Spine
SPINE Volume 39, Number 7, pp 603-611 ©2014, Lippincott Williams & Wilkins
HEALTH SERVICES RESEARCH
July Effecf' in Elective Spine Surgery Analysis of the American College of Surgeons National Surgical Quality Improvement Program Database Daniel D. Bohl, MPH, Michael C. Fu, BS, lordan A. Gruskay, BA, Bryce A. Basques, BS, Nicholas S. Golinvaux, BA, and Jonathan N. Grauer, MD
Study Design. Retrospective cohort. Objective. To evaluate for the presence and magnitude of the "July effect" within elective spine surgery. Summary of Background Data. The July effect is the hypothetical increase in morbidity and mortality thought to be associated with the influx of new (or newly promoted) trainees during the first portion of the academic year. Studies evaluating for the presence and magnitude of the July effect have demonstrated conflicting results. Methods. We accessed the American College of Surgeons National Surgical Quality Improvement Program database from 2005-2010. Statistical analyses were conducted using bivariate and multivariate logistic regression. Results. A total of 14,986 cases met inclusion criteria and constitute the study population. Of these, 26.5% occurred in the first academic quarter and 25.3% had resident involvement. The rate of serious adverse events was 1.9 times higher and the rate of any adverse events was 1.6 times higher among cases with resident involvement than among those without (P < 0.001 for both). Among cases without resident involvement, the rates of serious adverse events and any adverse events did not differ by academic quarter. Similarly, among cases with resident involvement, the rates of serious adverse events and any adverse events did not differ by academic quarter. Conclusion. We could not demonstrate that the training of new (or newly promoted) residents is associated with an increase in the adverse events of spine surgery. Safeguards that have been put in place to ensure patient safety during this training period seem to be effective. Although adverse events were more common among cases with resident involvement than among cases without resident involvement, our data suggest that this association is more likely
From the Yale School of Medicine, New Haven, CT. Acknowledgment date: September 17, 2013. Revision date; December 15, 2013. Acceptance date: December 16, 2013. The manuscript submitted does not contain information about medical device(s)/drug(s). No funds were received in support of this work. Relevant financial activities outside the submitted work: consultancy, expert testimony, grants. Address correspondence and reprint requests to Jonathan N. Grauer, MD, Yale School of Medicine, 800 Howard Ave., New Haven, CT 06510; E-mail: [email protected] DOl: 10.1097/BRS.0000000000000196 Spine
a product of the riskier population of cases in which residents participate than of the resident involvement itself. Key words: July effect, elective spine surgery, seasonal variation, resident training, adverse events.
Level of Evidence: 3 Spine 2014;39:603-611
R
esident education is integral to the medical system. Often cited as analogous to an apprenticeship, this is the way that practical, hands-on education is imparted to those entering their chosen fields. Although there have been adjustments to the resident education process over the years, these principles are paramount to training the next generation of physicians and surgeons. Every year in July, senior residents graduate from their respective training programs and are replaced by new (or newly promoted) recruits. These residents may lack some of the knowledge, experience, decision-making capacity, and technical skill of their predecessors. Conventional wisdom has provided that this time of year is a poor time to be hospitalized, and particularly to have surgery. The popular literature has recently given increased attention to this concern.'-^ Regulations regarding resident work hours highlight how seriously the public and professional societies take the potential that physician training could negatively impact patient care. These regulations not only place strains on residency programs to prepare their trainees and on hospitals to care for their patients, but also have tremendous financial implications for the country as a whole. Indeed, the regulations proposed by the 2001 Accreditation Council for Graduate Medical Education alone are estimated to cost the US medical system between $820 million and $1.64 billion per year.^ The evidence on whether the yearly transition from more experienced to less experienced trainees is responsible for an increase in medical errors is mixed. In favor of this phenomenon, a study of all US death certificates from 1979 to 2006 found that in counties with teaching hospitals, fatal medication errors spiked by 10% in July relative to other months.'* Conversely, there was no July spike in counties without teaching hospitals, and the greater the concentration of teaching hospitals in a region, the greater the July spike. www.spinejournal.com
603
HEALTH SERVICES RESEARCH
Similarly, a study examining all patients from 2001 to 2004 available through the American College of Surgeons National Surgical Quality Improvement Program (ACS NSQIP) database found that there was an 18% higher risk of postoperative morbidity and a 41% higher risk of mortality across all major surgical procedures between July 1 and August 30 than between April 15 and June 15.^ Conversely, a recent study using the same database found that during the first academic quarter, the mortality rate, "serious adverse event" (SAE) rate, and morbidity rate among the 10 most common inpatient operative procedures from 2005-2007 was not different between the first academic quarter (July, August, and September) as compared with the rest of the year.* More than 30 other studies span this range of findings and there is no consensus.^ Within orthopedic surgery, less work has been done. One study of 722 patients sustaining femoral neck or intertrochanteric fractures between 2000 and 2010 at a single institution found that the mortality rate and complication rate were not different during July and August compared with the rest of the year.^ On the contrary, a study using the National Inpatient Sample (NIS) from 1988 to 2003 found that the relative risk of in-hospital mortality after hip fracture during July and August [vs. the remainder of the year) was significantly higher at teaching hospitals than at nonteaching hospitals.' Three studies have examined the "July effect" within spine surgery. Recently, one group of authors using the NIS studied spine surgery hospitalizations from 2001 to 2008 and showed that, compared with patients admitted in other months, patients who were admitted in July did not show a difference in in-hospital mortality, reaction to implanted device/instrumentation, or postoperative wound dehiscence.'" However, a study using the same database found that patients undergoing surgery for metastatic spinal disease at teaching hospitals in July had higher rates of in-hospital mortality and intraoperative complications." Additionally, a group examined spine surgery cases at a single tertiary referral institution between January 2005 and December 2009 and showed that spine surgical procedures performed during the summer and fall months were associated with a significantly higher incidence of wound infection than the winter and spring.'^ Given the heterogeneity of results both within and outside of orthopedic and spine surgery, and given the great interest on the part of the public, the patient, hospital administrators, and residency programs, we queried a widely used national database (ACS NSQIP) to investigate whether the rates of adverse events associated with spine surgery and the subsequent 30 days are higher during the first part of the academic year.
MATERIALS AND METHODS To conduct this study, we accessed the ACS NSQIP database, which is available to our institution as a data-contributing member. The ACS NSQIP is a prospective, multi-institutional outcomes program, the details of which are available elsewhere.''•" Briefly, the ACS NSQIP collects data on 135 Health Insurance Portability and Accountability Act-compliant 604
www.spinejournaLcom
"July Effect" in Elective Spine Surgery • Bohl et al
variables, including preoperative comorbidities, intraoperative variables, and 30-day postoperative morbidity and mortality outcomes for patients undergoing surgical procedures in both the inpatient and outpatient settings. We studied data from 2005 to 2010, the 6 most recent years for which data were available. Using Current Procedural Terminology codes, we identified all patients who underwent 1 of 7 different spine procedures: anterior cervical discectomy/fusion, posterior cervical fusion, anterior lumbar fusion, posterior lumbar fusion, transforaminal lumbar interbody fusion/posterior lumbar interbody fusion, lumbar discectomy, and lumbar decompression. To focus on the typical patient undergoing elective spine surgery, and to minimize potential seasonal confounders, we excluded from the analysis patients who underwent combination lumbar anterior and posterior approaches, patients who underwent spinal surgery for deformity, patients who had simultaneous nonorthopedic surgical procedures (/.e., appendectomy), and patients undergoing emergency surgery. Academic quarter and resident involvement were variables that were extracted directly from the data set. American Society of Anesthesiologists (ASA) score was also directly extracted. A variable for complex case was generated on the basis of whether or not there were 4 or more Current Procedural Terminology codes used. Each patient was also assigned a comorbidity score based on the Charlson Comorbidity Index (CCI) modified to fit the available data.'^ Studies have demonstrated that such modified CCIs have been similar in efficiency and prognosis to the original CCI,'*-'^ and the modified CCI we used has been previously used with the ACS NSQIP dataset.*" The comorbidities used to determine the modified CCI were chronic obstructive pulmonary disease, esophageal varices, ascites, peripheral vascular disease, cerebrovascular disease, hemiplegia, myocardial infarction, congestive heart failure, end-stage renal disease, dementia, diabetes mellitus, cancer, and age. Each of these comorbidities was assigned a specific point value; the values were summed for a total number indicative of each patient's overall 10-year risk of mortality. Two composite and 23 individual adverse event outcomes were studied. The first composite outcome was a SAE. Consistent with previous definition,*^ we defined this to be the occurrence of any of the following: mortality, coma more than 24 hours, on ventilator more than 48 hours, unplanned intubation, stroke/cerebrovascular accident, pulmonary embolism, cardiac arrest, myocardial infarction, acute renal failure, sepsis, septic shock, or return to the operating room. The second composite outcome was "any adverse event" (AAE), which we defined as the occurrence of any of the following: any of the SAEs listed in the prior sentence, wound disruption, superficial surgical site infection, deep surgical site infection, organ/ space infection (other than surgical site), urinary tract infection, pneumonia, transfusion of 5 or more units of packed red blood cells, progressive renal insufficiency, graft/prosthesis/ flap failure, peripheral nerve injury, and deep vein thrombosis/thrombophlebitis. The individual outcomes were the 23 individual adverse events listed in the earlier text. Statistical April 2014
HEALTH SERVICES RESEARCH
analyses were conducted using STATA version 11.2 (StataCorp, LP, College Station, TX). All tests were 2 tailed, and the level of significance was set at F = 0.05. First, rates of resident involvement in cases were stratified and compared by procedure type, modified CCI, ASA score, complexity of case, and academic quarter using bivariate logistic regression. Second, the rates of SAEs and AAEs were compared between cases with and without residents using bivariate logistic regression. The third set of analyses was conducted first among cases without resident involvement and then among cases with resident involvement. The rates of SAEs and AAEs were compared between cases occurring during the first academic quarter and cases occurring during the rest of the year. These comparisons were conducted using multivariate logistic regression with controls for procedure type, modified CCI, ASA score, and complexity of case. Rates of individual adverse events were compared in an analogous manner; however, because some of these individual adverse events were very rare, such comparisons were only made for event types that had 30 or more events occur. To avoid a type 1 error, for the regressions on the individual adverse events, Bonferroni adjustments were made, lowering the P values required for significance to P = 0.005 for regressions among cases without resident involvement and P = 0.007 for regressions among cases with resident involvement.
RESULTS A total of 14,986 cases met inclusion criteria and constitute the study population. Of these, 26.5% occurred in the first academic quarter and 25.3% had resident involvement. The distribution of the 7 types of spine procedures by academic quarter is shown in Eigure 1. Rates of resident involvement in cases varied by procedure type, modified CCI, ASA score, complexity of case, and academic quarter (Table 1). Specifically, cases with higher modified CCI, higher ASA score, and greater complexity were more likely to have residents involved (P < 0.001 for each). Cases that occurred during the first academic quarter were more likely to have residents involved than those that occurred during the rest of the year (26.5% vs. 24.8%, P = 0.038). The rate of SAEs was 1.9 times higher among cases with resident involvement than among cases without (6.7% vs. 3.6%, respectively; P < 0.001; Eigure 2). Similarly, the rate of AAEs was 1.6 times higher among cases with resident involvement than among cases without (13.8% vs. 8.1%, respectively; P < 0.001; Eigure 3). Among cases without resident involvement, the rates of SAEs and AAEs did not differ by academic quarter (P = 0.612 and P = 0.209, respectively; Table 2). Similarly, the rates of each of the individual adverse events did not differ by academic quarter (Table 3). Among cases with resident involvement, the rates of SAEs and AAEs did not differ by academic quarter (P = 0.096 and P = 0.157, respectively; Table 4). Similarly, the rates of each of the individual adverse events did not differ by academic quarter (Table 5). Spine
"July Effect" in Elective Spine Surgery • Bohl et al
3025201510o
Q. O
50 Jul-Sep
Oct-Dec
Jan-Mar
Apr-Jun
— — — - - Lumbar discectomy • — ^ ^ ^ — Anterior cervical discectomy/fusion Lumbar decompression _._._>—
Posterior lumbar fusion Transforaminal/posterior lumbar interbody fusion
— — — — - Posterior cervicai fusion Anterior lumbar fusion
Figure 1. Distribution of 7 types of spine procedures by academic quarter.
DISCUSSION The July effect is the hypothetical increase in morbidity and mortality thought to be associated with the influx of new (or newly promoted) trainees during the first portion of the academic year. Studies evaluating for the presence and magnitude of the July effect within both medicine and surgery have been conflicted in their results.''"'^ We evaluated for the presence and magnitude of the July effect within spine surgery using the ACS NSQIP, a database for which such an analysis has not been previously presented. We found that there is no seasonal increase in the rate of adverse events; that is, in the manner that the expression is most commonly used, we found that in spine surgery, the July effect does not occur. Specifically, SAEs and AAEs were no more likely to occur during the first academic quarter than during the rest of the year, whether the population was restricted to only cases without or only cases with residents. The composite outcomes (SAEs and AAEs) could miss specific increases in individual adverse events, including mortality, so using identical methods (with statistical adjustment to avoid type I error), we evaluated for seasonal changes in the composite outcomes' components and demonstrated that there were no such associations, reinforcing our primary result. As a result, it seems that safeguards that have been put into place to minimize the number and impact of any errors have been effective. The most visible of these safeguards is the resident duty hour restrictions imposed by the Accreditation Council for Graduate Medical Education in 2003.'* These regulations limit the work week to 80 hours, require a 10-hour rest period between duty periods, and limit continuous duty to 24 hours. Similarly, in 2004, the Joint Commission imposed the Universal Protocol, a set of regulations designed to make surgery safer." As part of the Universal Protocol, the attending surgeon is required to be present www.spinejournal.com
605
HEALTH SERVICES RESEARCH
TABLE!.
"July Effect" in Elective Spine Surgery • Bohl et al
tes of Resident Involvement in Resident Involvement Rate* (%)
Overall
ORt
P
25.3
> 0 Q) CO >
Procedure type
SPINE Volume 39, Number 7, pp 603-611 ©2014, Lippincott Williams & Wilkins
HEALTH SERVICES RESEARCH
July Effecf' in Elective Spine Surgery Analysis of the American College of Surgeons National Surgical Quality Improvement Program Database Daniel D. Bohl, MPH, Michael C. Fu, BS, lordan A. Gruskay, BA, Bryce A. Basques, BS, Nicholas S. Golinvaux, BA, and Jonathan N. Grauer, MD
Study Design. Retrospective cohort. Objective. To evaluate for the presence and magnitude of the "July effect" within elective spine surgery. Summary of Background Data. The July effect is the hypothetical increase in morbidity and mortality thought to be associated with the influx of new (or newly promoted) trainees during the first portion of the academic year. Studies evaluating for the presence and magnitude of the July effect have demonstrated conflicting results. Methods. We accessed the American College of Surgeons National Surgical Quality Improvement Program database from 2005-2010. Statistical analyses were conducted using bivariate and multivariate logistic regression. Results. A total of 14,986 cases met inclusion criteria and constitute the study population. Of these, 26.5% occurred in the first academic quarter and 25.3% had resident involvement. The rate of serious adverse events was 1.9 times higher and the rate of any adverse events was 1.6 times higher among cases with resident involvement than among those without (P < 0.001 for both). Among cases without resident involvement, the rates of serious adverse events and any adverse events did not differ by academic quarter. Similarly, among cases with resident involvement, the rates of serious adverse events and any adverse events did not differ by academic quarter. Conclusion. We could not demonstrate that the training of new (or newly promoted) residents is associated with an increase in the adverse events of spine surgery. Safeguards that have been put in place to ensure patient safety during this training period seem to be effective. Although adverse events were more common among cases with resident involvement than among cases without resident involvement, our data suggest that this association is more likely
From the Yale School of Medicine, New Haven, CT. Acknowledgment date: September 17, 2013. Revision date; December 15, 2013. Acceptance date: December 16, 2013. The manuscript submitted does not contain information about medical device(s)/drug(s). No funds were received in support of this work. Relevant financial activities outside the submitted work: consultancy, expert testimony, grants. Address correspondence and reprint requests to Jonathan N. Grauer, MD, Yale School of Medicine, 800 Howard Ave., New Haven, CT 06510; E-mail: [email protected] DOl: 10.1097/BRS.0000000000000196 Spine
a product of the riskier population of cases in which residents participate than of the resident involvement itself. Key words: July effect, elective spine surgery, seasonal variation, resident training, adverse events.
Level of Evidence: 3 Spine 2014;39:603-611
R
esident education is integral to the medical system. Often cited as analogous to an apprenticeship, this is the way that practical, hands-on education is imparted to those entering their chosen fields. Although there have been adjustments to the resident education process over the years, these principles are paramount to training the next generation of physicians and surgeons. Every year in July, senior residents graduate from their respective training programs and are replaced by new (or newly promoted) recruits. These residents may lack some of the knowledge, experience, decision-making capacity, and technical skill of their predecessors. Conventional wisdom has provided that this time of year is a poor time to be hospitalized, and particularly to have surgery. The popular literature has recently given increased attention to this concern.'-^ Regulations regarding resident work hours highlight how seriously the public and professional societies take the potential that physician training could negatively impact patient care. These regulations not only place strains on residency programs to prepare their trainees and on hospitals to care for their patients, but also have tremendous financial implications for the country as a whole. Indeed, the regulations proposed by the 2001 Accreditation Council for Graduate Medical Education alone are estimated to cost the US medical system between $820 million and $1.64 billion per year.^ The evidence on whether the yearly transition from more experienced to less experienced trainees is responsible for an increase in medical errors is mixed. In favor of this phenomenon, a study of all US death certificates from 1979 to 2006 found that in counties with teaching hospitals, fatal medication errors spiked by 10% in July relative to other months.'* Conversely, there was no July spike in counties without teaching hospitals, and the greater the concentration of teaching hospitals in a region, the greater the July spike. www.spinejournal.com
603
HEALTH SERVICES RESEARCH
Similarly, a study examining all patients from 2001 to 2004 available through the American College of Surgeons National Surgical Quality Improvement Program (ACS NSQIP) database found that there was an 18% higher risk of postoperative morbidity and a 41% higher risk of mortality across all major surgical procedures between July 1 and August 30 than between April 15 and June 15.^ Conversely, a recent study using the same database found that during the first academic quarter, the mortality rate, "serious adverse event" (SAE) rate, and morbidity rate among the 10 most common inpatient operative procedures from 2005-2007 was not different between the first academic quarter (July, August, and September) as compared with the rest of the year.* More than 30 other studies span this range of findings and there is no consensus.^ Within orthopedic surgery, less work has been done. One study of 722 patients sustaining femoral neck or intertrochanteric fractures between 2000 and 2010 at a single institution found that the mortality rate and complication rate were not different during July and August compared with the rest of the year.^ On the contrary, a study using the National Inpatient Sample (NIS) from 1988 to 2003 found that the relative risk of in-hospital mortality after hip fracture during July and August [vs. the remainder of the year) was significantly higher at teaching hospitals than at nonteaching hospitals.' Three studies have examined the "July effect" within spine surgery. Recently, one group of authors using the NIS studied spine surgery hospitalizations from 2001 to 2008 and showed that, compared with patients admitted in other months, patients who were admitted in July did not show a difference in in-hospital mortality, reaction to implanted device/instrumentation, or postoperative wound dehiscence.'" However, a study using the same database found that patients undergoing surgery for metastatic spinal disease at teaching hospitals in July had higher rates of in-hospital mortality and intraoperative complications." Additionally, a group examined spine surgery cases at a single tertiary referral institution between January 2005 and December 2009 and showed that spine surgical procedures performed during the summer and fall months were associated with a significantly higher incidence of wound infection than the winter and spring.'^ Given the heterogeneity of results both within and outside of orthopedic and spine surgery, and given the great interest on the part of the public, the patient, hospital administrators, and residency programs, we queried a widely used national database (ACS NSQIP) to investigate whether the rates of adverse events associated with spine surgery and the subsequent 30 days are higher during the first part of the academic year.
MATERIALS AND METHODS To conduct this study, we accessed the ACS NSQIP database, which is available to our institution as a data-contributing member. The ACS NSQIP is a prospective, multi-institutional outcomes program, the details of which are available elsewhere.''•" Briefly, the ACS NSQIP collects data on 135 Health Insurance Portability and Accountability Act-compliant 604
www.spinejournaLcom
"July Effect" in Elective Spine Surgery • Bohl et al
variables, including preoperative comorbidities, intraoperative variables, and 30-day postoperative morbidity and mortality outcomes for patients undergoing surgical procedures in both the inpatient and outpatient settings. We studied data from 2005 to 2010, the 6 most recent years for which data were available. Using Current Procedural Terminology codes, we identified all patients who underwent 1 of 7 different spine procedures: anterior cervical discectomy/fusion, posterior cervical fusion, anterior lumbar fusion, posterior lumbar fusion, transforaminal lumbar interbody fusion/posterior lumbar interbody fusion, lumbar discectomy, and lumbar decompression. To focus on the typical patient undergoing elective spine surgery, and to minimize potential seasonal confounders, we excluded from the analysis patients who underwent combination lumbar anterior and posterior approaches, patients who underwent spinal surgery for deformity, patients who had simultaneous nonorthopedic surgical procedures (/.e., appendectomy), and patients undergoing emergency surgery. Academic quarter and resident involvement were variables that were extracted directly from the data set. American Society of Anesthesiologists (ASA) score was also directly extracted. A variable for complex case was generated on the basis of whether or not there were 4 or more Current Procedural Terminology codes used. Each patient was also assigned a comorbidity score based on the Charlson Comorbidity Index (CCI) modified to fit the available data.'^ Studies have demonstrated that such modified CCIs have been similar in efficiency and prognosis to the original CCI,'*-'^ and the modified CCI we used has been previously used with the ACS NSQIP dataset.*" The comorbidities used to determine the modified CCI were chronic obstructive pulmonary disease, esophageal varices, ascites, peripheral vascular disease, cerebrovascular disease, hemiplegia, myocardial infarction, congestive heart failure, end-stage renal disease, dementia, diabetes mellitus, cancer, and age. Each of these comorbidities was assigned a specific point value; the values were summed for a total number indicative of each patient's overall 10-year risk of mortality. Two composite and 23 individual adverse event outcomes were studied. The first composite outcome was a SAE. Consistent with previous definition,*^ we defined this to be the occurrence of any of the following: mortality, coma more than 24 hours, on ventilator more than 48 hours, unplanned intubation, stroke/cerebrovascular accident, pulmonary embolism, cardiac arrest, myocardial infarction, acute renal failure, sepsis, septic shock, or return to the operating room. The second composite outcome was "any adverse event" (AAE), which we defined as the occurrence of any of the following: any of the SAEs listed in the prior sentence, wound disruption, superficial surgical site infection, deep surgical site infection, organ/ space infection (other than surgical site), urinary tract infection, pneumonia, transfusion of 5 or more units of packed red blood cells, progressive renal insufficiency, graft/prosthesis/ flap failure, peripheral nerve injury, and deep vein thrombosis/thrombophlebitis. The individual outcomes were the 23 individual adverse events listed in the earlier text. Statistical April 2014
HEALTH SERVICES RESEARCH
analyses were conducted using STATA version 11.2 (StataCorp, LP, College Station, TX). All tests were 2 tailed, and the level of significance was set at F = 0.05. First, rates of resident involvement in cases were stratified and compared by procedure type, modified CCI, ASA score, complexity of case, and academic quarter using bivariate logistic regression. Second, the rates of SAEs and AAEs were compared between cases with and without residents using bivariate logistic regression. The third set of analyses was conducted first among cases without resident involvement and then among cases with resident involvement. The rates of SAEs and AAEs were compared between cases occurring during the first academic quarter and cases occurring during the rest of the year. These comparisons were conducted using multivariate logistic regression with controls for procedure type, modified CCI, ASA score, and complexity of case. Rates of individual adverse events were compared in an analogous manner; however, because some of these individual adverse events were very rare, such comparisons were only made for event types that had 30 or more events occur. To avoid a type 1 error, for the regressions on the individual adverse events, Bonferroni adjustments were made, lowering the P values required for significance to P = 0.005 for regressions among cases without resident involvement and P = 0.007 for regressions among cases with resident involvement.
RESULTS A total of 14,986 cases met inclusion criteria and constitute the study population. Of these, 26.5% occurred in the first academic quarter and 25.3% had resident involvement. The distribution of the 7 types of spine procedures by academic quarter is shown in Eigure 1. Rates of resident involvement in cases varied by procedure type, modified CCI, ASA score, complexity of case, and academic quarter (Table 1). Specifically, cases with higher modified CCI, higher ASA score, and greater complexity were more likely to have residents involved (P < 0.001 for each). Cases that occurred during the first academic quarter were more likely to have residents involved than those that occurred during the rest of the year (26.5% vs. 24.8%, P = 0.038). The rate of SAEs was 1.9 times higher among cases with resident involvement than among cases without (6.7% vs. 3.6%, respectively; P < 0.001; Eigure 2). Similarly, the rate of AAEs was 1.6 times higher among cases with resident involvement than among cases without (13.8% vs. 8.1%, respectively; P < 0.001; Eigure 3). Among cases without resident involvement, the rates of SAEs and AAEs did not differ by academic quarter (P = 0.612 and P = 0.209, respectively; Table 2). Similarly, the rates of each of the individual adverse events did not differ by academic quarter (Table 3). Among cases with resident involvement, the rates of SAEs and AAEs did not differ by academic quarter (P = 0.096 and P = 0.157, respectively; Table 4). Similarly, the rates of each of the individual adverse events did not differ by academic quarter (Table 5). Spine
"July Effect" in Elective Spine Surgery • Bohl et al
3025201510o
Q. O
50 Jul-Sep
Oct-Dec
Jan-Mar
Apr-Jun
— — — - - Lumbar discectomy • — ^ ^ ^ — Anterior cervical discectomy/fusion Lumbar decompression _._._>—
Posterior lumbar fusion Transforaminal/posterior lumbar interbody fusion
— — — — - Posterior cervicai fusion Anterior lumbar fusion
Figure 1. Distribution of 7 types of spine procedures by academic quarter.
DISCUSSION The July effect is the hypothetical increase in morbidity and mortality thought to be associated with the influx of new (or newly promoted) trainees during the first portion of the academic year. Studies evaluating for the presence and magnitude of the July effect within both medicine and surgery have been conflicted in their results.''"'^ We evaluated for the presence and magnitude of the July effect within spine surgery using the ACS NSQIP, a database for which such an analysis has not been previously presented. We found that there is no seasonal increase in the rate of adverse events; that is, in the manner that the expression is most commonly used, we found that in spine surgery, the July effect does not occur. Specifically, SAEs and AAEs were no more likely to occur during the first academic quarter than during the rest of the year, whether the population was restricted to only cases without or only cases with residents. The composite outcomes (SAEs and AAEs) could miss specific increases in individual adverse events, including mortality, so using identical methods (with statistical adjustment to avoid type I error), we evaluated for seasonal changes in the composite outcomes' components and demonstrated that there were no such associations, reinforcing our primary result. As a result, it seems that safeguards that have been put into place to minimize the number and impact of any errors have been effective. The most visible of these safeguards is the resident duty hour restrictions imposed by the Accreditation Council for Graduate Medical Education in 2003.'* These regulations limit the work week to 80 hours, require a 10-hour rest period between duty periods, and limit continuous duty to 24 hours. Similarly, in 2004, the Joint Commission imposed the Universal Protocol, a set of regulations designed to make surgery safer." As part of the Universal Protocol, the attending surgeon is required to be present www.spinejournal.com
605
HEALTH SERVICES RESEARCH
TABLE!.
"July Effect" in Elective Spine Surgery • Bohl et al
tes of Resident Involvement in Resident Involvement Rate* (%)
Overall
ORt
P
25.3
> 0 Q) CO >
Procedure type
