User login
BOSTON – Resident participation in emergency general surgery cases was independently associated with a host of complications – pulmonary embolism, surgical site infections, and unplanned reoperation in a secondary analysis of the American College of Surgeons prospective National Surgical Quality Improvement database.
Adequate exposure of residents to emergency general surgery is crucial for surgical training, but academic operating teams should be mindful of this association, Dr. George Kasotakis said at the annual meeting of the American Surgical Association.
Three papers have shown resident participation modestly increases complications in elective surgery, but their impact is not well understood for emergency general surgery, where patient physiology is typically deranged and timely surgery is imperative.
Dr. Kasotakis and his colleagues at Boston University identified 141,010 patients who underwent emergency general surgery procedures in the 2005-2010 American College of Surgeons prospective National Surgical Quality Improvement database. Because of the nonrandom assignment of more complex cases to resident participation, patients were matched 1:1 based on age; gender; use of alcohol, tobacco, and steroids; inpatient status; obesity; diabetes; renal failure; cardiopulmonary disease; and expected probability for morbidity and mortality. Regression models were fitted for each outcome and adjusted for the same risk factors and operative time.
The most common procedures were appendectomy (40%), exploratory laparotomy (8.75%), bowel resection (9.2%), cholecystectomy (6%), and lysis of adhesions (6%).
Thirty-day mortality was similar with and without residents (3.25% vs. 2.96%; P = .082), but hospital length of stay was longer by about a half a day with residents (4.97 days vs. 4.59 days; P = .019), said Dr. Kasotakis, an acute care surgeon and intensivist.
Resident participation added about 20 minutes to operative (75 minutes vs. 59 minutes; P less than .001) and anesthesia (122 minutes vs. 100 minutes; P less than .001) times.
Intraoperative transfusions were more common with residents (3.43% vs. 2.55%; P less than .001), perhaps because of longer operating room times, and, as a result, fewer postoperative transfusions were needed (1.12% vs. 1.28% P = .031), he said. Unplanned reoperations, however, were more common with residents, as well (4.22% vs. 3.80%; P = .002).
Postsurgical superficial wound infections (3.5% vs. 2.78%; P less than .001) and organ space surgical site infections (2.27% vs. 1.77%; P less than .001) were more common in the resident group, while wound dehiscence was not (0.63% vs. 0.69%; P = .266), Dr. Kasotakis noted.
Pulmonary complications were significantly more common in the resident group including postoperative pneumonia (1.85% vs. 1.67%; P = .04), reintubation (1.64% vs. 1.15%; P less than .001), and mechanical ventilation for more than 48 hours (2.87% vs. 2.06%; P less than .001).
The same was true for deep vein thrombosis (DVT) (0.80% vs. 0.62%; P = .002) and pulmonary embolism (PE) (0.43% vs. 0.28%; P less than .001).
Urinary tract infections (UTI) were higher with resident participation (1.45% vs. 1.14%; P less than .001), as was sepsis (2.42% vs. 2.13%; P = .005), likely because of the increase in surgical infections, Dr. Kasotakis said.
Thankfully, significant cardiac complications and septic shock were not more common with residents, he said.
Adjusted analyses
After adjustment for operative duration, case complexity and pre-existing comorbidities, residents did not increase length of stay (odds ratio, 0.07; P = .242) or septic events (OR, 1.07; P = .155), but their participation was still independently associated with about 20% more superficial surgical site infections (odds ratio, 1.23; P less than .001), organ space infections (OR, 1.21; P less than .001), UTIs (OR, 1.23; P = .001), and intraoperative transfusions (OR, 1.20; P = .001), he said.
Also, about 8% more patients required a return trip to the operating room when residents participated (OR, 1.08; P = .041).
"These outcomes can perhaps be attributed to their underdeveloped surgical skills," Dr. Kasotakis said.
The incidence of DVT and PE were also higher by about 25% (OR, 1.25; P = .011) and 40% (OR, 1.42; P = .005), respectively, perhaps because of delayed DVT prophylaxis initiation because of concerns of hemostasis or missed doses due to additional return trips to the emergency department, he suggested.
Interestingly, reintubation and prolonged mechanical ventilation rates were increased by about 40% (OR, 1.38; OR, 1.43; both P less than .001), perhaps because of prolonged operative times or greater resuscitation requirements, he added.
Dr. Kasotakis was quick to point out that this was a secondary analysis of a data set not originally intended to assess the effect of trainee participation, that no information was available on the degree of resident involvement during surgery or in perioperative care, and that participating institutions were skewed toward tertiary centers, which typically receive more complex cases.
"Staff surgeons should supervise as needed and minimize unnecessary [emergency department] time. And residents, for their part, should be well prepared for emergency procedures through simulation training and aim to maximize their operating room efficiency," he suggested.
The results sparked a flurry of rebuttals led off by discussant Dr. Julie Ann Sosa, Duke University, Durham, N.C.,who said they conflict with other analyses showing little to no impact from residents in elective cases.
"If not interpreted with care, policy makers, payers, and the public could construe that surgical care at academic health centers is compromised by trainees, which could have unfortunate ramifications for everyone in the room as well as the trainees and the patients," she said.
Dr. Sosa expressed concern about drawing causal inferences from an observational study in the setting of possible selection bias and said attempts to match for case complexity using CPT codes do not necessarily account for say, "the difference between a routine appendectomy that takes 15-30 minutes and a complex one that takes 3 hours for a perforation."
Some attendees questioned why the authors didn’t match the institutions in the analysis and chose to ascribe all of the outcomes to residents, with a round of applause following the suggestion that the paper should be titled "Academic centers increase emergency surgery complications." Other attendees questioned whether the poor outcomes reflect resident training and supervision.
Dr. George Velmahos, Massachusetts General Hospital, Boston, questioned whether hospitals have a medical/legal responsibility to inform patients that a resident is in the operating room and may impact outcomes.
Dr. Kasotakis said that institutions may want to add a clause to consent paperwork stating that residents and trainees will be participating.
The complete manuscript of this study and its presentation at the American Surgical Association’s 134th Annual Meeting is anticipated to be published in the Annals of Surgery, pending editorial review.
Dr. Kasotakis reported no conflicts.
 |
| Dr. Sapan S. Desai |
While it is tempting to value the very high sample sizes reported in studies that utilize administrative database sources such as NSQIP, it is with great caution that we should interpret the results of these studies. Databases are a valuable source of determining incidence, costs, and even some correlations among various factors. Correlations, however, do not imply causation. For instance, one could say that there is a high correlation between the number of taxis on the street after 3 a.m. and the number of crimes, but that does not mean that taxi drivers transform into criminals during the hour of the wolf.
Databases are a blunt tool for identifying trends in data over large samples and over many years. However, as noted in the study presented by Dr. Kasotakis, the NSQIP does not track the extent of resident participation in operative cases. In fact, the exact variable that is tracked is entitled "ATTEND" in position 16 of the NSQIP and is as follows: 1) Attending & Resident in OR, 2) Attending Alone, and 3) Attending Not Present, but Available. It does not characterize the amount of the case completed by the resident, nor does it state the level of training of the resident. It does not mention whether medical students, visitors, nurse trainees, PAs, new scrub techs, or other personnel are present. In fact, there is no way to know who did what during the case, which is a requirement in order to be able to reliably and accurately determine that the resident is the culprit for longer case times and more intraoperative transfusions. This specific variable has insufficient granularity: the conclusion that "[adding] a clause to consent paperwork stating that residents and trainees will be participating in the case" due to the "underdeveloped surgical skills" of trainees, among other factors, is invalid.
As Dr. Sosa indicated, it is dangerous to imply that these correlations somehow imply causation, as this may lead to adverse impacts on training and negatively impact patient care far more than a possible 20 minute increase in the duration of an operation. The sole utility of this paper is that it is hypothesis generating; only a properly designed prospective study can truly study the positive and negative impacts on patient care due to resident education in the operating room. The question is, do we really want to know the answer to this question and will it truly affect what we do now?
Sapan S. Desai, M.D., is the resident medical editor for Vascular Specialist.
| |
| Dr. Sapan S. Desai |
While it is tempting to value the very high sample sizes reported in studies that utilize administrative database sources such as NSQIP, it is with great caution that we should interpret the results of these studies. Databases are a valuable source of determining incidence, costs, and even some correlations among various factors. Correlations, however, do not imply causation. For instance, one could say that there is a high correlation between the number of taxis on the street after 3 a.m. and the number of crimes, but that does not mean that taxi drivers transform into criminals during the hour of the wolf.
Databases are a blunt tool for identifying trends in data over large samples and over many years. However, as noted in the study presented by Dr. Kasotakis, the NSQIP does not track the extent of resident participation in operative cases. In fact, the exact variable that is tracked is entitled "ATTEND" in position 16 of the NSQIP and is as follows: 1) Attending & Resident in OR, 2) Attending Alone, and 3) Attending Not Present, but Available. It does not characterize the amount of the case completed by the resident, nor does it state the level of training of the resident. It does not mention whether medical students, visitors, nurse trainees, PAs, new scrub techs, or other personnel are present. In fact, there is no way to know who did what during the case, which is a requirement in order to be able to reliably and accurately determine that the resident is the culprit for longer case times and more intraoperative transfusions. This specific variable has insufficient granularity: the conclusion that "[adding] a clause to consent paperwork stating that residents and trainees will be participating in the case" due to the "underdeveloped surgical skills" of trainees, among other factors, is invalid.
As Dr. Sosa indicated, it is dangerous to imply that these correlations somehow imply causation, as this may lead to adverse impacts on training and negatively impact patient care far more than a possible 20 minute increase in the duration of an operation. The sole utility of this paper is that it is hypothesis generating; only a properly designed prospective study can truly study the positive and negative impacts on patient care due to resident education in the operating room. The question is, do we really want to know the answer to this question and will it truly affect what we do now?
Sapan S. Desai, M.D., is the resident medical editor for Vascular Specialist.
| |
| Dr. Sapan S. Desai |
While it is tempting to value the very high sample sizes reported in studies that utilize administrative database sources such as NSQIP, it is with great caution that we should interpret the results of these studies. Databases are a valuable source of determining incidence, costs, and even some correlations among various factors. Correlations, however, do not imply causation. For instance, one could say that there is a high correlation between the number of taxis on the street after 3 a.m. and the number of crimes, but that does not mean that taxi drivers transform into criminals during the hour of the wolf.
Databases are a blunt tool for identifying trends in data over large samples and over many years. However, as noted in the study presented by Dr. Kasotakis, the NSQIP does not track the extent of resident participation in operative cases. In fact, the exact variable that is tracked is entitled "ATTEND" in position 16 of the NSQIP and is as follows: 1) Attending & Resident in OR, 2) Attending Alone, and 3) Attending Not Present, but Available. It does not characterize the amount of the case completed by the resident, nor does it state the level of training of the resident. It does not mention whether medical students, visitors, nurse trainees, PAs, new scrub techs, or other personnel are present. In fact, there is no way to know who did what during the case, which is a requirement in order to be able to reliably and accurately determine that the resident is the culprit for longer case times and more intraoperative transfusions. This specific variable has insufficient granularity: the conclusion that "[adding] a clause to consent paperwork stating that residents and trainees will be participating in the case" due to the "underdeveloped surgical skills" of trainees, among other factors, is invalid.
As Dr. Sosa indicated, it is dangerous to imply that these correlations somehow imply causation, as this may lead to adverse impacts on training and negatively impact patient care far more than a possible 20 minute increase in the duration of an operation. The sole utility of this paper is that it is hypothesis generating; only a properly designed prospective study can truly study the positive and negative impacts on patient care due to resident education in the operating room. The question is, do we really want to know the answer to this question and will it truly affect what we do now?
Sapan S. Desai, M.D., is the resident medical editor for Vascular Specialist.
BOSTON – Resident participation in emergency general surgery cases was independently associated with a host of complications – pulmonary embolism, surgical site infections, and unplanned reoperation in a secondary analysis of the American College of Surgeons prospective National Surgical Quality Improvement database.
Adequate exposure of residents to emergency general surgery is crucial for surgical training, but academic operating teams should be mindful of this association, Dr. George Kasotakis said at the annual meeting of the American Surgical Association.
Three papers have shown resident participation modestly increases complications in elective surgery, but their impact is not well understood for emergency general surgery, where patient physiology is typically deranged and timely surgery is imperative.
Dr. Kasotakis and his colleagues at Boston University identified 141,010 patients who underwent emergency general surgery procedures in the 2005-2010 American College of Surgeons prospective National Surgical Quality Improvement database. Because of the nonrandom assignment of more complex cases to resident participation, patients were matched 1:1 based on age; gender; use of alcohol, tobacco, and steroids; inpatient status; obesity; diabetes; renal failure; cardiopulmonary disease; and expected probability for morbidity and mortality. Regression models were fitted for each outcome and adjusted for the same risk factors and operative time.
The most common procedures were appendectomy (40%), exploratory laparotomy (8.75%), bowel resection (9.2%), cholecystectomy (6%), and lysis of adhesions (6%).
Thirty-day mortality was similar with and without residents (3.25% vs. 2.96%; P = .082), but hospital length of stay was longer by about a half a day with residents (4.97 days vs. 4.59 days; P = .019), said Dr. Kasotakis, an acute care surgeon and intensivist.
Resident participation added about 20 minutes to operative (75 minutes vs. 59 minutes; P less than .001) and anesthesia (122 minutes vs. 100 minutes; P less than .001) times.
Intraoperative transfusions were more common with residents (3.43% vs. 2.55%; P less than .001), perhaps because of longer operating room times, and, as a result, fewer postoperative transfusions were needed (1.12% vs. 1.28% P = .031), he said. Unplanned reoperations, however, were more common with residents, as well (4.22% vs. 3.80%; P = .002).
Postsurgical superficial wound infections (3.5% vs. 2.78%; P less than .001) and organ space surgical site infections (2.27% vs. 1.77%; P less than .001) were more common in the resident group, while wound dehiscence was not (0.63% vs. 0.69%; P = .266), Dr. Kasotakis noted.
Pulmonary complications were significantly more common in the resident group including postoperative pneumonia (1.85% vs. 1.67%; P = .04), reintubation (1.64% vs. 1.15%; P less than .001), and mechanical ventilation for more than 48 hours (2.87% vs. 2.06%; P less than .001).
The same was true for deep vein thrombosis (DVT) (0.80% vs. 0.62%; P = .002) and pulmonary embolism (PE) (0.43% vs. 0.28%; P less than .001).
Urinary tract infections (UTI) were higher with resident participation (1.45% vs. 1.14%; P less than .001), as was sepsis (2.42% vs. 2.13%; P = .005), likely because of the increase in surgical infections, Dr. Kasotakis said.
Thankfully, significant cardiac complications and septic shock were not more common with residents, he said.
Adjusted analyses
After adjustment for operative duration, case complexity and pre-existing comorbidities, residents did not increase length of stay (odds ratio, 0.07; P = .242) or septic events (OR, 1.07; P = .155), but their participation was still independently associated with about 20% more superficial surgical site infections (odds ratio, 1.23; P less than .001), organ space infections (OR, 1.21; P less than .001), UTIs (OR, 1.23; P = .001), and intraoperative transfusions (OR, 1.20; P = .001), he said.
Also, about 8% more patients required a return trip to the operating room when residents participated (OR, 1.08; P = .041).
"These outcomes can perhaps be attributed to their underdeveloped surgical skills," Dr. Kasotakis said.
The incidence of DVT and PE were also higher by about 25% (OR, 1.25; P = .011) and 40% (OR, 1.42; P = .005), respectively, perhaps because of delayed DVT prophylaxis initiation because of concerns of hemostasis or missed doses due to additional return trips to the emergency department, he suggested.
Interestingly, reintubation and prolonged mechanical ventilation rates were increased by about 40% (OR, 1.38; OR, 1.43; both P less than .001), perhaps because of prolonged operative times or greater resuscitation requirements, he added.
Dr. Kasotakis was quick to point out that this was a secondary analysis of a data set not originally intended to assess the effect of trainee participation, that no information was available on the degree of resident involvement during surgery or in perioperative care, and that participating institutions were skewed toward tertiary centers, which typically receive more complex cases.
"Staff surgeons should supervise as needed and minimize unnecessary [emergency department] time. And residents, for their part, should be well prepared for emergency procedures through simulation training and aim to maximize their operating room efficiency," he suggested.
The results sparked a flurry of rebuttals led off by discussant Dr. Julie Ann Sosa, Duke University, Durham, N.C.,who said they conflict with other analyses showing little to no impact from residents in elective cases.
"If not interpreted with care, policy makers, payers, and the public could construe that surgical care at academic health centers is compromised by trainees, which could have unfortunate ramifications for everyone in the room as well as the trainees and the patients," she said.
Dr. Sosa expressed concern about drawing causal inferences from an observational study in the setting of possible selection bias and said attempts to match for case complexity using CPT codes do not necessarily account for say, "the difference between a routine appendectomy that takes 15-30 minutes and a complex one that takes 3 hours for a perforation."
Some attendees questioned why the authors didn’t match the institutions in the analysis and chose to ascribe all of the outcomes to residents, with a round of applause following the suggestion that the paper should be titled "Academic centers increase emergency surgery complications." Other attendees questioned whether the poor outcomes reflect resident training and supervision.
Dr. George Velmahos, Massachusetts General Hospital, Boston, questioned whether hospitals have a medical/legal responsibility to inform patients that a resident is in the operating room and may impact outcomes.
Dr. Kasotakis said that institutions may want to add a clause to consent paperwork stating that residents and trainees will be participating.
The complete manuscript of this study and its presentation at the American Surgical Association’s 134th Annual Meeting is anticipated to be published in the Annals of Surgery, pending editorial review.
Dr. Kasotakis reported no conflicts.
BOSTON – Resident participation in emergency general surgery cases was independently associated with a host of complications – pulmonary embolism, surgical site infections, and unplanned reoperation in a secondary analysis of the American College of Surgeons prospective National Surgical Quality Improvement database.
Adequate exposure of residents to emergency general surgery is crucial for surgical training, but academic operating teams should be mindful of this association, Dr. George Kasotakis said at the annual meeting of the American Surgical Association.
Three papers have shown resident participation modestly increases complications in elective surgery, but their impact is not well understood for emergency general surgery, where patient physiology is typically deranged and timely surgery is imperative.
Dr. Kasotakis and his colleagues at Boston University identified 141,010 patients who underwent emergency general surgery procedures in the 2005-2010 American College of Surgeons prospective National Surgical Quality Improvement database. Because of the nonrandom assignment of more complex cases to resident participation, patients were matched 1:1 based on age; gender; use of alcohol, tobacco, and steroids; inpatient status; obesity; diabetes; renal failure; cardiopulmonary disease; and expected probability for morbidity and mortality. Regression models were fitted for each outcome and adjusted for the same risk factors and operative time.
The most common procedures were appendectomy (40%), exploratory laparotomy (8.75%), bowel resection (9.2%), cholecystectomy (6%), and lysis of adhesions (6%).
Thirty-day mortality was similar with and without residents (3.25% vs. 2.96%; P = .082), but hospital length of stay was longer by about a half a day with residents (4.97 days vs. 4.59 days; P = .019), said Dr. Kasotakis, an acute care surgeon and intensivist.
Resident participation added about 20 minutes to operative (75 minutes vs. 59 minutes; P less than .001) and anesthesia (122 minutes vs. 100 minutes; P less than .001) times.
Intraoperative transfusions were more common with residents (3.43% vs. 2.55%; P less than .001), perhaps because of longer operating room times, and, as a result, fewer postoperative transfusions were needed (1.12% vs. 1.28% P = .031), he said. Unplanned reoperations, however, were more common with residents, as well (4.22% vs. 3.80%; P = .002).
Postsurgical superficial wound infections (3.5% vs. 2.78%; P less than .001) and organ space surgical site infections (2.27% vs. 1.77%; P less than .001) were more common in the resident group, while wound dehiscence was not (0.63% vs. 0.69%; P = .266), Dr. Kasotakis noted.
Pulmonary complications were significantly more common in the resident group including postoperative pneumonia (1.85% vs. 1.67%; P = .04), reintubation (1.64% vs. 1.15%; P less than .001), and mechanical ventilation for more than 48 hours (2.87% vs. 2.06%; P less than .001).
The same was true for deep vein thrombosis (DVT) (0.80% vs. 0.62%; P = .002) and pulmonary embolism (PE) (0.43% vs. 0.28%; P less than .001).
Urinary tract infections (UTI) were higher with resident participation (1.45% vs. 1.14%; P less than .001), as was sepsis (2.42% vs. 2.13%; P = .005), likely because of the increase in surgical infections, Dr. Kasotakis said.
Thankfully, significant cardiac complications and septic shock were not more common with residents, he said.
Adjusted analyses
After adjustment for operative duration, case complexity and pre-existing comorbidities, residents did not increase length of stay (odds ratio, 0.07; P = .242) or septic events (OR, 1.07; P = .155), but their participation was still independently associated with about 20% more superficial surgical site infections (odds ratio, 1.23; P less than .001), organ space infections (OR, 1.21; P less than .001), UTIs (OR, 1.23; P = .001), and intraoperative transfusions (OR, 1.20; P = .001), he said.
Also, about 8% more patients required a return trip to the operating room when residents participated (OR, 1.08; P = .041).
"These outcomes can perhaps be attributed to their underdeveloped surgical skills," Dr. Kasotakis said.
The incidence of DVT and PE were also higher by about 25% (OR, 1.25; P = .011) and 40% (OR, 1.42; P = .005), respectively, perhaps because of delayed DVT prophylaxis initiation because of concerns of hemostasis or missed doses due to additional return trips to the emergency department, he suggested.
Interestingly, reintubation and prolonged mechanical ventilation rates were increased by about 40% (OR, 1.38; OR, 1.43; both P less than .001), perhaps because of prolonged operative times or greater resuscitation requirements, he added.
Dr. Kasotakis was quick to point out that this was a secondary analysis of a data set not originally intended to assess the effect of trainee participation, that no information was available on the degree of resident involvement during surgery or in perioperative care, and that participating institutions were skewed toward tertiary centers, which typically receive more complex cases.
"Staff surgeons should supervise as needed and minimize unnecessary [emergency department] time. And residents, for their part, should be well prepared for emergency procedures through simulation training and aim to maximize their operating room efficiency," he suggested.
The results sparked a flurry of rebuttals led off by discussant Dr. Julie Ann Sosa, Duke University, Durham, N.C.,who said they conflict with other analyses showing little to no impact from residents in elective cases.
"If not interpreted with care, policy makers, payers, and the public could construe that surgical care at academic health centers is compromised by trainees, which could have unfortunate ramifications for everyone in the room as well as the trainees and the patients," she said.
Dr. Sosa expressed concern about drawing causal inferences from an observational study in the setting of possible selection bias and said attempts to match for case complexity using CPT codes do not necessarily account for say, "the difference between a routine appendectomy that takes 15-30 minutes and a complex one that takes 3 hours for a perforation."
Some attendees questioned why the authors didn’t match the institutions in the analysis and chose to ascribe all of the outcomes to residents, with a round of applause following the suggestion that the paper should be titled "Academic centers increase emergency surgery complications." Other attendees questioned whether the poor outcomes reflect resident training and supervision.
Dr. George Velmahos, Massachusetts General Hospital, Boston, questioned whether hospitals have a medical/legal responsibility to inform patients that a resident is in the operating room and may impact outcomes.
Dr. Kasotakis said that institutions may want to add a clause to consent paperwork stating that residents and trainees will be participating.
The complete manuscript of this study and its presentation at the American Surgical Association’s 134th Annual Meeting is anticipated to be published in the Annals of Surgery, pending editorial review.
Dr. Kasotakis reported no conflicts.
Major finding: Rates were significantly higher with resident participation than without for DVT (0.80% vs. 0.62%; P = .002) and PE (0.43% vs. 0.28%; P less than .001).
Data source: A secondary analysis of 141,010 emergency general surgery cases in the prospective ACS NSQIP database.
Disclosures: Dr. Kasotakis and his coauthors reported no conflicting interests.
