be challenging for leaders in overwhelmed health systems. However, we hope that leaders at all levels can apply the lessons we have learned: to ask hospitalists how they are feeling, share their own feelings, initiate difficult conversations when needed, anticipate the emotional effects of operational changes, encourage expressions of emotion in multiple venues, nurture hospitalists who need more formal support, and appreciate frontline hospitalists. While the emotional needs of hospitalists will undoubtedly change over time as the pandemic evolves, we suspect that these strategies will continue to be important over the coming weeks, months, and longer as we settle into the postpandemic world.

Recent experiences in the transportation industry highlight the importance of getting right the regulation of decision-support systems in high-stakes environments. Two tragic plane crashes resulted in 346 deaths and were deemed, in part, to be related to a cockpit alert system that overwhelmed pilots with multiple notifications.¹ Similarly, a driverless car struck and killed a pedestrian in the street, in part because the car was not programmed to look for humans outside of a crosswalk.² These two bellwether events offer poignant lessons for the healthcare industry in which human lives also depend on decision-support systems.

Clinical decision-support (CDS) systems are computerized applications, often embedded in an electronic health record (EHR), that provide information to clinicians to inform care. Although CDS systems have been used for many years,³ they have never been subjected to any enforcement of formal testing requirements. However, a draft guidance document released in 2019 from the Food and Drug Administration (FDA) outlined new directions for the regulation of CDS systems.⁴ Although the FDA has thus far focused regulatory efforts on predictive systems developed by private manufacturers,^5,6 this new document provides examples of software that would require regulation for CDS systems that hospitals are already using. Thus, this new guidance raises critical questions—will hospitals themselves be evaluated like private manufacturers, be exempted from federal regulation, or require their own specialized regulation? The FDA has not yet clarified its approach to hospitals or hospital-developed CDS systems, which leaves open numerous possibilities in a rapidly evolving regulatory environment.

Although the FDA has officially regulated CDS systems under section 201(h) of the Federal Food, Drug, and Cosmetic Act (1938), only recently has the FDA begun to sketch the shape of its regulatory efforts. This trend to actually regulate CDS systems began with the 21st Century Cures Act (2016) that amended the definition of software systems that qualify as medical devices and outlined criteria under which a system may be exempt from FDA oversight. For example, regulation would not apply to systems that support “population health” or a “healthy lifestyle” or to ones that qualify as “electronic patient records” as long as they do not “interpret or analyze” data within them.⁷ Following the rapid proliferation of many machine learning and other predictive technologies with medical applications, the FDA began the voluntary Digital Health Software Precertification (Pre-Cert) Program in 2017. Through this program, the FDA selected nine companies from more than 100 applicants and certified them across five domains of excellence. Notably, the Pre-Cert Program currently allows for certification of software manufacturers themselves and does not approve or test actual software devices directly. This regulatory pathway will eventually allow manufacturers to apply under a modified premarket review process for individual software as a medical device (SaMD) that use artificial intelligence (AI) and machine learning. In the meantime, however, many hospitals have developed and deployed their own predictive CDS systems that cross the boundaries into the FDA’s purview and, indeed, do “interpret or analyze” data for real-time EHR alerts, population health management, and other applications.

Regulatory oversight for hospitals could provide quality or safety standards where currently there are none. However, such regulations could also interfere with existing local care practices, hinder rapid development of new CDS systems, and may be perceived as interfering in hospital operations. With the current enthusiasm for AI-based technologies and the concurrent lack of evidence to suggest their effectiveness in practice, regulation could also prompt necessary scrutiny of potential harms of CDS systems, an area with even less evidence. At the same time, CDS developers—private or hospital based—may be able to avoid regulation for some devices with well-placed disclaimers about the intended use of the CDS, one of the FDA criteria for determining the degree of oversight. If the FDA were to regulate hospitals or hospital-developed CDS systems, there are several unanswered questions to consider so that such regulations have their intended impact.

First, does the FDA intend to regulate hospitals and hospital-developed software at all? The framework for determining whether a CDS system will be regulated depends on the severity of the clinical scenario, the ability to independently evaluate the model output, and the intended user (Table). Notably, many types of CDS systems that would require regulation under this framework are already commonplace. For example, the FDA intends to regulate software that “identifies patients who may exhibit signs of opioid addiction,” a scenario similar to prediction models already developed at academic hospitals.⁸ The FDA also plans to regulate a software device even if it is not a CDS system if it is “intended to generate an alarm or an alert to notify a caregiver of a life-threatening condition, such as stroke, and the caregiver relies primarily on this alarm or alert to make a treatment decision.” Although there are no published reports of stroke-specific early warning systems in use, analogous nonspecific and sepsis-specific early warning systems to prompt urgent clinical care have been deployed by hospitals directly⁹ and developed for embedding in commercial EHRs.¹⁰ Hospitals need clarification on the FDA’s regulatory intentions for such CDS systems. FDA regulation of hospitals and hospital-developed CDS systems would fill a critical oversight need and potentially strengthen processes to improve safety and effectiveness. But burdensome regulations may also restrain hospitals from tackling complex problems in medicine for which they are uniquely suited.

Such a regulatory environment may be especially prohibitive for safety-net hospitals that could find themselves at a disadvantage in developing their own CDS systems relative to large academic medical centers that are typically endowed with greater resources. Additionally, CDS systems developed at academic medical centers may not generalize well to populations in the community setting, which could further deepen disparities in access to cutting-edge technologies. For example, racial bias in treatment and referral patterns could bias training labels for CDS systems focused on population health management.¹¹ Similarly, the composition of patient skin color in one population may distort predictions of a model in another with a different distribution of skin color, even when the primary outcome of a prediction model is gender.¹² Additional regulatory steps may apply for models that are adapted to new populations or recalibrated across locations and time.¹³ Until there is more data on the clinical impact of such CDS systems, it is unknown how potential differences in evaluation and approval would actually affect clinical outcomes.

Second, would hospitals be eligible for the Pre-Cert program, and if so, would they be held to the same standards as a private technology manufacturer? The domains of excellence required for precertification approval such as “patient safety,” “clinical responsibility,” and “proactive culture” are aligned with the efforts of hospitals that are already overseen and accredited by organizations like the Joint Commission on Accreditation of Healthcare Organizations and the American Nurses Credentialing Center. There is limited motivation for the FDA to be in the business of regulating these aspects of hospital functions. However, while domains like “product quality” and “cybersecurity” may be less familiar to some hospitals, these existing credentialing bodies may be better suited than the FDA to set and enforce standards for hospitals. In contrast, private manufacturers may have deep expertise in these latter domains. Therefore, as with public-private partnerships for the development of predictive radiology applications,¹⁴ synergies between hospitals and manufacturers may also prove useful for obtaining approvals in a competitive marketplace. Simultaneously, such collaborations would continue to raise questions about conflicts of interest and data privacy.

Finally, regardless of how the FDA will regulate hospitals, what will become of predictive CDS systems that fall outside of the FDA’s scope? Hospitals will continue to find themselves in the position of self-regulation without clear guidance. Although the FDA suggests that developers of unregulated CDS systems still follow best practices for software validation and cybersecurity, existing guidance documents in these domains do not cover the full range of concerns relevant to the development, deployment, and oversight of AI-based CDS systems in the clinical domain. Nor do most hospitals have the infrastructure or expertise to oversee their own CDS systems. Disparate recommendations for development, training, and oversight of AI-based medical systems have emerged but have yet to be endorsed by a federal regulatory body or become part of the hospital accreditation process.¹⁵ Optimal local oversight would require a collaboration between clinical experts, hospital operations leaders, statisticians, data scientists, and ethics experts to ensure effectiveness, safety, and fairness.

Hospitals will remain at the forefront of developing and implementing predictive CDS systems. The proposed FDA regulatory framework would mark an important step toward realizing benefit from such systems, but the FDA needs to clarify the requirements for hospitals and hospital-developed CDS systems to ensure reasonable standards that account for their differences from private software manufacturers. Should the FDA choose to focus regulation on private manufacturers only, hospitals leaders may both feel more empowered to develop their own local CDS tools and feel more comfortable buying CDS systems from vendors that have been precertified. This strategy would provide an optimal balance of assurance and flexibility while maintaining quality standards that ultimately improve patient care.

Recent experiences in the transportation industry highlight the importance of getting right the regulation of decision-support systems in high-stakes environments. Two tragic plane crashes resulted in 346 deaths and were deemed, in part, to be related to a cockpit alert system that overwhelmed pilots with multiple notifications.¹ Similarly, a driverless car struck and killed a pedestrian in the street, in part because the car was not programmed to look for humans outside of a crosswalk.² These two bellwether events offer poignant lessons for the healthcare industry in which human lives also depend on decision-support systems.

Clinical decision-support (CDS) systems are computerized applications, often embedded in an electronic health record (EHR), that provide information to clinicians to inform care. Although CDS systems have been used for many years,³ they have never been subjected to any enforcement of formal testing requirements. However, a draft guidance document released in 2019 from the Food and Drug Administration (FDA) outlined new directions for the regulation of CDS systems.⁴ Although the FDA has thus far focused regulatory efforts on predictive systems developed by private manufacturers,^5,6 this new document provides examples of software that would require regulation for CDS systems that hospitals are already using. Thus, this new guidance raises critical questions—will hospitals themselves be evaluated like private manufacturers, be exempted from federal regulation, or require their own specialized regulation? The FDA has not yet clarified its approach to hospitals or hospital-developed CDS systems, which leaves open numerous possibilities in a rapidly evolving regulatory environment.

Although the FDA has officially regulated CDS systems under section 201(h) of the Federal Food, Drug, and Cosmetic Act (1938), only recently has the FDA begun to sketch the shape of its regulatory efforts. This trend to actually regulate CDS systems began with the 21st Century Cures Act (2016) that amended the definition of software systems that qualify as medical devices and outlined criteria under which a system may be exempt from FDA oversight. For example, regulation would not apply to systems that support “population health” or a “healthy lifestyle” or to ones that qualify as “electronic patient records” as long as they do not “interpret or analyze” data within them.⁷ Following the rapid proliferation of many machine learning and other predictive technologies with medical applications, the FDA began the voluntary Digital Health Software Precertification (Pre-Cert) Program in 2017. Through this program, the FDA selected nine companies from more than 100 applicants and certified them across five domains of excellence. Notably, the Pre-Cert Program currently allows for certification of software manufacturers themselves and does not approve or test actual software devices directly. This regulatory pathway will eventually allow manufacturers to apply under a modified premarket review process for individual software as a medical device (SaMD) that use artificial intelligence (AI) and machine learning. In the meantime, however, many hospitals have developed and deployed their own predictive CDS systems that cross the boundaries into the FDA’s purview and, indeed, do “interpret or analyze” data for real-time EHR alerts, population health management, and other applications.

Regulatory oversight for hospitals could provide quality or safety standards where currently there are none. However, such regulations could also interfere with existing local care practices, hinder rapid development of new CDS systems, and may be perceived as interfering in hospital operations. With the current enthusiasm for AI-based technologies and the concurrent lack of evidence to suggest their effectiveness in practice, regulation could also prompt necessary scrutiny of potential harms of CDS systems, an area with even less evidence. At the same time, CDS developers—private or hospital based—may be able to avoid regulation for some devices with well-placed disclaimers about the intended use of the CDS, one of the FDA criteria for determining the degree of oversight. If the FDA were to regulate hospitals or hospital-developed CDS systems, there are several unanswered questions to consider so that such regulations have their intended impact.

First, does the FDA intend to regulate hospitals and hospital-developed software at all? The framework for determining whether a CDS system will be regulated depends on the severity of the clinical scenario, the ability to independently evaluate the model output, and the intended user (Table). Notably, many types of CDS systems that would require regulation under this framework are already commonplace. For example, the FDA intends to regulate software that “identifies patients who may exhibit signs of opioid addiction,” a scenario similar to prediction models already developed at academic hospitals.⁸ The FDA also plans to regulate a software device even if it is not a CDS system if it is “intended to generate an alarm or an alert to notify a caregiver of a life-threatening condition, such as stroke, and the caregiver relies primarily on this alarm or alert to make a treatment decision.” Although there are no published reports of stroke-specific early warning systems in use, analogous nonspecific and sepsis-specific early warning systems to prompt urgent clinical care have been deployed by hospitals directly⁹ and developed for embedding in commercial EHRs.¹⁰ Hospitals need clarification on the FDA’s regulatory intentions for such CDS systems. FDA regulation of hospitals and hospital-developed CDS systems would fill a critical oversight need and potentially strengthen processes to improve safety and effectiveness. But burdensome regulations may also restrain hospitals from tackling complex problems in medicine for which they are uniquely suited.

Such a regulatory environment may be especially prohibitive for safety-net hospitals that could find themselves at a disadvantage in developing their own CDS systems relative to large academic medical centers that are typically endowed with greater resources. Additionally, CDS systems developed at academic medical centers may not generalize well to populations in the community setting, which could further deepen disparities in access to cutting-edge technologies. For example, racial bias in treatment and referral patterns could bias training labels for CDS systems focused on population health management.¹¹ Similarly, the composition of patient skin color in one population may distort predictions of a model in another with a different distribution of skin color, even when the primary outcome of a prediction model is gender.¹² Additional regulatory steps may apply for models that are adapted to new populations or recalibrated across locations and time.¹³ Until there is more data on the clinical impact of such CDS systems, it is unknown how potential differences in evaluation and approval would actually affect clinical outcomes.

Second, would hospitals be eligible for the Pre-Cert program, and if so, would they be held to the same standards as a private technology manufacturer? The domains of excellence required for precertification approval such as “patient safety,” “clinical responsibility,” and “proactive culture” are aligned with the efforts of hospitals that are already overseen and accredited by organizations like the Joint Commission on Accreditation of Healthcare Organizations and the American Nurses Credentialing Center. There is limited motivation for the FDA to be in the business of regulating these aspects of hospital functions. However, while domains like “product quality” and “cybersecurity” may be less familiar to some hospitals, these existing credentialing bodies may be better suited than the FDA to set and enforce standards for hospitals. In contrast, private manufacturers may have deep expertise in these latter domains. Therefore, as with public-private partnerships for the development of predictive radiology applications,¹⁴ synergies between hospitals and manufacturers may also prove useful for obtaining approvals in a competitive marketplace. Simultaneously, such collaborations would continue to raise questions about conflicts of interest and data privacy.

Finally, regardless of how the FDA will regulate hospitals, what will become of predictive CDS systems that fall outside of the FDA’s scope? Hospitals will continue to find themselves in the position of self-regulation without clear guidance. Although the FDA suggests that developers of unregulated CDS systems still follow best practices for software validation and cybersecurity, existing guidance documents in these domains do not cover the full range of concerns relevant to the development, deployment, and oversight of AI-based CDS systems in the clinical domain. Nor do most hospitals have the infrastructure or expertise to oversee their own CDS systems. Disparate recommendations for development, training, and oversight of AI-based medical systems have emerged but have yet to be endorsed by a federal regulatory body or become part of the hospital accreditation process.¹⁵ Optimal local oversight would require a collaboration between clinical experts, hospital operations leaders, statisticians, data scientists, and ethics experts to ensure effectiveness, safety, and fairness.

Hospitals will remain at the forefront of developing and implementing predictive CDS systems. The proposed FDA regulatory framework would mark an important step toward realizing benefit from such systems, but the FDA needs to clarify the requirements for hospitals and hospital-developed CDS systems to ensure reasonable standards that account for their differences from private software manufacturers. Should the FDA choose to focus regulation on private manufacturers only, hospitals leaders may both feel more empowered to develop their own local CDS tools and feel more comfortable buying CDS systems from vendors that have been precertified. This strategy would provide an optimal balance of assurance and flexibility while maintaining quality standards that ultimately improve patient care.

Recent experiences in the transportation industry highlight the importance of getting right the regulation of decision-support systems in high-stakes environments. Two tragic plane crashes resulted in 346 deaths and were deemed, in part, to be related to a cockpit alert system that overwhelmed pilots with multiple notifications.¹ Similarly, a driverless car struck and killed a pedestrian in the street, in part because the car was not programmed to look for humans outside of a crosswalk.² These two bellwether events offer poignant lessons for the healthcare industry in which human lives also depend on decision-support systems.

Clinical decision-support (CDS) systems are computerized applications, often embedded in an electronic health record (EHR), that provide information to clinicians to inform care. Although CDS systems have been used for many years,³ they have never been subjected to any enforcement of formal testing requirements. However, a draft guidance document released in 2019 from the Food and Drug Administration (FDA) outlined new directions for the regulation of CDS systems.⁴ Although the FDA has thus far focused regulatory efforts on predictive systems developed by private manufacturers,^5,6 this new document provides examples of software that would require regulation for CDS systems that hospitals are already using. Thus, this new guidance raises critical questions—will hospitals themselves be evaluated like private manufacturers, be exempted from federal regulation, or require their own specialized regulation? The FDA has not yet clarified its approach to hospitals or hospital-developed CDS systems, which leaves open numerous possibilities in a rapidly evolving regulatory environment.

Although the FDA has officially regulated CDS systems under section 201(h) of the Federal Food, Drug, and Cosmetic Act (1938), only recently has the FDA begun to sketch the shape of its regulatory efforts. This trend to actually regulate CDS systems began with the 21st Century Cures Act (2016) that amended the definition of software systems that qualify as medical devices and outlined criteria under which a system may be exempt from FDA oversight. For example, regulation would not apply to systems that support “population health” or a “healthy lifestyle” or to ones that qualify as “electronic patient records” as long as they do not “interpret or analyze” data within them.⁷ Following the rapid proliferation of many machine learning and other predictive technologies with medical applications, the FDA began the voluntary Digital Health Software Precertification (Pre-Cert) Program in 2017. Through this program, the FDA selected nine companies from more than 100 applicants and certified them across five domains of excellence. Notably, the Pre-Cert Program currently allows for certification of software manufacturers themselves and does not approve or test actual software devices directly. This regulatory pathway will eventually allow manufacturers to apply under a modified premarket review process for individual software as a medical device (SaMD) that use artificial intelligence (AI) and machine learning. In the meantime, however, many hospitals have developed and deployed their own predictive CDS systems that cross the boundaries into the FDA’s purview and, indeed, do “interpret or analyze” data for real-time EHR alerts, population health management, and other applications.

Regulatory oversight for hospitals could provide quality or safety standards where currently there are none. However, such regulations could also interfere with existing local care practices, hinder rapid development of new CDS systems, and may be perceived as interfering in hospital operations. With the current enthusiasm for AI-based technologies and the concurrent lack of evidence to suggest their effectiveness in practice, regulation could also prompt necessary scrutiny of potential harms of CDS systems, an area with even less evidence. At the same time, CDS developers—private or hospital based—may be able to avoid regulation for some devices with well-placed disclaimers about the intended use of the CDS, one of the FDA criteria for determining the degree of oversight. If the FDA were to regulate hospitals or hospital-developed CDS systems, there are several unanswered questions to consider so that such regulations have their intended impact.

First, does the FDA intend to regulate hospitals and hospital-developed software at all? The framework for determining whether a CDS system will be regulated depends on the severity of the clinical scenario, the ability to independently evaluate the model output, and the intended user (Table). Notably, many types of CDS systems that would require regulation under this framework are already commonplace. For example, the FDA intends to regulate software that “identifies patients who may exhibit signs of opioid addiction,” a scenario similar to prediction models already developed at academic hospitals.⁸ The FDA also plans to regulate a software device even if it is not a CDS system if it is “intended to generate an alarm or an alert to notify a caregiver of a life-threatening condition, such as stroke, and the caregiver relies primarily on this alarm or alert to make a treatment decision.” Although there are no published reports of stroke-specific early warning systems in use, analogous nonspecific and sepsis-specific early warning systems to prompt urgent clinical care have been deployed by hospitals directly⁹ and developed for embedding in commercial EHRs.¹⁰ Hospitals need clarification on the FDA’s regulatory intentions for such CDS systems. FDA regulation of hospitals and hospital-developed CDS systems would fill a critical oversight need and potentially strengthen processes to improve safety and effectiveness. But burdensome regulations may also restrain hospitals from tackling complex problems in medicine for which they are uniquely suited.

Such a regulatory environment may be especially prohibitive for safety-net hospitals that could find themselves at a disadvantage in developing their own CDS systems relative to large academic medical centers that are typically endowed with greater resources. Additionally, CDS systems developed at academic medical centers may not generalize well to populations in the community setting, which could further deepen disparities in access to cutting-edge technologies. For example, racial bias in treatment and referral patterns could bias training labels for CDS systems focused on population health management.¹¹ Similarly, the composition of patient skin color in one population may distort predictions of a model in another with a different distribution of skin color, even when the primary outcome of a prediction model is gender.¹² Additional regulatory steps may apply for models that are adapted to new populations or recalibrated across locations and time.¹³ Until there is more data on the clinical impact of such CDS systems, it is unknown how potential differences in evaluation and approval would actually affect clinical outcomes.

Second, would hospitals be eligible for the Pre-Cert program, and if so, would they be held to the same standards as a private technology manufacturer? The domains of excellence required for precertification approval such as “patient safety,” “clinical responsibility,” and “proactive culture” are aligned with the efforts of hospitals that are already overseen and accredited by organizations like the Joint Commission on Accreditation of Healthcare Organizations and the American Nurses Credentialing Center. There is limited motivation for the FDA to be in the business of regulating these aspects of hospital functions. However, while domains like “product quality” and “cybersecurity” may be less familiar to some hospitals, these existing credentialing bodies may be better suited than the FDA to set and enforce standards for hospitals. In contrast, private manufacturers may have deep expertise in these latter domains. Therefore, as with public-private partnerships for the development of predictive radiology applications,¹⁴ synergies between hospitals and manufacturers may also prove useful for obtaining approvals in a competitive marketplace. Simultaneously, such collaborations would continue to raise questions about conflicts of interest and data privacy.

Finally, regardless of how the FDA will regulate hospitals, what will become of predictive CDS systems that fall outside of the FDA’s scope? Hospitals will continue to find themselves in the position of self-regulation without clear guidance. Although the FDA suggests that developers of unregulated CDS systems still follow best practices for software validation and cybersecurity, existing guidance documents in these domains do not cover the full range of concerns relevant to the development, deployment, and oversight of AI-based CDS systems in the clinical domain. Nor do most hospitals have the infrastructure or expertise to oversee their own CDS systems. Disparate recommendations for development, training, and oversight of AI-based medical systems have emerged but have yet to be endorsed by a federal regulatory body or become part of the hospital accreditation process.¹⁵ Optimal local oversight would require a collaboration between clinical experts, hospital operations leaders, statisticians, data scientists, and ethics experts to ensure effectiveness, safety, and fairness.

Hospitals will remain at the forefront of developing and implementing predictive CDS systems. The proposed FDA regulatory framework would mark an important step toward realizing benefit from such systems, but the FDA needs to clarify the requirements for hospitals and hospital-developed CDS systems to ensure reasonable standards that account for their differences from private software manufacturers. Should the FDA choose to focus regulation on private manufacturers only, hospitals leaders may both feel more empowered to develop their own local CDS tools and feel more comfortable buying CDS systems from vendors that have been precertified. This strategy would provide an optimal balance of assurance and flexibility while maintaining quality standards that ultimately improve patient care.

There is growing momentum to hold drug manufacturers accountable for the more than 400,000 US opioid overdose deaths that have occurred since 1999.¹ As state lawsuits against pharmaceutical manufacturers and distributors wind their way through the legal system, hospitals—which may benefit from settlement funds—have been paying close attention. Recently, former Governor John Kasich (R-Ohio), West Virginia University president E. Gordon Gee, and America’s Essential Hospitals argued that adequately compensating hospitals for the costs of being on the crisis’ “front lines” requires prioritizing them as settlement fund recipients.²

Hospitals should be laying the groundwork for how settlement funds might be used. They may consider enhancing some of the most promising, evidence-based services for individuals with opioid use disorders (OUDs), including improving treatment for commonly associated health conditions such as HIV and hepatitis C virus (HCV); expanding ambulatory long-term antibiotic treatment for endocarditis and other intravenous drug use–associated infections; more broadly adopting harm-reduction practices such as naloxone coprescribing; and applying best practices to caring for substance-exposed infants. They could also develop clinical services not already provided, including creating programs for OUD management during pregnancy and initiating medication for OUD in inpatient, emergency department, and ambulatory settings. In short, hospitals play a critical role in engaging people with OUD in treatment at every possible opportunity.³

When considering how to most effectively use opioid settlement funding, hospitals may consider adding or expanding these much-needed clinical services to address opioid-­related harms; however, their efforts should not stop there. Investments made outside hospital walls could have a significant effect on the public’s health, especially if they target social determinants of health. By tackling factors in the pathway to developing OUD, such as lack of meaningful employment, affordable housing, and mental health care, hospitals can move beyond treating the downstream consequences of addiction and toward preventing community-level opioid-related harms. To accomplish this daunting goal, hospitals will need to strengthen existing relationships with community partners and build new ones. Yet in a 2015 study, only 54% of nonprofit hospitals proposed a strategy to address the overdose crisis that involved community partnering.⁴

In this Perspective, we describe the following three strategies hospitals can use to multiply the reach of their opioid settlement funding by addressing root causes of opioid use through primary prevention: (1) supporting economic opportunities in their communities, (2) expanding affordable housing options in surrounding neighborhoods, and (3) building capacity in ambulatory practices and pharmacies to prevent OUD (Table).

Lack of economic opportunity is one of many root causes of opioid use. For example, a recent study found that automotive assembly plant closures were associated with increases in opioid overdose mortality.⁵ To tackle this complex issue, hospitals can play a crucial role in expanding employment and career advancement options for members of their local communities. Specifically, hospitals can do the following:

• Create jobs within the healthcare system and preferentially recruit and hire from surrounding neighborhoods
• Establish structured career development programs to build skills among entry-level healthcare employees
• Award contracts of varying sizes to locally owned businesses
• Employ individuals with lived experience with substance use disorders, such as peer recovery coaches⁶

To illustrate how health systems are investing in enhancing career opportunities for members of their communities, hundreds of institutions have implemented “School at Work,” a 6-month career development program for entry-level healthcare employees.⁷ The hospitals’ Human Resources department trains participants in communication skills, reading and writing, patient safety and satisfaction, medical terminology, and strategies for success and career advancement. Evaluations of this program have demonstrated improved employee outcomes and a favorable return on investment for hospitals.⁸

As “anchor institutions” and large employers in many communities, hospitals can simultaneously enhance their own workforce and offer employment opportunities that can help break the cycle of addiction that commonly traps individuals and families in communities affected by the overdose crisis.

Hospitals are increasingly supporting interventions that fall outside their traditional purview as they seek to improve population health, such as developing safe green outdoor spaces and increasing access to healthy food options by supporting local farmers markets and grocers.⁹ Stable, decent, and affordable housing is critically important to health and well-being,¹⁰ and there is a well-documented association of opioid use disorder and opioid misuse with housing instability.¹¹ Given evidence of improved outcomes with hospital-led housing interventions,¹² a growing number of hospitals are partnering with housing authorities and community groups to help do the following¹³:

• Contribute to supportive housing options
• Provide environmental health assessments, repairs, and renovations
• Buy or develop affordable housing units

Boston Medical Center, where one in four inpatients are experiencing homelessnes and one in three pediatric emergency department patients are housing insecure, provides an example of how a hospital has invested in housing.¹⁴ In 2017, the hospital began a 5-year, $6.5 million investment in community partnerships in surrounding neighborhoods. Instead of building housing units or acting as a landlord, the hospital chose to invest funding in creative ways to increase the pool of affordable housing. It invested $1 million to rehabilitate permanent, supportive housing units for individuals with mental health conditions in a nearby Boston neighborhood and in a housing stabilization program for people with complex medical issues including substance use disorder. It provided resources to a homeless shelter near the hospital and to the Boston Health Care for the Homeless Program, which provides healthcare to individuals with housing instability. It also funded a community wellness advocate based at the hospital, who received training in substance use disorders and served as a liaison between the hospital and the Boston Housing Authority.

Housing instability is just one of the social determinants of health that hospitals have the capacity to address as they consider where to invest their opioid settlement funds.

Finally, hospitals can partner with community ambulatory practices and pharmacies to prevent the progression to problematic opioid use and OUD. Specifically, hospitals can do as follows:

• Provide evidence-based training to community providers on safe prescribing practices for acute and chronic pain management, as well as postoperative, postprocedural, and postpartum pain management
• Support ambulatory providers in expanding office-based mental health treatment through direct care via telemedicine and in building mental health treatment capacity through consultation, continuing medical education, and telementorship (eg, Project ECHO¹⁵)
• Support ambulatory providers to implement risk reduction strategies to prevent initiation of problematic opioid use, particularly among adolescents and young adults
• Partner with local pharmacies to promote point-of-­prescription counseling on the risks and benefits of opioids

Hospitals bring key strengths and resources to these prevention-­oriented partnerships. First, they may have resources available for clinical research, implementation support, program evaluation, and quality improvement, bringing such expertise to partnerships with ambulatory practices and pharmacies. They likely have specific expertise among their staff, including areas such as pain management, obstetric care, pediatrics, and adolescent medicine, and can provide specialists for consultation services or telementoring initiatives. They also can organize continuing medical education and can offer in-service training at local practices and pharmacies.

Project ECHO is one example of telementoring to build capacity among community providers to manage chronic pain and address addiction and other related harms.¹⁶ The Project ECHO model includes virtual sessions with didactic content and case presentations during which specialists mentor community clinicians. Specific to primary prevention, telementoring has been shown to improve access to evidence-based treatment of chronic pain and mental health conditions,^17,18 which could prevent the development of OUD. By equipping community clinicians with tools to prevent the development of problematic opioid use, hospitals can help reduce the downstream burden of OUD and its associated morbidity, mortality, and costs.

The opioid crisis has devastated families, reduced life expectancy in certain communities,¹⁹ and had a substantial financial impact on hospitals—resulting in an estimated $11 billion in costs to US hospitals each year.²⁰ This ongoing crisis is only going to be compounded by the recent emergence of the SARS-CoV-2 virus. Hospital resources are being strained in unprecedented ways, which has required unprecedented responses in order to continue to serve their communities. Supporting economic opportunity, stable housing, and mental health treatment will be challenging in this new environment but has never been more urgently needed. If opioid settlement funds are targeted to US hospitals, they should be held accountable for where funds are spent because they have a unique opportunity to focus on primary prevention in their communities—confronting OUD before it begins.²¹ However, if hospitals use opioid settlement funding only to continue to provide services already offered, or fail to make bold investments in their communities, this public health crisis will continue to strain the resources of those providing clinical care on the front lines.

The authors wish to thank Hilary Peterson of the RAND Corporation for preparing the paper for submission. She was not compensated for her contribution.

The authors report being supported by grants from the National Institute on Drug Abuse of the National Institutes of Health under awards R21DA045212 (Dr Faherty), K23DA045085 (Dr Hadland), L40DA042434 (Dr Hadland), K23DA038720 (Dr Patrick), R01DA045729 (Dr Patrick), and P50DA046351 (Dr Stein). Dr Hadland also reports grant support from the Thrasher Research Fund and the Academic Pediatric Association. The content of this article is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.

There is growing momentum to hold drug manufacturers accountable for the more than 400,000 US opioid overdose deaths that have occurred since 1999.¹ As state lawsuits against pharmaceutical manufacturers and distributors wind their way through the legal system, hospitals—which may benefit from settlement funds—have been paying close attention. Recently, former Governor John Kasich (R-Ohio), West Virginia University president E. Gordon Gee, and America’s Essential Hospitals argued that adequately compensating hospitals for the costs of being on the crisis’ “front lines” requires prioritizing them as settlement fund recipients.²

Hospitals should be laying the groundwork for how settlement funds might be used. They may consider enhancing some of the most promising, evidence-based services for individuals with opioid use disorders (OUDs), including improving treatment for commonly associated health conditions such as HIV and hepatitis C virus (HCV); expanding ambulatory long-term antibiotic treatment for endocarditis and other intravenous drug use–associated infections; more broadly adopting harm-reduction practices such as naloxone coprescribing; and applying best practices to caring for substance-exposed infants. They could also develop clinical services not already provided, including creating programs for OUD management during pregnancy and initiating medication for OUD in inpatient, emergency department, and ambulatory settings. In short, hospitals play a critical role in engaging people with OUD in treatment at every possible opportunity.³

When considering how to most effectively use opioid settlement funding, hospitals may consider adding or expanding these much-needed clinical services to address opioid-­related harms; however, their efforts should not stop there. Investments made outside hospital walls could have a significant effect on the public’s health, especially if they target social determinants of health. By tackling factors in the pathway to developing OUD, such as lack of meaningful employment, affordable housing, and mental health care, hospitals can move beyond treating the downstream consequences of addiction and toward preventing community-level opioid-related harms. To accomplish this daunting goal, hospitals will need to strengthen existing relationships with community partners and build new ones. Yet in a 2015 study, only 54% of nonprofit hospitals proposed a strategy to address the overdose crisis that involved community partnering.⁴

In this Perspective, we describe the following three strategies hospitals can use to multiply the reach of their opioid settlement funding by addressing root causes of opioid use through primary prevention: (1) supporting economic opportunities in their communities, (2) expanding affordable housing options in surrounding neighborhoods, and (3) building capacity in ambulatory practices and pharmacies to prevent OUD (Table).

Lack of economic opportunity is one of many root causes of opioid use. For example, a recent study found that automotive assembly plant closures were associated with increases in opioid overdose mortality.⁵ To tackle this complex issue, hospitals can play a crucial role in expanding employment and career advancement options for members of their local communities. Specifically, hospitals can do the following:

• Create jobs within the healthcare system and preferentially recruit and hire from surrounding neighborhoods
• Establish structured career development programs to build skills among entry-level healthcare employees
• Award contracts of varying sizes to locally owned businesses
• Employ individuals with lived experience with substance use disorders, such as peer recovery coaches⁶

To illustrate how health systems are investing in enhancing career opportunities for members of their communities, hundreds of institutions have implemented “School at Work,” a 6-month career development program for entry-level healthcare employees.⁷ The hospitals’ Human Resources department trains participants in communication skills, reading and writing, patient safety and satisfaction, medical terminology, and strategies for success and career advancement. Evaluations of this program have demonstrated improved employee outcomes and a favorable return on investment for hospitals.⁸

As “anchor institutions” and large employers in many communities, hospitals can simultaneously enhance their own workforce and offer employment opportunities that can help break the cycle of addiction that commonly traps individuals and families in communities affected by the overdose crisis.

Hospitals are increasingly supporting interventions that fall outside their traditional purview as they seek to improve population health, such as developing safe green outdoor spaces and increasing access to healthy food options by supporting local farmers markets and grocers.⁹ Stable, decent, and affordable housing is critically important to health and well-being,¹⁰ and there is a well-documented association of opioid use disorder and opioid misuse with housing instability.¹¹ Given evidence of improved outcomes with hospital-led housing interventions,¹² a growing number of hospitals are partnering with housing authorities and community groups to help do the following¹³:

• Contribute to supportive housing options
• Provide environmental health assessments, repairs, and renovations
• Buy or develop affordable housing units

Boston Medical Center, where one in four inpatients are experiencing homelessnes and one in three pediatric emergency department patients are housing insecure, provides an example of how a hospital has invested in housing.¹⁴ In 2017, the hospital began a 5-year, $6.5 million investment in community partnerships in surrounding neighborhoods. Instead of building housing units or acting as a landlord, the hospital chose to invest funding in creative ways to increase the pool of affordable housing. It invested $1 million to rehabilitate permanent, supportive housing units for individuals with mental health conditions in a nearby Boston neighborhood and in a housing stabilization program for people with complex medical issues including substance use disorder. It provided resources to a homeless shelter near the hospital and to the Boston Health Care for the Homeless Program, which provides healthcare to individuals with housing instability. It also funded a community wellness advocate based at the hospital, who received training in substance use disorders and served as a liaison between the hospital and the Boston Housing Authority.

Housing instability is just one of the social determinants of health that hospitals have the capacity to address as they consider where to invest their opioid settlement funds.

Finally, hospitals can partner with community ambulatory practices and pharmacies to prevent the progression to problematic opioid use and OUD. Specifically, hospitals can do as follows:

• Provide evidence-based training to community providers on safe prescribing practices for acute and chronic pain management, as well as postoperative, postprocedural, and postpartum pain management
• Support ambulatory providers in expanding office-based mental health treatment through direct care via telemedicine and in building mental health treatment capacity through consultation, continuing medical education, and telementorship (eg, Project ECHO¹⁵)
• Support ambulatory providers to implement risk reduction strategies to prevent initiation of problematic opioid use, particularly among adolescents and young adults
• Partner with local pharmacies to promote point-of-­prescription counseling on the risks and benefits of opioids

Hospitals bring key strengths and resources to these prevention-­oriented partnerships. First, they may have resources available for clinical research, implementation support, program evaluation, and quality improvement, bringing such expertise to partnerships with ambulatory practices and pharmacies. They likely have specific expertise among their staff, including areas such as pain management, obstetric care, pediatrics, and adolescent medicine, and can provide specialists for consultation services or telementoring initiatives. They also can organize continuing medical education and can offer in-service training at local practices and pharmacies.

Project ECHO is one example of telementoring to build capacity among community providers to manage chronic pain and address addiction and other related harms.¹⁶ The Project ECHO model includes virtual sessions with didactic content and case presentations during which specialists mentor community clinicians. Specific to primary prevention, telementoring has been shown to improve access to evidence-based treatment of chronic pain and mental health conditions,^17,18 which could prevent the development of OUD. By equipping community clinicians with tools to prevent the development of problematic opioid use, hospitals can help reduce the downstream burden of OUD and its associated morbidity, mortality, and costs.

The opioid crisis has devastated families, reduced life expectancy in certain communities,¹⁹ and had a substantial financial impact on hospitals—resulting in an estimated $11 billion in costs to US hospitals each year.²⁰ This ongoing crisis is only going to be compounded by the recent emergence of the SARS-CoV-2 virus. Hospital resources are being strained in unprecedented ways, which has required unprecedented responses in order to continue to serve their communities. Supporting economic opportunity, stable housing, and mental health treatment will be challenging in this new environment but has never been more urgently needed. If opioid settlement funds are targeted to US hospitals, they should be held accountable for where funds are spent because they have a unique opportunity to focus on primary prevention in their communities—confronting OUD before it begins.²¹ However, if hospitals use opioid settlement funding only to continue to provide services already offered, or fail to make bold investments in their communities, this public health crisis will continue to strain the resources of those providing clinical care on the front lines.

The authors wish to thank Hilary Peterson of the RAND Corporation for preparing the paper for submission. She was not compensated for her contribution.

The authors report being supported by grants from the National Institute on Drug Abuse of the National Institutes of Health under awards R21DA045212 (Dr Faherty), K23DA045085 (Dr Hadland), L40DA042434 (Dr Hadland), K23DA038720 (Dr Patrick), R01DA045729 (Dr Patrick), and P50DA046351 (Dr Stein). Dr Hadland also reports grant support from the Thrasher Research Fund and the Academic Pediatric Association. The content of this article is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.

There is growing momentum to hold drug manufacturers accountable for the more than 400,000 US opioid overdose deaths that have occurred since 1999.¹ As state lawsuits against pharmaceutical manufacturers and distributors wind their way through the legal system, hospitals—which may benefit from settlement funds—have been paying close attention. Recently, former Governor John Kasich (R-Ohio), West Virginia University president E. Gordon Gee, and America’s Essential Hospitals argued that adequately compensating hospitals for the costs of being on the crisis’ “front lines” requires prioritizing them as settlement fund recipients.²

Hospitals should be laying the groundwork for how settlement funds might be used. They may consider enhancing some of the most promising, evidence-based services for individuals with opioid use disorders (OUDs), including improving treatment for commonly associated health conditions such as HIV and hepatitis C virus (HCV); expanding ambulatory long-term antibiotic treatment for endocarditis and other intravenous drug use–associated infections; more broadly adopting harm-reduction practices such as naloxone coprescribing; and applying best practices to caring for substance-exposed infants. They could also develop clinical services not already provided, including creating programs for OUD management during pregnancy and initiating medication for OUD in inpatient, emergency department, and ambulatory settings. In short, hospitals play a critical role in engaging people with OUD in treatment at every possible opportunity.³

When considering how to most effectively use opioid settlement funding, hospitals may consider adding or expanding these much-needed clinical services to address opioid-­related harms; however, their efforts should not stop there. Investments made outside hospital walls could have a significant effect on the public’s health, especially if they target social determinants of health. By tackling factors in the pathway to developing OUD, such as lack of meaningful employment, affordable housing, and mental health care, hospitals can move beyond treating the downstream consequences of addiction and toward preventing community-level opioid-related harms. To accomplish this daunting goal, hospitals will need to strengthen existing relationships with community partners and build new ones. Yet in a 2015 study, only 54% of nonprofit hospitals proposed a strategy to address the overdose crisis that involved community partnering.⁴

In this Perspective, we describe the following three strategies hospitals can use to multiply the reach of their opioid settlement funding by addressing root causes of opioid use through primary prevention: (1) supporting economic opportunities in their communities, (2) expanding affordable housing options in surrounding neighborhoods, and (3) building capacity in ambulatory practices and pharmacies to prevent OUD (Table).

Lack of economic opportunity is one of many root causes of opioid use. For example, a recent study found that automotive assembly plant closures were associated with increases in opioid overdose mortality.⁵ To tackle this complex issue, hospitals can play a crucial role in expanding employment and career advancement options for members of their local communities. Specifically, hospitals can do the following:

• Create jobs within the healthcare system and preferentially recruit and hire from surrounding neighborhoods
• Establish structured career development programs to build skills among entry-level healthcare employees
• Award contracts of varying sizes to locally owned businesses
• Employ individuals with lived experience with substance use disorders, such as peer recovery coaches⁶

To illustrate how health systems are investing in enhancing career opportunities for members of their communities, hundreds of institutions have implemented “School at Work,” a 6-month career development program for entry-level healthcare employees.⁷ The hospitals’ Human Resources department trains participants in communication skills, reading and writing, patient safety and satisfaction, medical terminology, and strategies for success and career advancement. Evaluations of this program have demonstrated improved employee outcomes and a favorable return on investment for hospitals.⁸

As “anchor institutions” and large employers in many communities, hospitals can simultaneously enhance their own workforce and offer employment opportunities that can help break the cycle of addiction that commonly traps individuals and families in communities affected by the overdose crisis.

Hospitals are increasingly supporting interventions that fall outside their traditional purview as they seek to improve population health, such as developing safe green outdoor spaces and increasing access to healthy food options by supporting local farmers markets and grocers.⁹ Stable, decent, and affordable housing is critically important to health and well-being,¹⁰ and there is a well-documented association of opioid use disorder and opioid misuse with housing instability.¹¹ Given evidence of improved outcomes with hospital-led housing interventions,¹² a growing number of hospitals are partnering with housing authorities and community groups to help do the following¹³:

• Contribute to supportive housing options
• Provide environmental health assessments, repairs, and renovations
• Buy or develop affordable housing units

Boston Medical Center, where one in four inpatients are experiencing homelessnes and one in three pediatric emergency department patients are housing insecure, provides an example of how a hospital has invested in housing.¹⁴ In 2017, the hospital began a 5-year, $6.5 million investment in community partnerships in surrounding neighborhoods. Instead of building housing units or acting as a landlord, the hospital chose to invest funding in creative ways to increase the pool of affordable housing. It invested $1 million to rehabilitate permanent, supportive housing units for individuals with mental health conditions in a nearby Boston neighborhood and in a housing stabilization program for people with complex medical issues including substance use disorder. It provided resources to a homeless shelter near the hospital and to the Boston Health Care for the Homeless Program, which provides healthcare to individuals with housing instability. It also funded a community wellness advocate based at the hospital, who received training in substance use disorders and served as a liaison between the hospital and the Boston Housing Authority.

Housing instability is just one of the social determinants of health that hospitals have the capacity to address as they consider where to invest their opioid settlement funds.

Finally, hospitals can partner with community ambulatory practices and pharmacies to prevent the progression to problematic opioid use and OUD. Specifically, hospitals can do as follows:

• Provide evidence-based training to community providers on safe prescribing practices for acute and chronic pain management, as well as postoperative, postprocedural, and postpartum pain management
• Support ambulatory providers in expanding office-based mental health treatment through direct care via telemedicine and in building mental health treatment capacity through consultation, continuing medical education, and telementorship (eg, Project ECHO¹⁵)
• Support ambulatory providers to implement risk reduction strategies to prevent initiation of problematic opioid use, particularly among adolescents and young adults
• Partner with local pharmacies to promote point-of-­prescription counseling on the risks and benefits of opioids

Hospitals bring key strengths and resources to these prevention-­oriented partnerships. First, they may have resources available for clinical research, implementation support, program evaluation, and quality improvement, bringing such expertise to partnerships with ambulatory practices and pharmacies. They likely have specific expertise among their staff, including areas such as pain management, obstetric care, pediatrics, and adolescent medicine, and can provide specialists for consultation services or telementoring initiatives. They also can organize continuing medical education and can offer in-service training at local practices and pharmacies.

Project ECHO is one example of telementoring to build capacity among community providers to manage chronic pain and address addiction and other related harms.¹⁶ The Project ECHO model includes virtual sessions with didactic content and case presentations during which specialists mentor community clinicians. Specific to primary prevention, telementoring has been shown to improve access to evidence-based treatment of chronic pain and mental health conditions,^17,18 which could prevent the development of OUD. By equipping community clinicians with tools to prevent the development of problematic opioid use, hospitals can help reduce the downstream burden of OUD and its associated morbidity, mortality, and costs.

The opioid crisis has devastated families, reduced life expectancy in certain communities,¹⁹ and had a substantial financial impact on hospitals—resulting in an estimated $11 billion in costs to US hospitals each year.²⁰ This ongoing crisis is only going to be compounded by the recent emergence of the SARS-CoV-2 virus. Hospital resources are being strained in unprecedented ways, which has required unprecedented responses in order to continue to serve their communities. Supporting economic opportunity, stable housing, and mental health treatment will be challenging in this new environment but has never been more urgently needed. If opioid settlement funds are targeted to US hospitals, they should be held accountable for where funds are spent because they have a unique opportunity to focus on primary prevention in their communities—confronting OUD before it begins.²¹ However, if hospitals use opioid settlement funding only to continue to provide services already offered, or fail to make bold investments in their communities, this public health crisis will continue to strain the resources of those providing clinical care on the front lines.

The authors wish to thank Hilary Peterson of the RAND Corporation for preparing the paper for submission. She was not compensated for her contribution.

The authors report being supported by grants from the National Institute on Drug Abuse of the National Institutes of Health under awards R21DA045212 (Dr Faherty), K23DA045085 (Dr Hadland), L40DA042434 (Dr Hadland), K23DA038720 (Dr Patrick), R01DA045729 (Dr Patrick), and P50DA046351 (Dr Stein). Dr Hadland also reports grant support from the Thrasher Research Fund and the Academic Pediatric Association. The content of this article is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.

Frailty is a highly prevalent syndrome in nursing homes, occurring in at least 50% of patients.¹ The frailty phenotype has been described by Fried and colleagues as impairment in ≥ 3 of 5 domains: unintentional weight loss, self-reported exhaustion, muscle weakness, slow gait speed, and low physical activity. By this definition, frailty is highly associated with poor quality of life and mortality.^2,3

In recent years, there has been evolving evidence of a relationship between frailty and chronic systemic inflammation.^4-6 Some degree of chronic inflammation is likely inherent to the aging process and increases the risk of frailty (so-called inflammaging) but is seen to a greater degree in many pathologic conditions in nursing homes, including cancer, organ failure, and chronic infection.^4,6-8

Dysphagia also is highly prevalent in nursing homes, affecting up to 60% of patients and is a strong predictor of hospital utilization and of mortality.^9,10 Overt aspiration pneumonitis and pneumonia are perhaps the best studied sequelae, but chronic occult microaspiration also is prevalent in this population.¹¹ Just as normal systemic inflammatory changes in aging may increase vulnerability to frailty with additional illness burden, normal aging changes in swallowing function may increase vulnerability to dysphagia and to microaspiration with additional illness burden.^12,13 In older adults, important risk factors for microaspiration include not only overt dysphagia, dementia, and other neurologic illnesses, but also general debility, weakness, and immobility.¹⁴

Matsuse and colleagues have described diffuse aspiration bronchiolitis (DAB) in patients with chronic microaspiration.¹⁴ DAB often goes undiagnosed.^14-16 As in frailty, weight loss and chronic anemia may be seen, and many of these patients are bedridden.^14,17 Episodes of macroaspiration and overt lobar pneumonia also may occur.¹⁴ Lung biopsy or autopsy reveals chronic bronchiolar inflammation and sometimes pulmonary fibrosis, but to date there have been no reports suggesting chronic systemic inflammation or elevated proinflammatory cytokines.^14,15,17 We present 3 patients with progressive weight loss, functional decline, and frailty in whom chronic microaspiration likely played a significant role.

Case 1 Presentation

A 68-year-old man with a 6-year history of rapidly progressive Parkinson disease was admitted to the Haley’s Cove Community Living Center (CLC) on the James A. Haley Veterans’ Hospital campus in Tampa, Florida for long-term care. The patient’s medical history also was significant for bipolar illness and for small cell carcinoma of the lung in sustained remission.

Medications included levodopa/carbidopa 50 mg/200 mg 4 times daily, entacapone 200 mg 4 times daily, lithium carbonate 600 mg every night at bedtime, lamotrigine 150 mg daily, quetiapine 200 mg every night at bedtime, pravastatin 40 mg every night at bedtime, omeprazole 20 mg daily, tamsulosin 0.4 mg every night at bedtime, and aspirin 81 mg daily. He initially did well, but after 6 months the nursing staff began to notice the patient coughing during and after meals. Speech pathology evaluation revealed moderate oropharyngeal dysphagia, and his diet was downgraded to nectar-thickened liquids.

Over the subsequent 10 months, he became progressively weaker in physical therapy and more inactive, with about a 20-lb weight loss and mild hypoalbuminemia of 3.0 gm/dL. He had developed 3 episodes of aspiration pneumonia during this period; a repeat swallow evaluation after the last episode revealed worsened dysphagia, and his physician suggested nil per os (NPO) status and an alternative feeding route. His guardian declined placement of a percutaneous endoscopic gastrostomy (PEG) tube, he was transferred to the inpatient hospice unit, and died 2 weeks later. An autopsy was declined.

Case 2 Presentation

A 66-year-old man with a medical history of multiple traumatic brain injuries (TBIs) was admitted to the CLC for long-term care. Sequelae of the TBIs included moderate dementia, spastic paraparesis with multiple pressure injuries, a well-controlled seizure disorder, and severe oropharyngeal dysphagia with NPO status and a percutaneous endoscopic gastrostomy (PEG) tube. His medical history included TBIs and hepatitis C virus infection; medications included levetiracetam 1,000 mg twice daily, lamotrigine 25 mg twice daily, and cholecalciferol 2,000 U daily. He had multiple stage III pressure injuries and an ischial stage IV injury at the time of admission.

His 11-month stay in the CLC was characterized by progressively worsening weakness and inactivity, with a 25-lb weight loss in spite of adequate tube feeding. Serum albumin remained in the 2.0 to 2.5 gm/dL range, hemoglobin in the 7 to 9 gm/dL range without any obvious source of anemia. Most of the pressure injuries worsened during his stay in spite of aggressive wound care, and he developed a second stage IV sacral wound. A single C-reactive protein (CRP) level 2 months prior to his death was markedly elevated at 19.5 mg/dL. In spite of maintaining NPO status, he developed 3 episodes of aspiration pneumonia, all of which responded well to treatment. Ultimately, he was found pulseless and apneic and resuscitation was unsuccessful. An autopsy revealed purulent material in the small airways.

Case 3 Presentation

A 65-year-old man with a long history of paranoid schizophrenia and severe gastroesophageal reflux disease had resided in the CLC for about 10 years. Medications included risperidone microspheres 37.5 mg every 2 weeks, valproic acid 500 mg 3 times daily and 1,000 mg every night at bedtime, lansoprazole 30 mg twice daily, ranitidine 150 mg every night at bedtime, sucralfate 1,000 mg 3 times daily, simvastatin 20 mg every night at bedtime, and tamsulosin 0.4 mg every night at bedtime. He had done well for many years but developed some drooling and a modest resting tremor (but no other signs of pseudoparkinsonism) about 8 years after admission.

There had been no changes to his risperidone dosage. He also lost about 20 lb over a period of 1 year and became increasingly weak and dependent in gait, serum albumin dropped as low as 1.6 gm/dL, hemoglobin dropped to the 7 to 8 gm/dL range (without any other obvious source of anemia), and he developed a gradually worsening right-sided pleural effusion. CRP was chronically elevated at this point, in the 6 to 15 mg/dL range and as high as 17.2 mg/dL. Ultimately, he developed 3 episodes of aspiration pneumonia over a period of 2 months. Swallowing evaluation at that time revealed severe oropharyngeal dysphagia and a PEG tube was placed. Due to concerns for possible antipsychotic-induced dysphagia, risperidone was discontinued, and quetiapine 400 mg a day was substituted. He did well over the subsequent year with no further pneumonia and advancement back to a regular diet. He regained all of the lost weight and began independent ambulation. Albumin improved to the 3 gm/dL range, hemoglobin to the 12 to 13 gm/dL range, and CRP had decreased to 0.7 mg/dL. The pleural effusion (believed to have been a parapneumonic effusion) had resolved.

Discussion

All 3 patients met the Fried criteria for frailty, although there were several confounding issues.² All 3 patients lost between 20 and 25 lb; all had clearly become weaker according to nursing and rehabilitation staff (although none were formally assessed for grip strength); and all had clear declines in their activity level. Patient 3 had a clear decrement in gait speed, but patient 1 had severe gait impairment due to Parkinson disease (although his gait in therapy had clearly worsened). Patient 2 was paraparetic and unable to ambulate. There also was evidence of limited biomarkers of systemic inflammation; all 3 patients’ albumin had decreased, and patients 2 and 3 had significant decrease in hemoglobin; but these commonplace clinical biomarkers are obviously multifactorially determined. We have limited data on our patients’ CRP levels; serial levels would have been more specific for systemic inflammation but were infrequently performed on the patients.

Multimorbidity and medical complexity are more the rule than the exception in frail geriatric patients,and it is difficult to separate the role of microaspiration from other confounding conditions that might have contributed to these patients’ evolving systemic inflammation and frailty.¹⁸ It might be argued that the decline for patient 1 was related to the underlying Parkinson disease (a progressive neurologic illness in which systemic inflammation has been reported), or that the decline of patient 2 was related to the worsening pressure injuries rather than to covert microaspiration.¹⁹ However, the TBIs for patient 2 and the schizophrenia for patient 3 would not be expected to be associated with frailty or with systemic inflammation. Furthermore, the frailty symptoms of patient 3 and inflammatory biomarkers improved after the risperidone, which was likely responsible for his microaspiration, was discontinued. All 3 patients were at risk for oropharyngeal dysphagia (antipsychotic medication is clearly associated with dysphagia²⁰); patient 2 demonstrated pathologic evidence of DAB at autopsy.

There is evolving evidence that chronic systemic inflammation and immune activation are key mechanisms in the pathogenesis of frailty.^4-6 It is known that elevated serum levels of proinflammatory cytokines, including tumor necrosis factor-α, interleukin-6, and CRP are directly associated with frailty and are inversely associated with levels of albumin, hemoglobin, insulin-like growth factor-1, and several micronutrients in frail individuals.^4-7,21,22 Chronic inflammation contributes to the pathophysiology of frailty through detrimental effects on a broad range of systems, including the musculoskeletal, endocrine, and hematopoietic systems and through nutritional dysregulation.^2,4,23 These changes may lead to further deleterious effects, creating a downward spiral of worsening frailty. For example, it seems likely that our patients’ progressive weakness further compromised airway protection, creating a vicious cycle of worsening microaspiration and chronic inflammation.

Conclusions

To date, the role of chronic microaspiration and DAB in chronic systemic inflammation or in frailty has not been explored. Given the prevalence of microaspiration in nursing home residents and the devastating consequences of frailty, though, this seems to be a crucial area of investigation. It is equally crucial for long-term care staff, both providers and nursing staff, to have a heightened awareness of covert microaspiration and a low threshold for referral to speech pathology for further investigation. Staff also should be aware of the utility of the Fried criteria to improve identification of frailty in general. It is probable that covert microaspiration will prove to be an important part of the differential diagnosis of frailty.

Frailty is a highly prevalent syndrome in nursing homes, occurring in at least 50% of patients.¹ The frailty phenotype has been described by Fried and colleagues as impairment in ≥ 3 of 5 domains: unintentional weight loss, self-reported exhaustion, muscle weakness, slow gait speed, and low physical activity. By this definition, frailty is highly associated with poor quality of life and mortality.^2,3

In recent years, there has been evolving evidence of a relationship between frailty and chronic systemic inflammation.^4-6 Some degree of chronic inflammation is likely inherent to the aging process and increases the risk of frailty (so-called inflammaging) but is seen to a greater degree in many pathologic conditions in nursing homes, including cancer, organ failure, and chronic infection.^4,6-8

Dysphagia also is highly prevalent in nursing homes, affecting up to 60% of patients and is a strong predictor of hospital utilization and of mortality.^9,10 Overt aspiration pneumonitis and pneumonia are perhaps the best studied sequelae, but chronic occult microaspiration also is prevalent in this population.¹¹ Just as normal systemic inflammatory changes in aging may increase vulnerability to frailty with additional illness burden, normal aging changes in swallowing function may increase vulnerability to dysphagia and to microaspiration with additional illness burden.^12,13 In older adults, important risk factors for microaspiration include not only overt dysphagia, dementia, and other neurologic illnesses, but also general debility, weakness, and immobility.¹⁴

Matsuse and colleagues have described diffuse aspiration bronchiolitis (DAB) in patients with chronic microaspiration.¹⁴ DAB often goes undiagnosed.^14-16 As in frailty, weight loss and chronic anemia may be seen, and many of these patients are bedridden.^14,17 Episodes of macroaspiration and overt lobar pneumonia also may occur.¹⁴ Lung biopsy or autopsy reveals chronic bronchiolar inflammation and sometimes pulmonary fibrosis, but to date there have been no reports suggesting chronic systemic inflammation or elevated proinflammatory cytokines.^14,15,17 We present 3 patients with progressive weight loss, functional decline, and frailty in whom chronic microaspiration likely played a significant role.

Case 1 Presentation

A 68-year-old man with a 6-year history of rapidly progressive Parkinson disease was admitted to the Haley’s Cove Community Living Center (CLC) on the James A. Haley Veterans’ Hospital campus in Tampa, Florida for long-term care. The patient’s medical history also was significant for bipolar illness and for small cell carcinoma of the lung in sustained remission.

Medications included levodopa/carbidopa 50 mg/200 mg 4 times daily, entacapone 200 mg 4 times daily, lithium carbonate 600 mg every night at bedtime, lamotrigine 150 mg daily, quetiapine 200 mg every night at bedtime, pravastatin 40 mg every night at bedtime, omeprazole 20 mg daily, tamsulosin 0.4 mg every night at bedtime, and aspirin 81 mg daily. He initially did well, but after 6 months the nursing staff began to notice the patient coughing during and after meals. Speech pathology evaluation revealed moderate oropharyngeal dysphagia, and his diet was downgraded to nectar-thickened liquids.

Over the subsequent 10 months, he became progressively weaker in physical therapy and more inactive, with about a 20-lb weight loss and mild hypoalbuminemia of 3.0 gm/dL. He had developed 3 episodes of aspiration pneumonia during this period; a repeat swallow evaluation after the last episode revealed worsened dysphagia, and his physician suggested nil per os (NPO) status and an alternative feeding route. His guardian declined placement of a percutaneous endoscopic gastrostomy (PEG) tube, he was transferred to the inpatient hospice unit, and died 2 weeks later. An autopsy was declined.

Case 2 Presentation

A 66-year-old man with a medical history of multiple traumatic brain injuries (TBIs) was admitted to the CLC for long-term care. Sequelae of the TBIs included moderate dementia, spastic paraparesis with multiple pressure injuries, a well-controlled seizure disorder, and severe oropharyngeal dysphagia with NPO status and a percutaneous endoscopic gastrostomy (PEG) tube. His medical history included TBIs and hepatitis C virus infection; medications included levetiracetam 1,000 mg twice daily, lamotrigine 25 mg twice daily, and cholecalciferol 2,000 U daily. He had multiple stage III pressure injuries and an ischial stage IV injury at the time of admission.

His 11-month stay in the CLC was characterized by progressively worsening weakness and inactivity, with a 25-lb weight loss in spite of adequate tube feeding. Serum albumin remained in the 2.0 to 2.5 gm/dL range, hemoglobin in the 7 to 9 gm/dL range without any obvious source of anemia. Most of the pressure injuries worsened during his stay in spite of aggressive wound care, and he developed a second stage IV sacral wound. A single C-reactive protein (CRP) level 2 months prior to his death was markedly elevated at 19.5 mg/dL. In spite of maintaining NPO status, he developed 3 episodes of aspiration pneumonia, all of which responded well to treatment. Ultimately, he was found pulseless and apneic and resuscitation was unsuccessful. An autopsy revealed purulent material in the small airways.

Case 3 Presentation

A 65-year-old man with a long history of paranoid schizophrenia and severe gastroesophageal reflux disease had resided in the CLC for about 10 years. Medications included risperidone microspheres 37.5 mg every 2 weeks, valproic acid 500 mg 3 times daily and 1,000 mg every night at bedtime, lansoprazole 30 mg twice daily, ranitidine 150 mg every night at bedtime, sucralfate 1,000 mg 3 times daily, simvastatin 20 mg every night at bedtime, and tamsulosin 0.4 mg every night at bedtime. He had done well for many years but developed some drooling and a modest resting tremor (but no other signs of pseudoparkinsonism) about 8 years after admission.

There had been no changes to his risperidone dosage. He also lost about 20 lb over a period of 1 year and became increasingly weak and dependent in gait, serum albumin dropped as low as 1.6 gm/dL, hemoglobin dropped to the 7 to 8 gm/dL range (without any other obvious source of anemia), and he developed a gradually worsening right-sided pleural effusion. CRP was chronically elevated at this point, in the 6 to 15 mg/dL range and as high as 17.2 mg/dL. Ultimately, he developed 3 episodes of aspiration pneumonia over a period of 2 months. Swallowing evaluation at that time revealed severe oropharyngeal dysphagia and a PEG tube was placed. Due to concerns for possible antipsychotic-induced dysphagia, risperidone was discontinued, and quetiapine 400 mg a day was substituted. He did well over the subsequent year with no further pneumonia and advancement back to a regular diet. He regained all of the lost weight and began independent ambulation. Albumin improved to the 3 gm/dL range, hemoglobin to the 12 to 13 gm/dL range, and CRP had decreased to 0.7 mg/dL. The pleural effusion (believed to have been a parapneumonic effusion) had resolved.

Discussion

All 3 patients met the Fried criteria for frailty, although there were several confounding issues.² All 3 patients lost between 20 and 25 lb; all had clearly become weaker according to nursing and rehabilitation staff (although none were formally assessed for grip strength); and all had clear declines in their activity level. Patient 3 had a clear decrement in gait speed, but patient 1 had severe gait impairment due to Parkinson disease (although his gait in therapy had clearly worsened). Patient 2 was paraparetic and unable to ambulate. There also was evidence of limited biomarkers of systemic inflammation; all 3 patients’ albumin had decreased, and patients 2 and 3 had significant decrease in hemoglobin; but these commonplace clinical biomarkers are obviously multifactorially determined. We have limited data on our patients’ CRP levels; serial levels would have been more specific for systemic inflammation but were infrequently performed on the patients.

Multimorbidity and medical complexity are more the rule than the exception in frail geriatric patients,and it is difficult to separate the role of microaspiration from other confounding conditions that might have contributed to these patients’ evolving systemic inflammation and frailty.¹⁸ It might be argued that the decline for patient 1 was related to the underlying Parkinson disease (a progressive neurologic illness in which systemic inflammation has been reported), or that the decline of patient 2 was related to the worsening pressure injuries rather than to covert microaspiration.¹⁹ However, the TBIs for patient 2 and the schizophrenia for patient 3 would not be expected to be associated with frailty or with systemic inflammation. Furthermore, the frailty symptoms of patient 3 and inflammatory biomarkers improved after the risperidone, which was likely responsible for his microaspiration, was discontinued. All 3 patients were at risk for oropharyngeal dysphagia (antipsychotic medication is clearly associated with dysphagia²⁰); patient 2 demonstrated pathologic evidence of DAB at autopsy.

There is evolving evidence that chronic systemic inflammation and immune activation are key mechanisms in the pathogenesis of frailty.^4-6 It is known that elevated serum levels of proinflammatory cytokines, including tumor necrosis factor-α, interleukin-6, and CRP are directly associated with frailty and are inversely associated with levels of albumin, hemoglobin, insulin-like growth factor-1, and several micronutrients in frail individuals.^4-7,21,22 Chronic inflammation contributes to the pathophysiology of frailty through detrimental effects on a broad range of systems, including the musculoskeletal, endocrine, and hematopoietic systems and through nutritional dysregulation.^2,4,23 These changes may lead to further deleterious effects, creating a downward spiral of worsening frailty. For example, it seems likely that our patients’ progressive weakness further compromised airway protection, creating a vicious cycle of worsening microaspiration and chronic inflammation.

Conclusions

To date, the role of chronic microaspiration and DAB in chronic systemic inflammation or in frailty has not been explored. Given the prevalence of microaspiration in nursing home residents and the devastating consequences of frailty, though, this seems to be a crucial area of investigation. It is equally crucial for long-term care staff, both providers and nursing staff, to have a heightened awareness of covert microaspiration and a low threshold for referral to speech pathology for further investigation. Staff also should be aware of the utility of the Fried criteria to improve identification of frailty in general. It is probable that covert microaspiration will prove to be an important part of the differential diagnosis of frailty.

Frailty is a highly prevalent syndrome in nursing homes, occurring in at least 50% of patients.¹ The frailty phenotype has been described by Fried and colleagues as impairment in ≥ 3 of 5 domains: unintentional weight loss, self-reported exhaustion, muscle weakness, slow gait speed, and low physical activity. By this definition, frailty is highly associated with poor quality of life and mortality.^2,3

In recent years, there has been evolving evidence of a relationship between frailty and chronic systemic inflammation.^4-6 Some degree of chronic inflammation is likely inherent to the aging process and increases the risk of frailty (so-called inflammaging) but is seen to a greater degree in many pathologic conditions in nursing homes, including cancer, organ failure, and chronic infection.^4,6-8

Dysphagia also is highly prevalent in nursing homes, affecting up to 60% of patients and is a strong predictor of hospital utilization and of mortality.^9,10 Overt aspiration pneumonitis and pneumonia are perhaps the best studied sequelae, but chronic occult microaspiration also is prevalent in this population.¹¹ Just as normal systemic inflammatory changes in aging may increase vulnerability to frailty with additional illness burden, normal aging changes in swallowing function may increase vulnerability to dysphagia and to microaspiration with additional illness burden.^12,13 In older adults, important risk factors for microaspiration include not only overt dysphagia, dementia, and other neurologic illnesses, but also general debility, weakness, and immobility.¹⁴

Matsuse and colleagues have described diffuse aspiration bronchiolitis (DAB) in patients with chronic microaspiration.¹⁴ DAB often goes undiagnosed.^14-16 As in frailty, weight loss and chronic anemia may be seen, and many of these patients are bedridden.^14,17 Episodes of macroaspiration and overt lobar pneumonia also may occur.¹⁴ Lung biopsy or autopsy reveals chronic bronchiolar inflammation and sometimes pulmonary fibrosis, but to date there have been no reports suggesting chronic systemic inflammation or elevated proinflammatory cytokines.^14,15,17 We present 3 patients with progressive weight loss, functional decline, and frailty in whom chronic microaspiration likely played a significant role.

Case 1 Presentation

A 68-year-old man with a 6-year history of rapidly progressive Parkinson disease was admitted to the Haley’s Cove Community Living Center (CLC) on the James A. Haley Veterans’ Hospital campus in Tampa, Florida for long-term care. The patient’s medical history also was significant for bipolar illness and for small cell carcinoma of the lung in sustained remission.

Medications included levodopa/carbidopa 50 mg/200 mg 4 times daily, entacapone 200 mg 4 times daily, lithium carbonate 600 mg every night at bedtime, lamotrigine 150 mg daily, quetiapine 200 mg every night at bedtime, pravastatin 40 mg every night at bedtime, omeprazole 20 mg daily, tamsulosin 0.4 mg every night at bedtime, and aspirin 81 mg daily. He initially did well, but after 6 months the nursing staff began to notice the patient coughing during and after meals. Speech pathology evaluation revealed moderate oropharyngeal dysphagia, and his diet was downgraded to nectar-thickened liquids.

Over the subsequent 10 months, he became progressively weaker in physical therapy and more inactive, with about a 20-lb weight loss and mild hypoalbuminemia of 3.0 gm/dL. He had developed 3 episodes of aspiration pneumonia during this period; a repeat swallow evaluation after the last episode revealed worsened dysphagia, and his physician suggested nil per os (NPO) status and an alternative feeding route. His guardian declined placement of a percutaneous endoscopic gastrostomy (PEG) tube, he was transferred to the inpatient hospice unit, and died 2 weeks later. An autopsy was declined.

Case 2 Presentation

A 66-year-old man with a medical history of multiple traumatic brain injuries (TBIs) was admitted to the CLC for long-term care. Sequelae of the TBIs included moderate dementia, spastic paraparesis with multiple pressure injuries, a well-controlled seizure disorder, and severe oropharyngeal dysphagia with NPO status and a percutaneous endoscopic gastrostomy (PEG) tube. His medical history included TBIs and hepatitis C virus infection; medications included levetiracetam 1,000 mg twice daily, lamotrigine 25 mg twice daily, and cholecalciferol 2,000 U daily. He had multiple stage III pressure injuries and an ischial stage IV injury at the time of admission.

His 11-month stay in the CLC was characterized by progressively worsening weakness and inactivity, with a 25-lb weight loss in spite of adequate tube feeding. Serum albumin remained in the 2.0 to 2.5 gm/dL range, hemoglobin in the 7 to 9 gm/dL range without any obvious source of anemia. Most of the pressure injuries worsened during his stay in spite of aggressive wound care, and he developed a second stage IV sacral wound. A single C-reactive protein (CRP) level 2 months prior to his death was markedly elevated at 19.5 mg/dL. In spite of maintaining NPO status, he developed 3 episodes of aspiration pneumonia, all of which responded well to treatment. Ultimately, he was found pulseless and apneic and resuscitation was unsuccessful. An autopsy revealed purulent material in the small airways.

Case 3 Presentation

A 65-year-old man with a long history of paranoid schizophrenia and severe gastroesophageal reflux disease had resided in the CLC for about 10 years. Medications included risperidone microspheres 37.5 mg every 2 weeks, valproic acid 500 mg 3 times daily and 1,000 mg every night at bedtime, lansoprazole 30 mg twice daily, ranitidine 150 mg every night at bedtime, sucralfate 1,000 mg 3 times daily, simvastatin 20 mg every night at bedtime, and tamsulosin 0.4 mg every night at bedtime. He had done well for many years but developed some drooling and a modest resting tremor (but no other signs of pseudoparkinsonism) about 8 years after admission.

There had been no changes to his risperidone dosage. He also lost about 20 lb over a period of 1 year and became increasingly weak and dependent in gait, serum albumin dropped as low as 1.6 gm/dL, hemoglobin dropped to the 7 to 8 gm/dL range (without any other obvious source of anemia), and he developed a gradually worsening right-sided pleural effusion. CRP was chronically elevated at this point, in the 6 to 15 mg/dL range and as high as 17.2 mg/dL. Ultimately, he developed 3 episodes of aspiration pneumonia over a period of 2 months. Swallowing evaluation at that time revealed severe oropharyngeal dysphagia and a PEG tube was placed. Due to concerns for possible antipsychotic-induced dysphagia, risperidone was discontinued, and quetiapine 400 mg a day was substituted. He did well over the subsequent year with no further pneumonia and advancement back to a regular diet. He regained all of the lost weight and began independent ambulation. Albumin improved to the 3 gm/dL range, hemoglobin to the 12 to 13 gm/dL range, and CRP had decreased to 0.7 mg/dL. The pleural effusion (believed to have been a parapneumonic effusion) had resolved.

Discussion

All 3 patients met the Fried criteria for frailty, although there were several confounding issues.² All 3 patients lost between 20 and 25 lb; all had clearly become weaker according to nursing and rehabilitation staff (although none were formally assessed for grip strength); and all had clear declines in their activity level. Patient 3 had a clear decrement in gait speed, but patient 1 had severe gait impairment due to Parkinson disease (although his gait in therapy had clearly worsened). Patient 2 was paraparetic and unable to ambulate. There also was evidence of limited biomarkers of systemic inflammation; all 3 patients’ albumin had decreased, and patients 2 and 3 had significant decrease in hemoglobin; but these commonplace clinical biomarkers are obviously multifactorially determined. We have limited data on our patients’ CRP levels; serial levels would have been more specific for systemic inflammation but were infrequently performed on the patients.

Multimorbidity and medical complexity are more the rule than the exception in frail geriatric patients,and it is difficult to separate the role of microaspiration from other confounding conditions that might have contributed to these patients’ evolving systemic inflammation and frailty.¹⁸ It might be argued that the decline for patient 1 was related to the underlying Parkinson disease (a progressive neurologic illness in which systemic inflammation has been reported), or that the decline of patient 2 was related to the worsening pressure injuries rather than to covert microaspiration.¹⁹ However, the TBIs for patient 2 and the schizophrenia for patient 3 would not be expected to be associated with frailty or with systemic inflammation. Furthermore, the frailty symptoms of patient 3 and inflammatory biomarkers improved after the risperidone, which was likely responsible for his microaspiration, was discontinued. All 3 patients were at risk for oropharyngeal dysphagia (antipsychotic medication is clearly associated with dysphagia²⁰); patient 2 demonstrated pathologic evidence of DAB at autopsy.

There is evolving evidence that chronic systemic inflammation and immune activation are key mechanisms in the pathogenesis of frailty.^4-6 It is known that elevated serum levels of proinflammatory cytokines, including tumor necrosis factor-α, interleukin-6, and CRP are directly associated with frailty and are inversely associated with levels of albumin, hemoglobin, insulin-like growth factor-1, and several micronutrients in frail individuals.^4-7,21,22 Chronic inflammation contributes to the pathophysiology of frailty through detrimental effects on a broad range of systems, including the musculoskeletal, endocrine, and hematopoietic systems and through nutritional dysregulation.^2,4,23 These changes may lead to further deleterious effects, creating a downward spiral of worsening frailty. For example, it seems likely that our patients’ progressive weakness further compromised airway protection, creating a vicious cycle of worsening microaspiration and chronic inflammation.

Conclusions

To date, the role of chronic microaspiration and DAB in chronic systemic inflammation or in frailty has not been explored. Given the prevalence of microaspiration in nursing home residents and the devastating consequences of frailty, though, this seems to be a crucial area of investigation. It is equally crucial for long-term care staff, both providers and nursing staff, to have a heightened awareness of covert microaspiration and a low threshold for referral to speech pathology for further investigation. Staff also should be aware of the utility of the Fried criteria to improve identification of frailty in general. It is probable that covert microaspiration will prove to be an important part of the differential diagnosis of frailty.

Hospitalizations related to ambulatory care sensitive conditions (ACSCs) are potentially avoidable if timely and effective care is provided to the patient. The Agency of Healthcare Research and Quality has identified type 2 diabetes mellitus (T2DM), chronic obstructive pulmonary disease (COPD), hypertension, congestive heart failure (CHF), urinary tract infections (UTIs), asthma, dehydration, bacterial pneumonia, angina without an inhospital procedure, and perforated appendix as ACSCs.^1,2 Identifying patients with ACSCs who are at risk for hospitalization is a potential measure to enhance primary care delivery and reduce preventable hospitalizations

The US Department of Veterans Affairs (VA) Clinical Pharmacy Practice Office implemented a guidance statement describing the role and impact of a clinical pharmacy specialist (CPS) in managing ACSCs.¹ Within the Veterans Health Administration, the CPS may function under a scope of practice within their area of expertise with the ability to prescribe medications, place consults, and order laboratory tests and additional referrals as appropriate. As hospitalizations related to ACSCs are potentially preventable with effective primary care, the CPS can play an essential primary care role to implement interventions targeted at reducing these hospitalizations.

At the William S. Middleton Memorial Veterans Hospital, in Madison, Wisconsin, multiple transitions of care and postdischarge services have been established to capture those patients who are at a high risk of rehospitalization. Studies have been completed regarding implementation of intensive case management programs for high-risk patients.³ Currently though, no standardized process or protocol exists that can identify and optimize primary care for patients with ACSCs who have been hospitalized but are predicted to be at low risk for rehospitalization. Although these patients may not require intensive case management like that of those at high risk, improvements can be made to optimize clinical resources, education, and patient self-monitoring to mitigate risk for hospitalization or rehospitalization. Therefore, this project aimed to evaluate the implementation of offering further referrals and care for patients who have been hospitalized but are considered low risk for hospitalization from ACSCs.

Methods

This quality improvement project to offer further referrals and care to patients considered low risk for hospitalization was implemented to enhance ambulatory-care provided services. All patients identified as being a low risk for hospitalization via a VA dashboard from July through September 2018 were included. Patients were identified based on age, chronic diseases, gender, and other patient-specific factors predetermined by the VA dashboard algorithm. Patients receiving hospice or palliative care and those no longer receiving primary care through the facility were excluded.

A pharmacy resident conducted a baseline chart review using a standardized template in the computerized patient record system (CPRS) to identify additional referrals or interventions a patient may benefit from based on any identified ACSC. Potential referral options included a CPS or nurse care manager disease management, whole health/wellness, educational classes, home monitoring equipment, specialty clinics, nutrition, cardiac or pulmonary rehabilitation, social work, and mental health. A pharmacy resident or the patient aligned care team (PACT) CPS reviewed the identified referrals with PACT members at interdisciplinary team meetings and determined which referrals to offer the patient. The pharmacy resident or designated PACT member reached out to the patient via telephone or during a clinic visit to offer and enter the referrals. If the patient agreed to any referrals, a chart review was conducted 3 months later to determine the percentage of initially agreed-upon referrals that the patient completed. Additionally, the number of emergency department (ED) visits and hospitalizations related to an ACSC at 3 months was collected.

Feasibility was assessed to evaluate potential service implementation and was measured by the time in minutes to complete the baseline chart review, time in minutes to offer referrals to the patient, and proportion of referrals that were completed at 3 months.⁴ As this quality improvement project was undertaken for programmatic evaluation, the University of Wisconsin-Madison Health Sciences Institutional Review Board determined that this project did not meet the federal definition of research and therefore review was not required. Data were analyzed using descriptive statistics.

Results

A total of 78 veterans who had ≥ 1 ACSC-related hospitalization in the past year and who were categorized as low risk were identified, and 69 veterans were reviewed. Nine patients were not included based on hospice care and no longer receiving primary care through the facility. Eight patients were found to have optimized care with no further action warranted after review. Based on their assigned PACT, there was a range of 0 to 5 patients identified per team. Fifty-one patients were contacted, and 37 accepted ≥ 1 referral. Most of the patients were white and male (Table). The most common ACSCs were hypertension (68%), COPD (46%), and T2DM (30%); additional ACSCs included angina (18%), pneumonia (15%), UTIs (10%), CHF (6%), and asthma, dehydration, and perforated appendix (1.5% for each). Any ACSC listed as a diagnosis for a patient was included, regardless of whether it was related to a hospitalization. Most referrals were offered by pharmacists (pharmacy resident, 41%; CPS, 29%), followed by the nurse care manager (18%) and the primary care provider (12%). One patient passed away related to heart failure complications prior to being contacted to offer additional referrals. Of the 9 patients that were unable to be contacted, 4 did not respond to 3 phone call attempts and 5 had no documentation of referrals being offered after the initial chart review and recommendation was completed.

Most of the initially accepted referrals (n = 68) were for CPS disease management, whole health/wellness, and educational classes (Figure). Of the 28 initially accepted referrals for CPS disease management, most were for COPD (10) and hypertension (8), followed by neuropathic pain (3), vitamin D deficiency (3), hyperlipidemia (2), and T2DM (2). At 3 months, all referrals were completed except for 1 hypertension, 1 vitamin D deficiency, and 2 hyperlipidemia referrals. There were 6 COPD, 4 T2DM self-management, and 1 chronic pain class referrals made with 3 COPD and 1 T2DM referrals completed at 3 months. Two tobacco treatment and 2 palliative care referrals were specialty referrals accepted by patients with 1 palliative care referral completed at 3 months.

In terms of feasibility, the chart review took an average (SD) of 13 (4) minutes, and contacting the patient to offer referrals took an average of 8 (5) minutes. Most of the accepted referrals were completed by 3 months (42/68, 62%).

Comparing the 3 months prior to and the 3 months after offering referrals, there was a cumulative quantitative decrease in the number of ED visits (5 to 1) and hospitalizations (11 to 5). The 1 ED visit was for a patient who was unable to be contacted to offer additional referrals as were 4 of the hospitalizations. One of the hospitalizations was for a patient who was deemed to have optimized care with no additional referrals necessary.

Discussion

Evaluation of the review and referral process for patients at low risk for hospitalization from an ACSC was a proactive approach toward optimizing primary care for veterans, and the process increased patient access to education and primary care. There was a high initial patient acceptance rate of referrals and a high completion rate when offered by PACT members. Based on the number of identified patients, the time spent completing chart reviews and contacting patients to offer referrals for each PACT CPS and team was feasible to conduct.

As there were 69 eligible patients identified over a 3-month period for a single VA facility, including all community-based outpatient clinics serving an estimated 130,000 veterans, the additional time and workload for an individual PACT to reach out to these patients is minimal. Completing the review and outreach process for an average of 21 minutes per patient for at most 5 patients per primary care provider team is feasible to complete during the recommended 4 hours of weekly CPS population health management responsibilities.

Limitations

Several limitations were identified with the implementation of the project. A variety of PACT members completed initial outreach to veterans regarding additional referrals, which may have resulted in a lack of consistency in the approach and discussion of offering referrals to patients. Although there may be a difference in how the team members made referral offers to patients and therefore varying acceptance rates by patients, the process was thought to be more generalizable to the PACT approach for providing care in the VA. In addition, the time to contact patients to offer referrals was not always documented in the electronic health record, making the documented time an estimate. Given that patients identified were managed by a variety of PACT members, there were differences noted among PACTs in terms of acceptability of offering referrals to patients.

While there was a decrease noted in ED visits and hospitalizations when comparing 3 months before and afterward, additional data are needed to provide further insight into this relationship. As the patients identified were at low risk for hospitalization from an ACSC and had 1 or 2 hospitalizations within the year prior, additional time is warranted to compare 12-month ED visits and hospitalization rates postintervention. Finally, these findings may be limited in generalizability to other health care systems as the project was conducted among a specific, veteran patient population with PACT CPSs practicing independently within an established broad scope of practice.

Future Directions

Future directions include incorporating the review and referral process into the PACT CPS population health management responsibilities as a way to use all PACT members to enhance primary care delivered to veterans. To further elucidate the relationship between the referral process and hospitalization rates, a longer data collection period is needed.

Conclusions

Identifying patients at risk for hospitalization from an ACSC via a review and referral process by using the VA PACT structure and team members was feasible and led to increased patient access to primary care and additional services. The PACT CPS would benefit from using a similar approach for population health management for low risk for hospitalization patients or other identified chronic conditions.

Acknowledgments

Presented at the Wisconsin Pharmacy Residency Conference at the Pharmacy Society of Wisconsin Educational Conference April 10, 2019, in Madison, Wisconsin.

Hospitalizations related to ambulatory care sensitive conditions (ACSCs) are potentially avoidable if timely and effective care is provided to the patient. The Agency of Healthcare Research and Quality has identified type 2 diabetes mellitus (T2DM), chronic obstructive pulmonary disease (COPD), hypertension, congestive heart failure (CHF), urinary tract infections (UTIs), asthma, dehydration, bacterial pneumonia, angina without an inhospital procedure, and perforated appendix as ACSCs.^1,2 Identifying patients with ACSCs who are at risk for hospitalization is a potential measure to enhance primary care delivery and reduce preventable hospitalizations

The US Department of Veterans Affairs (VA) Clinical Pharmacy Practice Office implemented a guidance statement describing the role and impact of a clinical pharmacy specialist (CPS) in managing ACSCs.¹ Within the Veterans Health Administration, the CPS may function under a scope of practice within their area of expertise with the ability to prescribe medications, place consults, and order laboratory tests and additional referrals as appropriate. As hospitalizations related to ACSCs are potentially preventable with effective primary care, the CPS can play an essential primary care role to implement interventions targeted at reducing these hospitalizations.

At the William S. Middleton Memorial Veterans Hospital, in Madison, Wisconsin, multiple transitions of care and postdischarge services have been established to capture those patients who are at a high risk of rehospitalization. Studies have been completed regarding implementation of intensive case management programs for high-risk patients.³ Currently though, no standardized process or protocol exists that can identify and optimize primary care for patients with ACSCs who have been hospitalized but are predicted to be at low risk for rehospitalization. Although these patients may not require intensive case management like that of those at high risk, improvements can be made to optimize clinical resources, education, and patient self-monitoring to mitigate risk for hospitalization or rehospitalization. Therefore, this project aimed to evaluate the implementation of offering further referrals and care for patients who have been hospitalized but are considered low risk for hospitalization from ACSCs.

Methods

This quality improvement project to offer further referrals and care to patients considered low risk for hospitalization was implemented to enhance ambulatory-care provided services. All patients identified as being a low risk for hospitalization via a VA dashboard from July through September 2018 were included. Patients were identified based on age, chronic diseases, gender, and other patient-specific factors predetermined by the VA dashboard algorithm. Patients receiving hospice or palliative care and those no longer receiving primary care through the facility were excluded.

A pharmacy resident conducted a baseline chart review using a standardized template in the computerized patient record system (CPRS) to identify additional referrals or interventions a patient may benefit from based on any identified ACSC. Potential referral options included a CPS or nurse care manager disease management, whole health/wellness, educational classes, home monitoring equipment, specialty clinics, nutrition, cardiac or pulmonary rehabilitation, social work, and mental health. A pharmacy resident or the patient aligned care team (PACT) CPS reviewed the identified referrals with PACT members at interdisciplinary team meetings and determined which referrals to offer the patient. The pharmacy resident or designated PACT member reached out to the patient via telephone or during a clinic visit to offer and enter the referrals. If the patient agreed to any referrals, a chart review was conducted 3 months later to determine the percentage of initially agreed-upon referrals that the patient completed. Additionally, the number of emergency department (ED) visits and hospitalizations related to an ACSC at 3 months was collected.

Feasibility was assessed to evaluate potential service implementation and was measured by the time in minutes to complete the baseline chart review, time in minutes to offer referrals to the patient, and proportion of referrals that were completed at 3 months.⁴ As this quality improvement project was undertaken for programmatic evaluation, the University of Wisconsin-Madison Health Sciences Institutional Review Board determined that this project did not meet the federal definition of research and therefore review was not required. Data were analyzed using descriptive statistics.

Results

A total of 78 veterans who had ≥ 1 ACSC-related hospitalization in the past year and who were categorized as low risk were identified, and 69 veterans were reviewed. Nine patients were not included based on hospice care and no longer receiving primary care through the facility. Eight patients were found to have optimized care with no further action warranted after review. Based on their assigned PACT, there was a range of 0 to 5 patients identified per team. Fifty-one patients were contacted, and 37 accepted ≥ 1 referral. Most of the patients were white and male (Table). The most common ACSCs were hypertension (68%), COPD (46%), and T2DM (30%); additional ACSCs included angina (18%), pneumonia (15%), UTIs (10%), CHF (6%), and asthma, dehydration, and perforated appendix (1.5% for each). Any ACSC listed as a diagnosis for a patient was included, regardless of whether it was related to a hospitalization. Most referrals were offered by pharmacists (pharmacy resident, 41%; CPS, 29%), followed by the nurse care manager (18%) and the primary care provider (12%). One patient passed away related to heart failure complications prior to being contacted to offer additional referrals. Of the 9 patients that were unable to be contacted, 4 did not respond to 3 phone call attempts and 5 had no documentation of referrals being offered after the initial chart review and recommendation was completed.

Most of the initially accepted referrals (n = 68) were for CPS disease management, whole health/wellness, and educational classes (Figure). Of the 28 initially accepted referrals for CPS disease management, most were for COPD (10) and hypertension (8), followed by neuropathic pain (3), vitamin D deficiency (3), hyperlipidemia (2), and T2DM (2). At 3 months, all referrals were completed except for 1 hypertension, 1 vitamin D deficiency, and 2 hyperlipidemia referrals. There were 6 COPD, 4 T2DM self-management, and 1 chronic pain class referrals made with 3 COPD and 1 T2DM referrals completed at 3 months. Two tobacco treatment and 2 palliative care referrals were specialty referrals accepted by patients with 1 palliative care referral completed at 3 months.

In terms of feasibility, the chart review took an average (SD) of 13 (4) minutes, and contacting the patient to offer referrals took an average of 8 (5) minutes. Most of the accepted referrals were completed by 3 months (42/68, 62%).

Comparing the 3 months prior to and the 3 months after offering referrals, there was a cumulative quantitative decrease in the number of ED visits (5 to 1) and hospitalizations (11 to 5). The 1 ED visit was for a patient who was unable to be contacted to offer additional referrals as were 4 of the hospitalizations. One of the hospitalizations was for a patient who was deemed to have optimized care with no additional referrals necessary.

Discussion

Evaluation of the review and referral process for patients at low risk for hospitalization from an ACSC was a proactive approach toward optimizing primary care for veterans, and the process increased patient access to education and primary care. There was a high initial patient acceptance rate of referrals and a high completion rate when offered by PACT members. Based on the number of identified patients, the time spent completing chart reviews and contacting patients to offer referrals for each PACT CPS and team was feasible to conduct.

As there were 69 eligible patients identified over a 3-month period for a single VA facility, including all community-based outpatient clinics serving an estimated 130,000 veterans, the additional time and workload for an individual PACT to reach out to these patients is minimal. Completing the review and outreach process for an average of 21 minutes per patient for at most 5 patients per primary care provider team is feasible to complete during the recommended 4 hours of weekly CPS population health management responsibilities.

Limitations

Several limitations were identified with the implementation of the project. A variety of PACT members completed initial outreach to veterans regarding additional referrals, which may have resulted in a lack of consistency in the approach and discussion of offering referrals to patients. Although there may be a difference in how the team members made referral offers to patients and therefore varying acceptance rates by patients, the process was thought to be more generalizable to the PACT approach for providing care in the VA. In addition, the time to contact patients to offer referrals was not always documented in the electronic health record, making the documented time an estimate. Given that patients identified were managed by a variety of PACT members, there were differences noted among PACTs in terms of acceptability of offering referrals to patients.

While there was a decrease noted in ED visits and hospitalizations when comparing 3 months before and afterward, additional data are needed to provide further insight into this relationship. As the patients identified were at low risk for hospitalization from an ACSC and had 1 or 2 hospitalizations within the year prior, additional time is warranted to compare 12-month ED visits and hospitalization rates postintervention. Finally, these findings may be limited in generalizability to other health care systems as the project was conducted among a specific, veteran patient population with PACT CPSs practicing independently within an established broad scope of practice.

Future Directions

Future directions include incorporating the review and referral process into the PACT CPS population health management responsibilities as a way to use all PACT members to enhance primary care delivered to veterans. To further elucidate the relationship between the referral process and hospitalization rates, a longer data collection period is needed.

Conclusions

Identifying patients at risk for hospitalization from an ACSC via a review and referral process by using the VA PACT structure and team members was feasible and led to increased patient access to primary care and additional services. The PACT CPS would benefit from using a similar approach for population health management for low risk for hospitalization patients or other identified chronic conditions.

Acknowledgments

Presented at the Wisconsin Pharmacy Residency Conference at the Pharmacy Society of Wisconsin Educational Conference April 10, 2019, in Madison, Wisconsin.

Hospitalizations related to ambulatory care sensitive conditions (ACSCs) are potentially avoidable if timely and effective care is provided to the patient. The Agency of Healthcare Research and Quality has identified type 2 diabetes mellitus (T2DM), chronic obstructive pulmonary disease (COPD), hypertension, congestive heart failure (CHF), urinary tract infections (UTIs), asthma, dehydration, bacterial pneumonia, angina without an inhospital procedure, and perforated appendix as ACSCs.^1,2 Identifying patients with ACSCs who are at risk for hospitalization is a potential measure to enhance primary care delivery and reduce preventable hospitalizations

The US Department of Veterans Affairs (VA) Clinical Pharmacy Practice Office implemented a guidance statement describing the role and impact of a clinical pharmacy specialist (CPS) in managing ACSCs.¹ Within the Veterans Health Administration, the CPS may function under a scope of practice within their area of expertise with the ability to prescribe medications, place consults, and order laboratory tests and additional referrals as appropriate. As hospitalizations related to ACSCs are potentially preventable with effective primary care, the CPS can play an essential primary care role to implement interventions targeted at reducing these hospitalizations.

At the William S. Middleton Memorial Veterans Hospital, in Madison, Wisconsin, multiple transitions of care and postdischarge services have been established to capture those patients who are at a high risk of rehospitalization. Studies have been completed regarding implementation of intensive case management programs for high-risk patients.³ Currently though, no standardized process or protocol exists that can identify and optimize primary care for patients with ACSCs who have been hospitalized but are predicted to be at low risk for rehospitalization. Although these patients may not require intensive case management like that of those at high risk, improvements can be made to optimize clinical resources, education, and patient self-monitoring to mitigate risk for hospitalization or rehospitalization. Therefore, this project aimed to evaluate the implementation of offering further referrals and care for patients who have been hospitalized but are considered low risk for hospitalization from ACSCs.

Methods

This quality improvement project to offer further referrals and care to patients considered low risk for hospitalization was implemented to enhance ambulatory-care provided services. All patients identified as being a low risk for hospitalization via a VA dashboard from July through September 2018 were included. Patients were identified based on age, chronic diseases, gender, and other patient-specific factors predetermined by the VA dashboard algorithm. Patients receiving hospice or palliative care and those no longer receiving primary care through the facility were excluded.

A pharmacy resident conducted a baseline chart review using a standardized template in the computerized patient record system (CPRS) to identify additional referrals or interventions a patient may benefit from based on any identified ACSC. Potential referral options included a CPS or nurse care manager disease management, whole health/wellness, educational classes, home monitoring equipment, specialty clinics, nutrition, cardiac or pulmonary rehabilitation, social work, and mental health. A pharmacy resident or the patient aligned care team (PACT) CPS reviewed the identified referrals with PACT members at interdisciplinary team meetings and determined which referrals to offer the patient. The pharmacy resident or designated PACT member reached out to the patient via telephone or during a clinic visit to offer and enter the referrals. If the patient agreed to any referrals, a chart review was conducted 3 months later to determine the percentage of initially agreed-upon referrals that the patient completed. Additionally, the number of emergency department (ED) visits and hospitalizations related to an ACSC at 3 months was collected.

Feasibility was assessed to evaluate potential service implementation and was measured by the time in minutes to complete the baseline chart review, time in minutes to offer referrals to the patient, and proportion of referrals that were completed at 3 months.⁴ As this quality improvement project was undertaken for programmatic evaluation, the University of Wisconsin-Madison Health Sciences Institutional Review Board determined that this project did not meet the federal definition of research and therefore review was not required. Data were analyzed using descriptive statistics.

Results

A total of 78 veterans who had ≥ 1 ACSC-related hospitalization in the past year and who were categorized as low risk were identified, and 69 veterans were reviewed. Nine patients were not included based on hospice care and no longer receiving primary care through the facility. Eight patients were found to have optimized care with no further action warranted after review. Based on their assigned PACT, there was a range of 0 to 5 patients identified per team. Fifty-one patients were contacted, and 37 accepted ≥ 1 referral. Most of the patients were white and male (Table). The most common ACSCs were hypertension (68%), COPD (46%), and T2DM (30%); additional ACSCs included angina (18%), pneumonia (15%), UTIs (10%), CHF (6%), and asthma, dehydration, and perforated appendix (1.5% for each). Any ACSC listed as a diagnosis for a patient was included, regardless of whether it was related to a hospitalization. Most referrals were offered by pharmacists (pharmacy resident, 41%; CPS, 29%), followed by the nurse care manager (18%) and the primary care provider (12%). One patient passed away related to heart failure complications prior to being contacted to offer additional referrals. Of the 9 patients that were unable to be contacted, 4 did not respond to 3 phone call attempts and 5 had no documentation of referrals being offered after the initial chart review and recommendation was completed.

Most of the initially accepted referrals (n = 68) were for CPS disease management, whole health/wellness, and educational classes (Figure). Of the 28 initially accepted referrals for CPS disease management, most were for COPD (10) and hypertension (8), followed by neuropathic pain (3), vitamin D deficiency (3), hyperlipidemia (2), and T2DM (2). At 3 months, all referrals were completed except for 1 hypertension, 1 vitamin D deficiency, and 2 hyperlipidemia referrals. There were 6 COPD, 4 T2DM self-management, and 1 chronic pain class referrals made with 3 COPD and 1 T2DM referrals completed at 3 months. Two tobacco treatment and 2 palliative care referrals were specialty referrals accepted by patients with 1 palliative care referral completed at 3 months.

In terms of feasibility, the chart review took an average (SD) of 13 (4) minutes, and contacting the patient to offer referrals took an average of 8 (5) minutes. Most of the accepted referrals were completed by 3 months (42/68, 62%).

Comparing the 3 months prior to and the 3 months after offering referrals, there was a cumulative quantitative decrease in the number of ED visits (5 to 1) and hospitalizations (11 to 5). The 1 ED visit was for a patient who was unable to be contacted to offer additional referrals as were 4 of the hospitalizations. One of the hospitalizations was for a patient who was deemed to have optimized care with no additional referrals necessary.

Discussion

Evaluation of the review and referral process for patients at low risk for hospitalization from an ACSC was a proactive approach toward optimizing primary care for veterans, and the process increased patient access to education and primary care. There was a high initial patient acceptance rate of referrals and a high completion rate when offered by PACT members. Based on the number of identified patients, the time spent completing chart reviews and contacting patients to offer referrals for each PACT CPS and team was feasible to conduct.

As there were 69 eligible patients identified over a 3-month period for a single VA facility, including all community-based outpatient clinics serving an estimated 130,000 veterans, the additional time and workload for an individual PACT to reach out to these patients is minimal. Completing the review and outreach process for an average of 21 minutes per patient for at most 5 patients per primary care provider team is feasible to complete during the recommended 4 hours of weekly CPS population health management responsibilities.

Limitations

Several limitations were identified with the implementation of the project. A variety of PACT members completed initial outreach to veterans regarding additional referrals, which may have resulted in a lack of consistency in the approach and discussion of offering referrals to patients. Although there may be a difference in how the team members made referral offers to patients and therefore varying acceptance rates by patients, the process was thought to be more generalizable to the PACT approach for providing care in the VA. In addition, the time to contact patients to offer referrals was not always documented in the electronic health record, making the documented time an estimate. Given that patients identified were managed by a variety of PACT members, there were differences noted among PACTs in terms of acceptability of offering referrals to patients.

While there was a decrease noted in ED visits and hospitalizations when comparing 3 months before and afterward, additional data are needed to provide further insight into this relationship. As the patients identified were at low risk for hospitalization from an ACSC and had 1 or 2 hospitalizations within the year prior, additional time is warranted to compare 12-month ED visits and hospitalization rates postintervention. Finally, these findings may be limited in generalizability to other health care systems as the project was conducted among a specific, veteran patient population with PACT CPSs practicing independently within an established broad scope of practice.

Future Directions

Future directions include incorporating the review and referral process into the PACT CPS population health management responsibilities as a way to use all PACT members to enhance primary care delivered to veterans. To further elucidate the relationship between the referral process and hospitalization rates, a longer data collection period is needed.

Conclusions

Identifying patients at risk for hospitalization from an ACSC via a review and referral process by using the VA PACT structure and team members was feasible and led to increased patient access to primary care and additional services. The PACT CPS would benefit from using a similar approach for population health management for low risk for hospitalization patients or other identified chronic conditions.

Acknowledgments

Presented at the Wisconsin Pharmacy Residency Conference at the Pharmacy Society of Wisconsin Educational Conference April 10, 2019, in Madison, Wisconsin.

The rise in prevalence of the community spread of coronavirus disease 2019 (COVID-19) in the US in early March 2020 led to hospital systems across the country preparing for an increase in critically ill patients.¹ The US Department of Veterans Affairs (VA) Ann Arbor Healthcare System (VAAAHS) anticipated an increased census of veterans who would need hospital admission for severe COVID-19 as well as the potential need to receive patients from community hospitals in Southeast Michigan, the location of one of the worst outbreaks in the US at that time.²

Through the facility’s incident command center, a hospital operations group identified the postanesthesia care unit (PACU) as a space to convert to an intensive care unit (ICU) for patients with COVID-19 needing mechanical ventilation. Other hospitals throughout the world have created similar makeshift ICUs to help care for the surge of patients with COVID-19, recognizing the high level of monitoring and resources available in the perioperative setting.^3-5 These ICUs have been successfully created in operating rooms,³ recovery rooms,⁵ and procedural settings.⁴

Between March 27, 2020 and April 25, 2020, a great multidisciplinary effort enabled the VAAAHS PACU-ICU to care for critically ill veterans with COVID-19 from Southeast Michigan as well as civilian transfers from overwhelmed neighboring community hospitals. This article will discuss planning considerations, including facility preparation, equipment, and staffing models. The unique challenges faced in managing an open-plan surge-capacity ICU also will be discussed as well as the solutions that were enacted.

Methods

Hospital Preparation

Maintaining a 2-zone model in which patients with COVID-19 and without COVID-19 could be cared for separately was of major importance. The VAAAHS traditional ICU was converted into a 16-bed COVID-19 ICU and staffed by the Pulmonary Critical Care Service. A separate wing of the hospital was converted into a 19-bed non-COVID-19 ICU, which also was staffed by the Pulmonary Critical Care Service that increased its staffing of residents, fellows, and attending physicians to meet the increasing clinical demands. Elective major surgery cases were postponed, and surgeons managed the care of postoperative surgical ICU patients. This arrangement allowed the existing 4 anesthesiologist intensivists to staff the PACU COVID-19 ICU.

Considerations, including space requirements, staffing, equipment, infection control requirements, and ability for facilities to engineer a negative pressure space were factored into the decision to convert the PACU to an additional 12-bed ICU. This effectively tripled the VAAAHS ICU capacity, enabling patient transfers from the John D. Dingell VA Medical Center in Detroit, Michigan, which was being impacted by a surge of cases in Detroit. In addition, this allowed for the opening of the hospital for both COVID-19 and non-COVID-19 ICU transfers from hospitals in Southeast Michigan in order to fulfill the fourth VA mission to provide care and support to state and local communities for emergency management, public health, and safety.

PACU Preparation

PACU was selected as an overflow ICU due to its open floor plan, allowing patients on ventilators to be seen from a central nursing station. This would allow for the safe use of ventilators without central alarm capabilities (especially anesthesia machines). Given the risk of a circuit disconnect, all ventilators without central alarm capabilities needed to be seen and heard within the space to ensure patient safety.

Facilities Management was able to construct temporary barriers with vinyl covered sheetrock and plexiglass to partition the central nursing workstation from the patient area in a U-shape (Figure 1). The patient area was turned into a negative pressure space where strict airborne precautions could be observed. Although the air handling unit serving this space is equipped with high efficiency particulate air (HEPA) filters, it was mechanically manipulated to ensure that all air coming from the space was discharged through exhaust and not recirculated into another occupied space within the hospital. Total air exchange rates were measured and calculated for both the positive and negative spaces to ensure they met or exceeded at least 6 air changes per hour, as recommended by Occupational Safety and Health Administration guidance.^6,7 A differential pressure indicator was installed to provide staff with the ability to monitor the pressure relationship between the 2 spaces in real time.

Twelve patient care beds were created. A traditionally engineered airborne infection isolation room in PACU served as a procedure room for aerosol-generating procedures, especially intubation, extubation, use of high-flow nasal cannula, and tracheostomy placement. Strict airborne precautions were taken within the patient area. The area inside the nursing station was positively pressurized to allow for surgical masks only to be required for the comfort of health care workers (Figure 2). A clear donning and doffing workflow was created for movement between the nursing area and the patient care area.

Personal Protective Equipment

Personal protective equipment (PPE) was of paramount importance in this open care unit. Airborne precautions were used in the entire patient care area. Powered air-purifying respirators (PAPRs) were used when possible to conserve the supply of N95 masks. Each health care worker was issued a reusable PAPR hood, which was cleaned by the user after each use by wiping the exterior of the entire hood with virucidal wipes. The brand and active ingredient of the virucidal wipes varied by availability of supplies, but the “virus kill time” was clearly labeled on each container. Each health care worker had a paper bag for storing his or her PAPR hood between usage to allow drying and ventilation. PAPR units were charged in between uses and shared by all clinical staff. Two layers of nonsterile gloves were worn.

Because of the open care area, attention had to be given to adhere to infection control policies if health care workers wanted to care for multiple patients while in the area. A new gown was placed over the existing gown, and the outer layer of gloves was removed. The under layer of gloves was then sanitized with hand sanitizer, and a new pair of outer gloves was then worn.

Equipment

Much of the ICU-level equipment needed was already present within the operating room (OR) area. Existing patient monitors were used and connected to a central monitoring station present in the nurses station. Relevant contents of the ICU storage room were duplicated and placed on shelves in the patient care area. Out-of-use anesthesia carts were used for a dedicated COVID-19 invasive line cart. A designated ultrasound with cardiac and vascular access probes was assigned to the PACU-ICU. Anesthesia machines were brought into the PACU-ICU and prepared with viral filters in line to prevent contamination of the machines, in keeping with national guidance from the American Society of Anesthesiologists and Anesthesia Patient Safety Foundation.⁸

Multidisciplinary Staffing Model

With the reduced surgical and procedural case load due to halting nonemergent operations, the Anesthesiology and Perioperative Care Service was able to staff the PACU-ICU with critical care anesthesiologists, nurse anesthetists, residents, and PACU and procedural nurses without hindering access to emergent surgeries. A separate preoperative area was maintained with an 8-bed capacity for both preoperative and postoperative management of non-COVID-19 surgical patients.

The staffing model was designed using guidance on the expansion of ICU staffing with non-ICU resources from the Society of Critical Care Medicine as well as local guidance on appropriate nursing ratios (Figure 3).⁹ Given the high acuity and dynamic nature of COVID-19 coupled with the unique considerations that exist using anesthesia machines as long-term ICU ventilators, 24-hour inhospital attending intensivist coverage was provided in the ICU by 4 critical care anesthesiologists who rotated between 12-hour day and night shifts. The critical care anesthesiologists led a team of anesthesiology and surgery residents and ICU advanced practice providers dedicated solely to the PACU-ICU. Non-ICU anesthesiologists helped with procedures such as intubation and invasive line placement and provided coverage of the ICU patients during sign-out and rounding. Certified registered nurse anesthetists (CRNAs) performed intubations and helped offload respiratory therapists (one of the resources most in shortage) by managing and weaning ventilators and were instrumental in prone positioning of patients. Dedicated ICU nurses were deployed every shift to oversee the unit and act as a resource to the PACU nurses. Fortunately, many PACU nurses had prior ICU training and experience, and nurses from outpatient areas also were recruited to help with patient care. Together, they provided direct patient care. OR nurses assisted with delivering supplies, medications and transporting specimens to the laboratory, as no formal hospital tube station was present in the PACU.

Because of the open-unit setting, nurses practiced bundled care and staggered their turns in the patient care area. For example, a nurse who entered to administer medication to patient A, could then receive communication to check the urine output for patient B and do so without completely doffing and redonning. This allowed preservation of PPE and reduced time in PPE for the health care providers (HCPs).

A scheduled daily meeting included staff from PACU-ICU; Medical ICU (MICU), which also treated patients with COVID-19; and the Palliative Care Service (Figure 4). Patients with single-organ failure were preferentially sent to PACU-ICU, as the ability to do renal replacement therapy (RRT) in an open unit proved difficult. The palliative care team and VAAAHS social workers assisted both MICU and PACU-ICU with communicating with patients’ families, which provided a great help during a clinically demanding time. Physical therapists increased their staffing of the ICU to specifically help with mobilization of patients with COVID-19 and acute respiratory distress syndrome, given the prolonged mechanical ventilation courses that were seen. Other consulting services frequently involved included infectious disease and nephrology.

Challenges and Solutions

Communication between staff located within the patient area and staff located in the nursing station was difficult given the loud noise generated by a PAPR and the plexiglass walls that separated the areas. Multiple techniques were attempted to overcome this. Dry erase boards were placed within the space to facilitate requests, but these were found to be time consuming. Two-way radios worked well if the users were wearing N95s but were harder to communicate when users were wearing PAPRs. Baby monitors were purchased to facilitate 2-way communication and were useful at times although quieter than desired. Vocera B3000N Communication Badges, which were already utilized in the perioperative period at the facility, could be utilized underneath PPE and were ultimately the best form of clear communication between staff within the patient care area and outside the negative pressure zone. In accordance with company guidance, these mobile devices were cleaned with virucidal wipes after use.¹⁰

Communication with patients’ families was critically important. The ICU team, palliative care team, or social workers made daily telephone calls to family members. The facility telehealth coordinator provided a designated tablet device to enable the intensivists to video conference with the patients’ families at bedside, utilizing virtual care manager appointments. This allowed families to see and interact with their loved ones despite the prohibition of family visitors. Every effort was made to utilize video calling daily; however, clinical demands as well as Internet and technological constraints from individual family members intermittently precluded video calls.

Clinical Challenges

Patients with severe COVID-19 infections requiring mechanical ventilation have proven to be exceptionally high-acuity patients with myriad organ-based complications reported.¹¹ Specific to our PACU-ICU, we determined that it was impractical to arrange for continuous RRT given the amount of training PACU nursing staff would have required and the limited ICU nursing staff in the PACU-ICU. Intermittent hemodialysis required replumbing for water supply and drainage but was ultimately not required as our facility expanded the number of continuous RRT machines available, allowing all patients in the COVID-19 ICU who required RRT to stay in the 16-bed ICU. Daily communication with the MICU allowed for safe transfer of patients with imminent needs for RRT to the MICU, providing a coordinated strategy for the deployment of scarce resources across our expanded ICU footprint.

Using anesthesia machines as ICU ventilators proved challenging, despite following best practice guidance.⁸ Notably, anesthesia machines are not actively humidified and require very high fresh gas flows, necessitating the addition of heat moisture exchangers (HME) to the circuit. Also, viral filters were placed in the circuit to prevent machine contamination. The addition of the HME and viral filters to each circuit increased the present dead space and led todifficulty in providing adequate ventilation to patients who already may have had a high proportion of physiologic dead space. The high fresh gas flows used still seemed inadequate in preventing moisture buildup in the machine parts, necessitating frequent exchanges of viral filters, HMEs, and circuits to prevent high peak airway pressures. In addition, anesthesia machines directly sample gas from the patient's breathing circuit, creating the risk for contamination of the space. This required a reconfiguration to allow for a suction scavenging system by VAAAHS biomedical engineers. Also, anesthesia machines are not designed for long-term ventilation and have different ventilation modes compared with modern ICU ventilators. Although they were used for several patients when the PACU-ICU opened, the hospital was able to acquire additional ICU ventilators, and extensive or prolonged use of anesthesia machine ventilators was avoided.

Infection Control

The open care setting provided unique infection control issues that had to be addressed.¹² The open setting allowed preservation of PPE and the ability for bundled care to be delivered easily. The VAAAHS infection control team worked closely with the ICU team to develop practices to ensure both patient and health care worker protection. Notable challenges included donning new gowns between patients when a PAPR was already being worn, leading to draping of new gowns over existing gowns when going between patients. True hand hygiene was also difficult, as health care workers did not want to completely remove gloves while in the patient care area. Layering of 2 pairs of gloves allowed the outer gloves to be removed after care of each patient, at which time alcohol gel was applied to the inner gloves, a new gown was placed over the existing gown, and a new pair of gloves was layered on top.

Although patients were intubated for long periods in the PACU-ICU, there was concern for increased risk of exposure of health care workers after extubation given the inability to contain the coughing patients within a private room. If a patient did well, they were transferred to a private room on the general medical floors within 24 hours of extubation to minimize this risk.

Privacy

The open care design meant less privacy for patients than would be provided in a private room. Curtains were drawn around patient beds as much as possible, especially for nursing care, but priority was given to visualization of the ventilator when a HCP was not present to ensure safety at all times. The majority of patients cared for in the PACU-ICU were intubated and sedated on arrival, but thankfully many were extubated. After extubation privacy in the open care area became more of an issue and may have led to more nighttime disturbances and substandard delirium prevention measures. Priority was given to expediting the transfer of these patients to private rooms on the general medical floor once their respiratory status was deemed stable.

Conclusions

The COVID-19 pandemic is truly an unprecedented event in our nation’s history, which has led to the first nationwide authorization of the fourth mission of VA to provide support for national, state, and local public health. The PACU-ICU was designed, engineered, built, and staffed by perioperative HCPs through an exceptional multidisciplinary effort in a matter of days. Through this dedication of health care workers and staff, the VAAAHS was able to care for critically ill veterans from Southeast Michigan and serve the community during a time of overwhelming demand on the national health care system.

Acknowledgments

The authors thank the outstanding team of administrators, engineers, physical therapists, pharmacists, nurses, advanced practice providers, CRNAs, respiratory therapists, and physicians who made it possible to respond to our veterans’ and our community’s needs in a time of unprecedented demand on our health care system. A special thank you to Eric Deters, Chief Strategy Officer; Brittany McClure, ICU Nurse Manager; and Mark Dotson, Chief Supply Chain Officer. It was a privilege to serve on this mission together.

The rise in prevalence of the community spread of coronavirus disease 2019 (COVID-19) in the US in early March 2020 led to hospital systems across the country preparing for an increase in critically ill patients.¹ The US Department of Veterans Affairs (VA) Ann Arbor Healthcare System (VAAAHS) anticipated an increased census of veterans who would need hospital admission for severe COVID-19 as well as the potential need to receive patients from community hospitals in Southeast Michigan, the location of one of the worst outbreaks in the US at that time.²

Through the facility’s incident command center, a hospital operations group identified the postanesthesia care unit (PACU) as a space to convert to an intensive care unit (ICU) for patients with COVID-19 needing mechanical ventilation. Other hospitals throughout the world have created similar makeshift ICUs to help care for the surge of patients with COVID-19, recognizing the high level of monitoring and resources available in the perioperative setting.^3-5 These ICUs have been successfully created in operating rooms,³ recovery rooms,⁵ and procedural settings.⁴

Between March 27, 2020 and April 25, 2020, a great multidisciplinary effort enabled the VAAAHS PACU-ICU to care for critically ill veterans with COVID-19 from Southeast Michigan as well as civilian transfers from overwhelmed neighboring community hospitals. This article will discuss planning considerations, including facility preparation, equipment, and staffing models. The unique challenges faced in managing an open-plan surge-capacity ICU also will be discussed as well as the solutions that were enacted.

Methods

Hospital Preparation

Maintaining a 2-zone model in which patients with COVID-19 and without COVID-19 could be cared for separately was of major importance. The VAAAHS traditional ICU was converted into a 16-bed COVID-19 ICU and staffed by the Pulmonary Critical Care Service. A separate wing of the hospital was converted into a 19-bed non-COVID-19 ICU, which also was staffed by the Pulmonary Critical Care Service that increased its staffing of residents, fellows, and attending physicians to meet the increasing clinical demands. Elective major surgery cases were postponed, and surgeons managed the care of postoperative surgical ICU patients. This arrangement allowed the existing 4 anesthesiologist intensivists to staff the PACU COVID-19 ICU.

Considerations, including space requirements, staffing, equipment, infection control requirements, and ability for facilities to engineer a negative pressure space were factored into the decision to convert the PACU to an additional 12-bed ICU. This effectively tripled the VAAAHS ICU capacity, enabling patient transfers from the John D. Dingell VA Medical Center in Detroit, Michigan, which was being impacted by a surge of cases in Detroit. In addition, this allowed for the opening of the hospital for both COVID-19 and non-COVID-19 ICU transfers from hospitals in Southeast Michigan in order to fulfill the fourth VA mission to provide care and support to state and local communities for emergency management, public health, and safety.

PACU Preparation

PACU was selected as an overflow ICU due to its open floor plan, allowing patients on ventilators to be seen from a central nursing station. This would allow for the safe use of ventilators without central alarm capabilities (especially anesthesia machines). Given the risk of a circuit disconnect, all ventilators without central alarm capabilities needed to be seen and heard within the space to ensure patient safety.

Facilities Management was able to construct temporary barriers with vinyl covered sheetrock and plexiglass to partition the central nursing workstation from the patient area in a U-shape (Figure 1). The patient area was turned into a negative pressure space where strict airborne precautions could be observed. Although the air handling unit serving this space is equipped with high efficiency particulate air (HEPA) filters, it was mechanically manipulated to ensure that all air coming from the space was discharged through exhaust and not recirculated into another occupied space within the hospital. Total air exchange rates were measured and calculated for both the positive and negative spaces to ensure they met or exceeded at least 6 air changes per hour, as recommended by Occupational Safety and Health Administration guidance.^6,7 A differential pressure indicator was installed to provide staff with the ability to monitor the pressure relationship between the 2 spaces in real time.

Twelve patient care beds were created. A traditionally engineered airborne infection isolation room in PACU served as a procedure room for aerosol-generating procedures, especially intubation, extubation, use of high-flow nasal cannula, and tracheostomy placement. Strict airborne precautions were taken within the patient area. The area inside the nursing station was positively pressurized to allow for surgical masks only to be required for the comfort of health care workers (Figure 2). A clear donning and doffing workflow was created for movement between the nursing area and the patient care area.

Personal Protective Equipment

Personal protective equipment (PPE) was of paramount importance in this open care unit. Airborne precautions were used in the entire patient care area. Powered air-purifying respirators (PAPRs) were used when possible to conserve the supply of N95 masks. Each health care worker was issued a reusable PAPR hood, which was cleaned by the user after each use by wiping the exterior of the entire hood with virucidal wipes. The brand and active ingredient of the virucidal wipes varied by availability of supplies, but the “virus kill time” was clearly labeled on each container. Each health care worker had a paper bag for storing his or her PAPR hood between usage to allow drying and ventilation. PAPR units were charged in between uses and shared by all clinical staff. Two layers of nonsterile gloves were worn.

Because of the open care area, attention had to be given to adhere to infection control policies if health care workers wanted to care for multiple patients while in the area. A new gown was placed over the existing gown, and the outer layer of gloves was removed. The under layer of gloves was then sanitized with hand sanitizer, and a new pair of outer gloves was then worn.

Equipment

Much of the ICU-level equipment needed was already present within the operating room (OR) area. Existing patient monitors were used and connected to a central monitoring station present in the nurses station. Relevant contents of the ICU storage room were duplicated and placed on shelves in the patient care area. Out-of-use anesthesia carts were used for a dedicated COVID-19 invasive line cart. A designated ultrasound with cardiac and vascular access probes was assigned to the PACU-ICU. Anesthesia machines were brought into the PACU-ICU and prepared with viral filters in line to prevent contamination of the machines, in keeping with national guidance from the American Society of Anesthesiologists and Anesthesia Patient Safety Foundation.⁸

Multidisciplinary Staffing Model

With the reduced surgical and procedural case load due to halting nonemergent operations, the Anesthesiology and Perioperative Care Service was able to staff the PACU-ICU with critical care anesthesiologists, nurse anesthetists, residents, and PACU and procedural nurses without hindering access to emergent surgeries. A separate preoperative area was maintained with an 8-bed capacity for both preoperative and postoperative management of non-COVID-19 surgical patients.

The staffing model was designed using guidance on the expansion of ICU staffing with non-ICU resources from the Society of Critical Care Medicine as well as local guidance on appropriate nursing ratios (Figure 3).⁹ Given the high acuity and dynamic nature of COVID-19 coupled with the unique considerations that exist using anesthesia machines as long-term ICU ventilators, 24-hour inhospital attending intensivist coverage was provided in the ICU by 4 critical care anesthesiologists who rotated between 12-hour day and night shifts. The critical care anesthesiologists led a team of anesthesiology and surgery residents and ICU advanced practice providers dedicated solely to the PACU-ICU. Non-ICU anesthesiologists helped with procedures such as intubation and invasive line placement and provided coverage of the ICU patients during sign-out and rounding. Certified registered nurse anesthetists (CRNAs) performed intubations and helped offload respiratory therapists (one of the resources most in shortage) by managing and weaning ventilators and were instrumental in prone positioning of patients. Dedicated ICU nurses were deployed every shift to oversee the unit and act as a resource to the PACU nurses. Fortunately, many PACU nurses had prior ICU training and experience, and nurses from outpatient areas also were recruited to help with patient care. Together, they provided direct patient care. OR nurses assisted with delivering supplies, medications and transporting specimens to the laboratory, as no formal hospital tube station was present in the PACU.

Because of the open-unit setting, nurses practiced bundled care and staggered their turns in the patient care area. For example, a nurse who entered to administer medication to patient A, could then receive communication to check the urine output for patient B and do so without completely doffing and redonning. This allowed preservation of PPE and reduced time in PPE for the health care providers (HCPs).

A scheduled daily meeting included staff from PACU-ICU; Medical ICU (MICU), which also treated patients with COVID-19; and the Palliative Care Service (Figure 4). Patients with single-organ failure were preferentially sent to PACU-ICU, as the ability to do renal replacement therapy (RRT) in an open unit proved difficult. The palliative care team and VAAAHS social workers assisted both MICU and PACU-ICU with communicating with patients’ families, which provided a great help during a clinically demanding time. Physical therapists increased their staffing of the ICU to specifically help with mobilization of patients with COVID-19 and acute respiratory distress syndrome, given the prolonged mechanical ventilation courses that were seen. Other consulting services frequently involved included infectious disease and nephrology.

Challenges and Solutions

Communication between staff located within the patient area and staff located in the nursing station was difficult given the loud noise generated by a PAPR and the plexiglass walls that separated the areas. Multiple techniques were attempted to overcome this. Dry erase boards were placed within the space to facilitate requests, but these were found to be time consuming. Two-way radios worked well if the users were wearing N95s but were harder to communicate when users were wearing PAPRs. Baby monitors were purchased to facilitate 2-way communication and were useful at times although quieter than desired. Vocera B3000N Communication Badges, which were already utilized in the perioperative period at the facility, could be utilized underneath PPE and were ultimately the best form of clear communication between staff within the patient care area and outside the negative pressure zone. In accordance with company guidance, these mobile devices were cleaned with virucidal wipes after use.¹⁰

Communication with patients’ families was critically important. The ICU team, palliative care team, or social workers made daily telephone calls to family members. The facility telehealth coordinator provided a designated tablet device to enable the intensivists to video conference with the patients’ families at bedside, utilizing virtual care manager appointments. This allowed families to see and interact with their loved ones despite the prohibition of family visitors. Every effort was made to utilize video calling daily; however, clinical demands as well as Internet and technological constraints from individual family members intermittently precluded video calls.

Clinical Challenges

Patients with severe COVID-19 infections requiring mechanical ventilation have proven to be exceptionally high-acuity patients with myriad organ-based complications reported.¹¹ Specific to our PACU-ICU, we determined that it was impractical to arrange for continuous RRT given the amount of training PACU nursing staff would have required and the limited ICU nursing staff in the PACU-ICU. Intermittent hemodialysis required replumbing for water supply and drainage but was ultimately not required as our facility expanded the number of continuous RRT machines available, allowing all patients in the COVID-19 ICU who required RRT to stay in the 16-bed ICU. Daily communication with the MICU allowed for safe transfer of patients with imminent needs for RRT to the MICU, providing a coordinated strategy for the deployment of scarce resources across our expanded ICU footprint.

Using anesthesia machines as ICU ventilators proved challenging, despite following best practice guidance.⁸ Notably, anesthesia machines are not actively humidified and require very high fresh gas flows, necessitating the addition of heat moisture exchangers (HME) to the circuit. Also, viral filters were placed in the circuit to prevent machine contamination. The addition of the HME and viral filters to each circuit increased the present dead space and led todifficulty in providing adequate ventilation to patients who already may have had a high proportion of physiologic dead space. The high fresh gas flows used still seemed inadequate in preventing moisture buildup in the machine parts, necessitating frequent exchanges of viral filters, HMEs, and circuits to prevent high peak airway pressures. In addition, anesthesia machines directly sample gas from the patient's breathing circuit, creating the risk for contamination of the space. This required a reconfiguration to allow for a suction scavenging system by VAAAHS biomedical engineers. Also, anesthesia machines are not designed for long-term ventilation and have different ventilation modes compared with modern ICU ventilators. Although they were used for several patients when the PACU-ICU opened, the hospital was able to acquire additional ICU ventilators, and extensive or prolonged use of anesthesia machine ventilators was avoided.

Infection Control

The open care setting provided unique infection control issues that had to be addressed.¹² The open setting allowed preservation of PPE and the ability for bundled care to be delivered easily. The VAAAHS infection control team worked closely with the ICU team to develop practices to ensure both patient and health care worker protection. Notable challenges included donning new gowns between patients when a PAPR was already being worn, leading to draping of new gowns over existing gowns when going between patients. True hand hygiene was also difficult, as health care workers did not want to completely remove gloves while in the patient care area. Layering of 2 pairs of gloves allowed the outer gloves to be removed after care of each patient, at which time alcohol gel was applied to the inner gloves, a new gown was placed over the existing gown, and a new pair of gloves was layered on top.

Although patients were intubated for long periods in the PACU-ICU, there was concern for increased risk of exposure of health care workers after extubation given the inability to contain the coughing patients within a private room. If a patient did well, they were transferred to a private room on the general medical floors within 24 hours of extubation to minimize this risk.

Privacy

The open care design meant less privacy for patients than would be provided in a private room. Curtains were drawn around patient beds as much as possible, especially for nursing care, but priority was given to visualization of the ventilator when a HCP was not present to ensure safety at all times. The majority of patients cared for in the PACU-ICU were intubated and sedated on arrival, but thankfully many were extubated. After extubation privacy in the open care area became more of an issue and may have led to more nighttime disturbances and substandard delirium prevention measures. Priority was given to expediting the transfer of these patients to private rooms on the general medical floor once their respiratory status was deemed stable.

Conclusions

The COVID-19 pandemic is truly an unprecedented event in our nation’s history, which has led to the first nationwide authorization of the fourth mission of VA to provide support for national, state, and local public health. The PACU-ICU was designed, engineered, built, and staffed by perioperative HCPs through an exceptional multidisciplinary effort in a matter of days. Through this dedication of health care workers and staff, the VAAAHS was able to care for critically ill veterans from Southeast Michigan and serve the community during a time of overwhelming demand on the national health care system.

Acknowledgments

The authors thank the outstanding team of administrators, engineers, physical therapists, pharmacists, nurses, advanced practice providers, CRNAs, respiratory therapists, and physicians who made it possible to respond to our veterans’ and our community’s needs in a time of unprecedented demand on our health care system. A special thank you to Eric Deters, Chief Strategy Officer; Brittany McClure, ICU Nurse Manager; and Mark Dotson, Chief Supply Chain Officer. It was a privilege to serve on this mission together.

The rise in prevalence of the community spread of coronavirus disease 2019 (COVID-19) in the US in early March 2020 led to hospital systems across the country preparing for an increase in critically ill patients.¹ The US Department of Veterans Affairs (VA) Ann Arbor Healthcare System (VAAAHS) anticipated an increased census of veterans who would need hospital admission for severe COVID-19 as well as the potential need to receive patients from community hospitals in Southeast Michigan, the location of one of the worst outbreaks in the US at that time.²

Through the facility’s incident command center, a hospital operations group identified the postanesthesia care unit (PACU) as a space to convert to an intensive care unit (ICU) for patients with COVID-19 needing mechanical ventilation. Other hospitals throughout the world have created similar makeshift ICUs to help care for the surge of patients with COVID-19, recognizing the high level of monitoring and resources available in the perioperative setting.^3-5 These ICUs have been successfully created in operating rooms,³ recovery rooms,⁵ and procedural settings.⁴

Between March 27, 2020 and April 25, 2020, a great multidisciplinary effort enabled the VAAAHS PACU-ICU to care for critically ill veterans with COVID-19 from Southeast Michigan as well as civilian transfers from overwhelmed neighboring community hospitals. This article will discuss planning considerations, including facility preparation, equipment, and staffing models. The unique challenges faced in managing an open-plan surge-capacity ICU also will be discussed as well as the solutions that were enacted.

Methods

Hospital Preparation

Maintaining a 2-zone model in which patients with COVID-19 and without COVID-19 could be cared for separately was of major importance. The VAAAHS traditional ICU was converted into a 16-bed COVID-19 ICU and staffed by the Pulmonary Critical Care Service. A separate wing of the hospital was converted into a 19-bed non-COVID-19 ICU, which also was staffed by the Pulmonary Critical Care Service that increased its staffing of residents, fellows, and attending physicians to meet the increasing clinical demands. Elective major surgery cases were postponed, and surgeons managed the care of postoperative surgical ICU patients. This arrangement allowed the existing 4 anesthesiologist intensivists to staff the PACU COVID-19 ICU.

Considerations, including space requirements, staffing, equipment, infection control requirements, and ability for facilities to engineer a negative pressure space were factored into the decision to convert the PACU to an additional 12-bed ICU. This effectively tripled the VAAAHS ICU capacity, enabling patient transfers from the John D. Dingell VA Medical Center in Detroit, Michigan, which was being impacted by a surge of cases in Detroit. In addition, this allowed for the opening of the hospital for both COVID-19 and non-COVID-19 ICU transfers from hospitals in Southeast Michigan in order to fulfill the fourth VA mission to provide care and support to state and local communities for emergency management, public health, and safety.

PACU Preparation

PACU was selected as an overflow ICU due to its open floor plan, allowing patients on ventilators to be seen from a central nursing station. This would allow for the safe use of ventilators without central alarm capabilities (especially anesthesia machines). Given the risk of a circuit disconnect, all ventilators without central alarm capabilities needed to be seen and heard within the space to ensure patient safety.

Facilities Management was able to construct temporary barriers with vinyl covered sheetrock and plexiglass to partition the central nursing workstation from the patient area in a U-shape (Figure 1). The patient area was turned into a negative pressure space where strict airborne precautions could be observed. Although the air handling unit serving this space is equipped with high efficiency particulate air (HEPA) filters, it was mechanically manipulated to ensure that all air coming from the space was discharged through exhaust and not recirculated into another occupied space within the hospital. Total air exchange rates were measured and calculated for both the positive and negative spaces to ensure they met or exceeded at least 6 air changes per hour, as recommended by Occupational Safety and Health Administration guidance.^6,7 A differential pressure indicator was installed to provide staff with the ability to monitor the pressure relationship between the 2 spaces in real time.

Twelve patient care beds were created. A traditionally engineered airborne infection isolation room in PACU served as a procedure room for aerosol-generating procedures, especially intubation, extubation, use of high-flow nasal cannula, and tracheostomy placement. Strict airborne precautions were taken within the patient area. The area inside the nursing station was positively pressurized to allow for surgical masks only to be required for the comfort of health care workers (Figure 2). A clear donning and doffing workflow was created for movement between the nursing area and the patient care area.

Personal Protective Equipment

Personal protective equipment (PPE) was of paramount importance in this open care unit. Airborne precautions were used in the entire patient care area. Powered air-purifying respirators (PAPRs) were used when possible to conserve the supply of N95 masks. Each health care worker was issued a reusable PAPR hood, which was cleaned by the user after each use by wiping the exterior of the entire hood with virucidal wipes. The brand and active ingredient of the virucidal wipes varied by availability of supplies, but the “virus kill time” was clearly labeled on each container. Each health care worker had a paper bag for storing his or her PAPR hood between usage to allow drying and ventilation. PAPR units were charged in between uses and shared by all clinical staff. Two layers of nonsterile gloves were worn.

Because of the open care area, attention had to be given to adhere to infection control policies if health care workers wanted to care for multiple patients while in the area. A new gown was placed over the existing gown, and the outer layer of gloves was removed. The under layer of gloves was then sanitized with hand sanitizer, and a new pair of outer gloves was then worn.

Equipment

Much of the ICU-level equipment needed was already present within the operating room (OR) area. Existing patient monitors were used and connected to a central monitoring station present in the nurses station. Relevant contents of the ICU storage room were duplicated and placed on shelves in the patient care area. Out-of-use anesthesia carts were used for a dedicated COVID-19 invasive line cart. A designated ultrasound with cardiac and vascular access probes was assigned to the PACU-ICU. Anesthesia machines were brought into the PACU-ICU and prepared with viral filters in line to prevent contamination of the machines, in keeping with national guidance from the American Society of Anesthesiologists and Anesthesia Patient Safety Foundation.⁸

Multidisciplinary Staffing Model

With the reduced surgical and procedural case load due to halting nonemergent operations, the Anesthesiology and Perioperative Care Service was able to staff the PACU-ICU with critical care anesthesiologists, nurse anesthetists, residents, and PACU and procedural nurses without hindering access to emergent surgeries. A separate preoperative area was maintained with an 8-bed capacity for both preoperative and postoperative management of non-COVID-19 surgical patients.

The staffing model was designed using guidance on the expansion of ICU staffing with non-ICU resources from the Society of Critical Care Medicine as well as local guidance on appropriate nursing ratios (Figure 3).⁹ Given the high acuity and dynamic nature of COVID-19 coupled with the unique considerations that exist using anesthesia machines as long-term ICU ventilators, 24-hour inhospital attending intensivist coverage was provided in the ICU by 4 critical care anesthesiologists who rotated between 12-hour day and night shifts. The critical care anesthesiologists led a team of anesthesiology and surgery residents and ICU advanced practice providers dedicated solely to the PACU-ICU. Non-ICU anesthesiologists helped with procedures such as intubation and invasive line placement and provided coverage of the ICU patients during sign-out and rounding. Certified registered nurse anesthetists (CRNAs) performed intubations and helped offload respiratory therapists (one of the resources most in shortage) by managing and weaning ventilators and were instrumental in prone positioning of patients. Dedicated ICU nurses were deployed every shift to oversee the unit and act as a resource to the PACU nurses. Fortunately, many PACU nurses had prior ICU training and experience, and nurses from outpatient areas also were recruited to help with patient care. Together, they provided direct patient care. OR nurses assisted with delivering supplies, medications and transporting specimens to the laboratory, as no formal hospital tube station was present in the PACU.

Because of the open-unit setting, nurses practiced bundled care and staggered their turns in the patient care area. For example, a nurse who entered to administer medication to patient A, could then receive communication to check the urine output for patient B and do so without completely doffing and redonning. This allowed preservation of PPE and reduced time in PPE for the health care providers (HCPs).

A scheduled daily meeting included staff from PACU-ICU; Medical ICU (MICU), which also treated patients with COVID-19; and the Palliative Care Service (Figure 4). Patients with single-organ failure were preferentially sent to PACU-ICU, as the ability to do renal replacement therapy (RRT) in an open unit proved difficult. The palliative care team and VAAAHS social workers assisted both MICU and PACU-ICU with communicating with patients’ families, which provided a great help during a clinically demanding time. Physical therapists increased their staffing of the ICU to specifically help with mobilization of patients with COVID-19 and acute respiratory distress syndrome, given the prolonged mechanical ventilation courses that were seen. Other consulting services frequently involved included infectious disease and nephrology.

Challenges and Solutions

Communication between staff located within the patient area and staff located in the nursing station was difficult given the loud noise generated by a PAPR and the plexiglass walls that separated the areas. Multiple techniques were attempted to overcome this. Dry erase boards were placed within the space to facilitate requests, but these were found to be time consuming. Two-way radios worked well if the users were wearing N95s but were harder to communicate when users were wearing PAPRs. Baby monitors were purchased to facilitate 2-way communication and were useful at times although quieter than desired. Vocera B3000N Communication Badges, which were already utilized in the perioperative period at the facility, could be utilized underneath PPE and were ultimately the best form of clear communication between staff within the patient care area and outside the negative pressure zone. In accordance with company guidance, these mobile devices were cleaned with virucidal wipes after use.¹⁰

Communication with patients’ families was critically important. The ICU team, palliative care team, or social workers made daily telephone calls to family members. The facility telehealth coordinator provided a designated tablet device to enable the intensivists to video conference with the patients’ families at bedside, utilizing virtual care manager appointments. This allowed families to see and interact with their loved ones despite the prohibition of family visitors. Every effort was made to utilize video calling daily; however, clinical demands as well as Internet and technological constraints from individual family members intermittently precluded video calls.

Clinical Challenges

Patients with severe COVID-19 infections requiring mechanical ventilation have proven to be exceptionally high-acuity patients with myriad organ-based complications reported.¹¹ Specific to our PACU-ICU, we determined that it was impractical to arrange for continuous RRT given the amount of training PACU nursing staff would have required and the limited ICU nursing staff in the PACU-ICU. Intermittent hemodialysis required replumbing for water supply and drainage but was ultimately not required as our facility expanded the number of continuous RRT machines available, allowing all patients in the COVID-19 ICU who required RRT to stay in the 16-bed ICU. Daily communication with the MICU allowed for safe transfer of patients with imminent needs for RRT to the MICU, providing a coordinated strategy for the deployment of scarce resources across our expanded ICU footprint.

Using anesthesia machines as ICU ventilators proved challenging, despite following best practice guidance.⁸ Notably, anesthesia machines are not actively humidified and require very high fresh gas flows, necessitating the addition of heat moisture exchangers (HME) to the circuit. Also, viral filters were placed in the circuit to prevent machine contamination. The addition of the HME and viral filters to each circuit increased the present dead space and led todifficulty in providing adequate ventilation to patients who already may have had a high proportion of physiologic dead space. The high fresh gas flows used still seemed inadequate in preventing moisture buildup in the machine parts, necessitating frequent exchanges of viral filters, HMEs, and circuits to prevent high peak airway pressures. In addition, anesthesia machines directly sample gas from the patient's breathing circuit, creating the risk for contamination of the space. This required a reconfiguration to allow for a suction scavenging system by VAAAHS biomedical engineers. Also, anesthesia machines are not designed for long-term ventilation and have different ventilation modes compared with modern ICU ventilators. Although they were used for several patients when the PACU-ICU opened, the hospital was able to acquire additional ICU ventilators, and extensive or prolonged use of anesthesia machine ventilators was avoided.

Infection Control

The open care setting provided unique infection control issues that had to be addressed.¹² The open setting allowed preservation of PPE and the ability for bundled care to be delivered easily. The VAAAHS infection control team worked closely with the ICU team to develop practices to ensure both patient and health care worker protection. Notable challenges included donning new gowns between patients when a PAPR was already being worn, leading to draping of new gowns over existing gowns when going between patients. True hand hygiene was also difficult, as health care workers did not want to completely remove gloves while in the patient care area. Layering of 2 pairs of gloves allowed the outer gloves to be removed after care of each patient, at which time alcohol gel was applied to the inner gloves, a new gown was placed over the existing gown, and a new pair of gloves was layered on top.

Although patients were intubated for long periods in the PACU-ICU, there was concern for increased risk of exposure of health care workers after extubation given the inability to contain the coughing patients within a private room. If a patient did well, they were transferred to a private room on the general medical floors within 24 hours of extubation to minimize this risk.

Privacy

The open care design meant less privacy for patients than would be provided in a private room. Curtains were drawn around patient beds as much as possible, especially for nursing care, but priority was given to visualization of the ventilator when a HCP was not present to ensure safety at all times. The majority of patients cared for in the PACU-ICU were intubated and sedated on arrival, but thankfully many were extubated. After extubation privacy in the open care area became more of an issue and may have led to more nighttime disturbances and substandard delirium prevention measures. Priority was given to expediting the transfer of these patients to private rooms on the general medical floor once their respiratory status was deemed stable.

Conclusions

The COVID-19 pandemic is truly an unprecedented event in our nation’s history, which has led to the first nationwide authorization of the fourth mission of VA to provide support for national, state, and local public health. The PACU-ICU was designed, engineered, built, and staffed by perioperative HCPs through an exceptional multidisciplinary effort in a matter of days. Through this dedication of health care workers and staff, the VAAAHS was able to care for critically ill veterans from Southeast Michigan and serve the community during a time of overwhelming demand on the national health care system.

Acknowledgments

The authors thank the outstanding team of administrators, engineers, physical therapists, pharmacists, nurses, advanced practice providers, CRNAs, respiratory therapists, and physicians who made it possible to respond to our veterans’ and our community’s needs in a time of unprecedented demand on our health care system. A special thank you to Eric Deters, Chief Strategy Officer; Brittany McClure, ICU Nurse Manager; and Mark Dotson, Chief Supply Chain Officer. It was a privilege to serve on this mission together.

Over the last several decades science has fallen off this country’s radar screen. Yes, STEM (science, technology, engineering, and mathematics) has recently had a brief moment in the spotlight as a buzzword de jour. But the critical importance of careful and systematic investigation into the world around us using observation and trial and error is a tough sell to a large segment of our population.

SDI Productions/iStock/Getty Images

The COVID-19 pandemic is providing an excellent opportunity for science and medicine to showcase their star qualities. Of course some people in leadership positions persist in disregarding the value of scientific investigation. But I get the feeling that the fear generated by the pandemic is creating some converts among many previous science skeptics. This gathering enthusiasm among the general population is a predictably slow process because that’s the way science works. It often doesn’t provide quick answers. And it is difficult for the nonscientist to see the beauty in the reality that the things we thought were true 2 months ago are likely to be proven wrong today as more observations accumulate.

Unfortunately, even in this time of renewal, science and medicine continue to generate a bumper crop of bad apples. A recent New York Times article examines the career of one such unscrupulous physician/scientist whose recent exploits threaten to undo much of the positive image the pandemic has cast on science (“The Doctor Behind the Disputed Covid Data,” by Ellen Gabler and Roni Caryn Rabin, The New York Times, July 27, 2020). The subject of the article is the physician who was responsible for providing some of the large data sets on which several papers were published about the apparent ineffectiveness and danger of using hydroxychloroquine in COVID-19 patients. The authenticity of the data sets recently has been seriously questioned, and the articles have been retracted by the journals in which they had appeared.

Based on numerous interviews with coworkers, the Times reporters present a strong case that this individual’s long history of unreliability make his association with allegedly fraudulent data set not surprising but maybe even predictable. At one point in his training, there appears to have been serious questions about advancing the physician to the next level. Despite these concerns, he was allowed to continue and complete his specialty training. It is of note that in his last year of clinical practice, the physician became the subject of three serious malpractice claims that question his competence.

I suspect that some of you have crossed paths with physicians whose competence and/or moral character you found concerning. Were they peers? Were you the individual’s supervisor or was he or she your mentor? How did you respond? Did anyone respond at all?

There has been a lot written and said in recent months about how and when to respond to respond to sexual harassment in the workplace. But I don’t recall reading any articles that discuss how one should respond to incompetence. Of course competency can be a relative term, but in most cases significant incompetence is hard to miss because it tends to be repeated.

It is easy for the airports and subway systems to post signs that say “If you see something say something.” It’s a different story for hospitals and medical schools that may have systems in place for reporting and following up on poor practice. But my sense is that there are too many cases that slip through the cracks.

This is another example of a problem for which I don’t have a solution. However, if this column prompts just one of you who sees something to say something then I have had a good day.
 

Dr. Wilkoff practiced primary care pediatrics in Brunswick, Maine, for nearly 40 years. He has authored several books on behavioral pediatrics, including “How to Say No to Your Toddler.” Email him at pdnews@mdedge.com.

Over the last several decades science has fallen off this country’s radar screen. Yes, STEM (science, technology, engineering, and mathematics) has recently had a brief moment in the spotlight as a buzzword de jour. But the critical importance of careful and systematic investigation into the world around us using observation and trial and error is a tough sell to a large segment of our population.

SDI Productions/iStock/Getty Images

The COVID-19 pandemic is providing an excellent opportunity for science and medicine to showcase their star qualities. Of course some people in leadership positions persist in disregarding the value of scientific investigation. But I get the feeling that the fear generated by the pandemic is creating some converts among many previous science skeptics. This gathering enthusiasm among the general population is a predictably slow process because that’s the way science works. It often doesn’t provide quick answers. And it is difficult for the nonscientist to see the beauty in the reality that the things we thought were true 2 months ago are likely to be proven wrong today as more observations accumulate.

Unfortunately, even in this time of renewal, science and medicine continue to generate a bumper crop of bad apples. A recent New York Times article examines the career of one such unscrupulous physician/scientist whose recent exploits threaten to undo much of the positive image the pandemic has cast on science (“The Doctor Behind the Disputed Covid Data,” by Ellen Gabler and Roni Caryn Rabin, The New York Times, July 27, 2020). The subject of the article is the physician who was responsible for providing some of the large data sets on which several papers were published about the apparent ineffectiveness and danger of using hydroxychloroquine in COVID-19 patients. The authenticity of the data sets recently has been seriously questioned, and the articles have been retracted by the journals in which they had appeared.

Based on numerous interviews with coworkers, the Times reporters present a strong case that this individual’s long history of unreliability make his association with allegedly fraudulent data set not surprising but maybe even predictable. At one point in his training, there appears to have been serious questions about advancing the physician to the next level. Despite these concerns, he was allowed to continue and complete his specialty training. It is of note that in his last year of clinical practice, the physician became the subject of three serious malpractice claims that question his competence.

I suspect that some of you have crossed paths with physicians whose competence and/or moral character you found concerning. Were they peers? Were you the individual’s supervisor or was he or she your mentor? How did you respond? Did anyone respond at all?

There has been a lot written and said in recent months about how and when to respond to respond to sexual harassment in the workplace. But I don’t recall reading any articles that discuss how one should respond to incompetence. Of course competency can be a relative term, but in most cases significant incompetence is hard to miss because it tends to be repeated.

It is easy for the airports and subway systems to post signs that say “If you see something say something.” It’s a different story for hospitals and medical schools that may have systems in place for reporting and following up on poor practice. But my sense is that there are too many cases that slip through the cracks.

This is another example of a problem for which I don’t have a solution. However, if this column prompts just one of you who sees something to say something then I have had a good day.
 

Dr. Wilkoff practiced primary care pediatrics in Brunswick, Maine, for nearly 40 years. He has authored several books on behavioral pediatrics, including “How to Say No to Your Toddler.” Email him at pdnews@mdedge.com.

Over the last several decades science has fallen off this country’s radar screen. Yes, STEM (science, technology, engineering, and mathematics) has recently had a brief moment in the spotlight as a buzzword de jour. But the critical importance of careful and systematic investigation into the world around us using observation and trial and error is a tough sell to a large segment of our population.

SDI Productions/iStock/Getty Images

The COVID-19 pandemic is providing an excellent opportunity for science and medicine to showcase their star qualities. Of course some people in leadership positions persist in disregarding the value of scientific investigation. But I get the feeling that the fear generated by the pandemic is creating some converts among many previous science skeptics. This gathering enthusiasm among the general population is a predictably slow process because that’s the way science works. It often doesn’t provide quick answers. And it is difficult for the nonscientist to see the beauty in the reality that the things we thought were true 2 months ago are likely to be proven wrong today as more observations accumulate.

Unfortunately, even in this time of renewal, science and medicine continue to generate a bumper crop of bad apples. A recent New York Times article examines the career of one such unscrupulous physician/scientist whose recent exploits threaten to undo much of the positive image the pandemic has cast on science (“The Doctor Behind the Disputed Covid Data,” by Ellen Gabler and Roni Caryn Rabin, The New York Times, July 27, 2020). The subject of the article is the physician who was responsible for providing some of the large data sets on which several papers were published about the apparent ineffectiveness and danger of using hydroxychloroquine in COVID-19 patients. The authenticity of the data sets recently has been seriously questioned, and the articles have been retracted by the journals in which they had appeared.

Based on numerous interviews with coworkers, the Times reporters present a strong case that this individual’s long history of unreliability make his association with allegedly fraudulent data set not surprising but maybe even predictable. At one point in his training, there appears to have been serious questions about advancing the physician to the next level. Despite these concerns, he was allowed to continue and complete his specialty training. It is of note that in his last year of clinical practice, the physician became the subject of three serious malpractice claims that question his competence.

I suspect that some of you have crossed paths with physicians whose competence and/or moral character you found concerning. Were they peers? Were you the individual’s supervisor or was he or she your mentor? How did you respond? Did anyone respond at all?

There has been a lot written and said in recent months about how and when to respond to respond to sexual harassment in the workplace. But I don’t recall reading any articles that discuss how one should respond to incompetence. Of course competency can be a relative term, but in most cases significant incompetence is hard to miss because it tends to be repeated.

It is easy for the airports and subway systems to post signs that say “If you see something say something.” It’s a different story for hospitals and medical schools that may have systems in place for reporting and following up on poor practice. But my sense is that there are too many cases that slip through the cracks.

This is another example of a problem for which I don’t have a solution. However, if this column prompts just one of you who sees something to say something then I have had a good day.
 

Dr. Wilkoff practiced primary care pediatrics in Brunswick, Maine, for nearly 40 years. He has authored several books on behavioral pediatrics, including “How to Say No to Your Toddler.” Email him at pdnews@mdedge.com.