Demystifying the “July Effect” in Plastic Surgery: A Multi-Institutional Study

Demystifying the “July Effect” in Plastic Surgery: A Multi-Institutional Study Abstract Background The “July Effect” refers to a theoretical increase in complications that may occur with the influx of inexperienced interns and residents at the beginning of each academic year in July. Objectives We endeavored to determine if a July Effect occurs in plastic surgery. Methods Plastic surgery procedures were isolated from the National Surgical Quality Improvement Program registry. Cases involving residents were grouped as either having occurred within the first academic quarter (AQ1) or remaining year (AQ2-4). Groups were propensity matched using patient/operative factors and procedure type to account for baseline differences. Univariate and multivariate regression analyses assessed differences in overall complications, surgical and medical complications, individual complications, length of hospital stay, and operative time. A comparison group comprised of procedures without resident involvement was also analyzed. Results There were 5967 cases with resident involvement, 5156 of which successfully matched. Both univariate and multivariate regression analyses revealed no significant differences between AQ1 and AQ2-4 in terms of overall, surgical, medical and individual complications, or length of hospital stay. There was a statistically significant, albeit not clinically significant, increase in operative time by 10 minutes per procedure during AQ1 in comparison to AQ2-4 (P = 0.001). For procedures lacking resident participation, there were no differences between AQ1 and AQ2-4 in terms of these outcomes. Conclusions A July Effect was not observed for plastic surgery procedures in our study, conceivably due to enhanced resident oversight and infrastructural safeguards. Patients electing to undergo plastic surgery early in the academic year can be reassured of their safety during this period. Level of Evidence: 2 In the United States, the new academic year begins every July and brings anticipation and a spirit of renewal as newly minted doctors and current residents and fellows are promoted to the next postgraduate level. At the peak of trauma season, house staff may find themselves on unfamiliar services, perhaps at different hospitals and universally in new roles with new responsibilities. Interns learn under rapid fire at the first call for lists of patients. Exemplary junior residents gain senior resident status, not only with increased operative autonomy, but also with the added responsibility for mentorship and oversight of their juniors and coordination of care with the rest of the healthcare team. Meanwhile, attending surgeons are developing familiarity with the new resident class, attempting to gauge their base knowledge and skills. Experienced academic surgeons, nurses, and ancillary staff recognize the challenges that the yearly July ritual brings. In recent years, popular media, as well as the scientific literature, have called into question the hypothetical increase in medical morbidity and mortality that purportedly occurs during this turnover of healthcare staff, which has come to be known as the “July Effect” or the “July Phenomenon.”1,2 One widely referenced study reported a 10% spike in fatal medication errors during the month of July.3 Since this 2010 study, investigations across a variety of specialties have attempted to examine this phenomenon, resulting in a controversial mix of reports in favor or against existence of a July Effect.4-22 Explanatory mechanisms for the July Effect have yet to be demonstrated; however, an assortment of etiologies have been posited. Most attention is placed on inexperienced interns and residents who could adversely affect the patient during preoperative evaluation, intraoperatively, and in postoperative care and discharge. In surgery, there is a learning curve for developing technical expertise, which may affect a surgeon’s complication rate.16-22 Each progressive year, residents and new attending physicians take on increased levels of responsibility, opening up the possibility for added morbidity and mortality. Notably, this is not a transition unique to residents. Other health professionals including nurses, operating room staff, respiratory therapists, and pharmacists may also be starting their career or beginning at a new institution with foreign protocols during this transition period.14,23 Although several surgical specialties, including cardiac, orthopaedic, and neurosurgery have examined the so-called July Effect, there are currently no publications detailing this phenomenon in plastic surgery.9,23-33 As a field defined by technically diverse and complex skills in the operating room and reliant upon nuanced flap assessment and wound care on the floor, plastic surgery may be especially vulnerable to the influx of new staff who have yet to reach mastery of these skills. Further, detecting a spike in complications during the early academic year could have considerable implications for patients undergoing elective procedures, who may choose to avoid operations during this time of year. In the present investigation, the authors sought to determine what extent the first academic quarter impacts outcomes in plastic surgery by providing a statistically powerful analysis using the National Surgical Quality Improvement Program clinical registry. METHODS Institutional Review Board The Northwestern University Institutional Review Board (IRB) has declared this retrospective investigation of de-identified patient data exempt from IRB review. Data Acquisition and Patient Selection Clinical data from the 2005 to 2013 American College of Surgeon’s National Surgical Quality Improvement Program (ACS-NSQIP) patient registry were accessed in July 2015 (J.B. and M.V.). The NSQIP registry has previously been described in detail.34,35 Briefly, variables collected include patient demographics, comorbidities, operative details, and 30-day complications. The registry comprises data which were independently abstracted prospectively by trained surgical reviewers and subjected to random auditing. These measures have ensured NSQIP data to be high quality and well standardized, as supported by an overall inter-rater disagreement rate of less than 1.6%.36 Initially, all operations performed by plastic surgeons in the data set were isolated. Only procedures which accounted for >1% of plastic surgery cases were included so that procedure type could be controlled for. Procedures were selected by primary Current Procedural Terminology (CPT) codes, which are displayed in Table 1. Next, because the July Effect is most often attributed to resident involvement, cases were divided into 2 cohorts—those which involved residents and those which did not. Patients for whom intraoperative resident involvement was unknown were excluded. A separate analysis of cases which did not involve residents enabled a comparative group. Patients were stratified by academic quarter in which the operation occurred. One group in each cohort assessed for the July Effect and included operations performed in the third quarter of the year (July-September), corresponding to the first academic quarter (AQ1). This is consistent with previous literature which has used AQ1 to approximate the time frame during which the July Effect is thought to occur.5,8,13,27,28 Prior studies have used the term “July Effect” to describe varying periods from one month to three months beginning in July when the graduate medical education cycle resets.5,8,9,13,23-33 Operations performed outside the window for the July Effect were those which occurred during the remaining three academic quarters (October-June, AQ2-4). These three academic quarters were grouped and analyzed together in order to achieve the statistical power and granularity necessary to perform a tightly controlled and informative propensity match against our primary index group, AQ1. Table 1. Procedure Listing by Current Procedural Terminology (CPT) Code Procedure  CPT code(s)  Breast reduction  19318  Prosthetic reconstruction  19357  Autologous reconstruction  19361, 19364, 19367, 19368, 19369  Mastopexy  19316  Augmentation  19324, 19325  Abdominoplasty/panniculectomy  15830, 15847  Delayed/immediate insertion of prosthesis  19340, 19342  Implant removal  19328, 19330  Capsulotomy/capsulectomy  19370, 19371  Breast Revision  19380  Nipple-areola reconstruction  19350  Free flap  15756, 15757, 15758, 15842, 20969, 20970, 20972, 20973, 43496, 49906  VHR/components separation  15734, 49560, 49561, 49565, 49566, 49568  Ganglion cyst removal  25111  Procedure  CPT code(s)  Breast reduction  19318  Prosthetic reconstruction  19357  Autologous reconstruction  19361, 19364, 19367, 19368, 19369  Mastopexy  19316  Augmentation  19324, 19325  Abdominoplasty/panniculectomy  15830, 15847  Delayed/immediate insertion of prosthesis  19340, 19342  Implant removal  19328, 19330  Capsulotomy/capsulectomy  19370, 19371  Breast Revision  19380  Nipple-areola reconstruction  19350  Free flap  15756, 15757, 15758, 15842, 20969, 20970, 20972, 20973, 43496, 49906  VHR/components separation  15734, 49560, 49561, 49565, 49566, 49568  Ganglion cyst removal  25111  VHR, ventral hernia repair View Large Risk Adjustment Variables and Outcomes of Interest Patient demographic and operative characteristics were used to compare baseline differences between groups and in propensity score matching as covariates. These variables included age, body mass index (BMI), sex, a modified Charlson comorbidity index (CCI), American Society of Anesthesiologists’ (ASA) status, surgical wound classification, smoking status, diabetes, dyspnea, hypertension, functional status (dependent vs independent), bleeding disorder, setting (inpatient vs outpatient), recent chemotherapy or radiotherapy (chemo or XRT), total relative value units (RVU) for the procedure, highest postgraduate year resident involved, and primary CPT (Table 1). NSQIP-specific definitions for these variables can be found in the user manual.37 While primary CPT allowed for significant control over procedure type, a sum of total RVUs allowed for a more robust and standardized measure of procedure intensity/complexity by accounting for concurrent procedures and their added risks. The modified CCI, which caters to the comorbidities made available through databases like NSQIP, has been demonstrated to perform similarly to the original CCI and has been employed in earlier NSQIP analyses.5,27,38-40 The modified CCI variable is not captured by NSQIP directly, and was computed as previously described. Briefly, each of the following comorbidities was assigned a point value: 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.5,27,39,40 Outcomes of interest included mortality, overall incidence of complication, surgical complications, medical complications, total operative time, and length of hospital stay. Surgical complications included surgical-site infection (SSI; superficial, deep, organ/space), wound dehiscence, graft/prosthesis/flap failure, and unplanned return to the operating room. Medical complications were coma >24 hours, cerebrovascular accident (CVA)/stroke, ventilator dependence >48 hours, unplanned intubation, progressive or acute renal failure, urinary tract infection (UTI), peripheral nerve injury, pneumonia, cardiac arrest, myocardial infarction (MI), bleeding requiring transfusion, sepsis or septic shock, and venous thromboembolism (VTE; DVT, deep venous thrombosis or PE, pulmonary embolism). Any complication included both surgical and medical complications. Readmissions were not included in this study due to insufficient cases available for which both readmission status and resident involvement status were reported. Propensity Score Matching Propensity score matching using the aforementioned patient/operative characteristics and procedure types as covariates was used to eliminate statistically significant differences in baseline characteristics between operations performed in AQ1 vs the remaining year (AQ2-4). Specifically, nearest-neighbor matching without replacement in a 3:1 ratio was used, as previously described.41-43 Essentially, propensity score matching ensured that any potentially confounding differences between the patient demographics, operative factors, or procedural makeup in AQ1 vs AQ2-4 were reduced in order to more precisely determine the extent to which AQ1 influenced complications. Subgroup Analyses In the first subanalysis, three separate subgroups were isolated from the matched cohort to explore whether an effect exists for a specific population. Univariate and multivariate analyses were conducted for: (1) cases involving interns (defined as a postgraduate year 1 resident being the most senior resident present); (2) those involving senior-level residents (defined as a postgraduate year 4-6 resident being the most senior resident present); and (3) those performed in the inpatient setting only, as these patients often receive postoperative care by residents. Pairwise comparisons between AQ1 and each individual quarter were also performed. AQ2, AQ3, and AQ4 were each matched separately 1:1 with AQ1 using the original unmatched cohort involving residents. Multivariate analyses were performed for incidence of any complication and total operative time. Conceivably, the difference between AQ1 and AQ4 would be expected to be the most dramatic, as residents fine tune both their operative skills and floor management as the year progresses. Finally, to investigate whether differences between AQ1 and AQ2-4 become more pronounced as procedural complexity increases, we performed a sensitivity analysis based on total case RVU. Using matched cases including residents, multivariate analyses of overall complications and total operative time were conducted for cohorts with increasing minimum total RVUs. The analysis was done at increments of 10 RVU, ranging from >10 RVU to >60 RVU. Statistical Analysis Univariate analysis (Fischer’s exact test or χ2 for categorical variables and independent t test for continuous variables) was used to assess for differences in baseline patient/operative characteristics and postoperative outcomes between groups. To adjust risk and determine the independent effect of AQ1, we utilized binary logistic regression for 30-day complications and linear regression for total operative time and length of hospital stay. Multivariate analyses were generated for matched cohorts, and limited to complications of >30 events. In all analyses, significance was considered as two-tailed at the α = 0.05 level. All statistics were performed using SPSS version 22 (IBM Corp., Armonk, NY) and R version 2.15.3 with the PS Matching package (R Fdn. for Statistical Computing, Vienna, Austria).44 RESULTS Characteristics of Cohort With Resident Involvement There were 9728 plastic surgery procedures in the data set with intraoperative resident involvement. After eliminating primary CPT codes comprising <1% of the cohort, 5967 remained, 1493 (25.0%) from AQ1 and 4474 (75.0%) from AQ2-4. Unmatched patients did not differ significantly in terms of baseline patient/operative characteristics by academic quarter; however, they did differ by number of breast reductions (27.9% AQ1 vs 24.0% AQ2-4; P = 0.003) and capsulotomy/capsulectomies (4.8% vs 6.7%; P = 0.008) performed (Supplementary Table 1). Propensity score matching successfully eliminated these differences and maintained similarity in all other patient/operative factors between AQ1 and AQ2-4 (Table 2). Mean patient age was 48.31 years (range, 16-89 years), and 5.1% were men. After matching, there were 5156 cases, 1443 (28.0%) from AQ1 and 3713 (72.0%) from AQ2-4. Table 2. Patient/Operative Characteristics by Academic Quarter for Matched Cohorts   Without residents (n = 6118)    With residents (n = 5156)    Characteristic  AQ1  AQ 2-4  P value  AQ1  AQ 2-4  P value  No. of patients  1616 (26.4%)  4502 (73.6%)  —  1443 (28.0%)  3713 (72.0%)  —  Demographic  Age (years)  47.00 (13.44)  46.66 (13.37)  0.389  48.34 (13.29)  48.30 (13.12)  0.922  BMI (kg/m2)  28.38 (6.86)  28.09 (6.79)  0.148  28.82 (6.81)  28.65 (6.84)  0.420  Male  3.0  2.6  0.323  5.2  5.1  0.833  CCI      0.558      0.929   0  54.7  56.2  —  48.7  49.1  —   1  23.3  23.2  —  25.7  26.3  —   2  13.1  12.8  —  15.2  14.1  —   3  5.6  5.3  —  6.7  6.7  —   4  2.7  2.2  —  2.4  2.6  —   5 or greater  0.7  0.4  —  1.3  1.2  —  ASA status      0.605      0.666   1  19.2  19.9  —  16.2  16.5  —   2  66.0  66.7  —  61.2  61.2  —   3 or greater  14.7  13.4  —  22.6  22.3  —  Wound classification      0.936      0.940   II -Clean-contaminated  3.6  3.4  —  3.9  3.8  —   III-Contaminated  1.1  1.1  —  1.7  1.7  —   IV-Dirty  1.5  1.3  —  2.3  2.0  —  Smoking  14.1  13.6  0.613  12.5  12.1  0.705  Diabetes  5.4  4.8  0.349  6.4  6.4  1.000  Dyspnea  2.2  1.9  0.533  3.0  2.6  0.389  Hypertension  23.1  21.7  0.235  24.3  23.6  0.636  Dependent  0.5  0.3  0.350  1.2  1.2  0.775  Bleeding disorder  0.6  0.7  0.861  1.3  1.2  0.780  Operative factor  Inpatient  19.4  17.3  0.063  37.2  36.9  0.847  Chemo or XRT  2.5  2.2  0.495  3.1  3.3  0.793  Total RVUs  19.31 (14.85)  19.33 (14.31)  0.955  27.80 (19.74)  28.15 (20.16)  0.576  Highest PGY  —  —  —  4.68 (2.17)  4.69 (2.20)  0.873  Procedure type  Breast reduction  27.6  28.3  0.628  28.3  26.1  0.115  Prosthetic reconstruction  5.6  5.6  0.950  7.8  7.9  0.954  Autologous reconstruction  4.6  4.5  0.835  10.7  11.2  0.692  Mastopexy  7.1  6.3  0.291  3.7  3.4  0.554  Augmentation  9.8  10.7  0.368  4.4  4.5  0.881  Abdominoplasty/panniculectomy  12.3  12.3  0.965  9.6  10.2  0.605  Delayed/immediate insertion of prosthesis  5.9  5.6  0.708  8.5  9.1  0.515  Implant removal  2.7  2.7  0.928  2.3  1.8  0.212  Capsulotomy/capsulectomy  7.3  7.2  0.866  5.0  5.5  0.535  Breast revision  7.7  8.0  —  8.0  8.3  0.821  Nipple-areola reconstruction  4.8  5.3  0.513  3.7  3.8  0.871  Free flap  0.5  0.4  0.830  2.2  2.5  0.542  VHR/components separation  2.2  1.8  0.287  4.5  4.4  0.940  Ganglion cyst removal  1.8  1.3  0.143  1.2  1.3  0.783    Without residents (n = 6118)    With residents (n = 5156)    Characteristic  AQ1  AQ 2-4  P value  AQ1  AQ 2-4  P value  No. of patients  1616 (26.4%)  4502 (73.6%)  —  1443 (28.0%)  3713 (72.0%)  —  Demographic  Age (years)  47.00 (13.44)  46.66 (13.37)  0.389  48.34 (13.29)  48.30 (13.12)  0.922  BMI (kg/m2)  28.38 (6.86)  28.09 (6.79)  0.148  28.82 (6.81)  28.65 (6.84)  0.420  Male  3.0  2.6  0.323  5.2  5.1  0.833  CCI      0.558      0.929   0  54.7  56.2  —  48.7  49.1  —   1  23.3  23.2  —  25.7  26.3  —   2  13.1  12.8  —  15.2  14.1  —   3  5.6  5.3  —  6.7  6.7  —   4  2.7  2.2  —  2.4  2.6  —   5 or greater  0.7  0.4  —  1.3  1.2  —  ASA status      0.605      0.666   1  19.2  19.9  —  16.2  16.5  —   2  66.0  66.7  —  61.2  61.2  —   3 or greater  14.7  13.4  —  22.6  22.3  —  Wound classification      0.936      0.940   II -Clean-contaminated  3.6  3.4  —  3.9  3.8  —   III-Contaminated  1.1  1.1  —  1.7  1.7  —   IV-Dirty  1.5  1.3  —  2.3  2.0  —  Smoking  14.1  13.6  0.613  12.5  12.1  0.705  Diabetes  5.4  4.8  0.349  6.4  6.4  1.000  Dyspnea  2.2  1.9  0.533  3.0  2.6  0.389  Hypertension  23.1  21.7  0.235  24.3  23.6  0.636  Dependent  0.5  0.3  0.350  1.2  1.2  0.775  Bleeding disorder  0.6  0.7  0.861  1.3  1.2  0.780  Operative factor  Inpatient  19.4  17.3  0.063  37.2  36.9  0.847  Chemo or XRT  2.5  2.2  0.495  3.1  3.3  0.793  Total RVUs  19.31 (14.85)  19.33 (14.31)  0.955  27.80 (19.74)  28.15 (20.16)  0.576  Highest PGY  —  —  —  4.68 (2.17)  4.69 (2.20)  0.873  Procedure type  Breast reduction  27.6  28.3  0.628  28.3  26.1  0.115  Prosthetic reconstruction  5.6  5.6  0.950  7.8  7.9  0.954  Autologous reconstruction  4.6  4.5  0.835  10.7  11.2  0.692  Mastopexy  7.1  6.3  0.291  3.7  3.4  0.554  Augmentation  9.8  10.7  0.368  4.4  4.5  0.881  Abdominoplasty/panniculectomy  12.3  12.3  0.965  9.6  10.2  0.605  Delayed/immediate insertion of prosthesis  5.9  5.6  0.708  8.5  9.1  0.515  Implant removal  2.7  2.7  0.928  2.3  1.8  0.212  Capsulotomy/capsulectomy  7.3  7.2  0.866  5.0  5.5  0.535  Breast revision  7.7  8.0  —  8.0  8.3  0.821  Nipple-areola reconstruction  4.8  5.3  0.513  3.7  3.8  0.871  Free flap  0.5  0.4  0.830  2.2  2.5  0.542  VHR/components separation  2.2  1.8  0.287  4.5  4.4  0.940  Ganglion cyst removal  1.8  1.3  0.143  1.2  1.3  0.783  *denotes statistical significance at P < 0.05. Categorical variables are reported as percentages. Continuous variables are reported as their means (standard deviation). View Large There were no differences in any unadjusted outcomes after matching (Table 3; Supplementary Table 2 for individual complications). Postmatch rate of any complication in AQ1 was 12.4% vs 11.8% in AQ2-4 (P = 0.566), mortality (0.1% vs 0.0%; P = 0.191), surgical complication (9.5% vs 9.0%; P = 0.591), medical complication (4.6% vs 4.2%; P = 0.446), total operative time (190.56 ± 155.57 vs 182.21 ± 150.31 minutes; P = 0.076), and length of hospital stay (2.03 ± 6.01 vs 2.08 ± 8.99 days; P = 0.830). Table 3. Unadjusted Complication Rates by Academic Quarter for Matched Cohorts Complication  Without residents    With residents    AQ 2-4  AQ1    AQ 2-4  AQ1    Count  %  Count  %  P value  Count  %  Count  %  P value  Any complication  247  5.5%  85  5.3%  0.798  438  11.8%  179  12.4%  0.566  Surgical complication  198  4.4%  70  4.3%  0.944  335  9.0%  137  9.5%  0.591  Medical complication  71  1.6%  23  1.4%  0.725  155  4.2%  67  4.6%  0.446  Return to OR  93  2.1%  33  2.0%  1.000  203  5.5%  63  4.4%  0.123  Death  4  0.1%  1  0.1%  1.000  1  0.0%  2  0.1%  0.191  Total operative time (minutes)  148.96  (89.66)  153.10  (101.94)  0.149  182.21  (150.31)  190.56  (155.57)  0.076  Length of hospital stay (days)  0.89  (3.68)  0.97  (3.87)  0.488  2.08  (8.99)  2.03  (6.01)  0.830  Complication  Without residents    With residents    AQ 2-4  AQ1    AQ 2-4  AQ1    Count  %  Count  %  P value  Count  %  Count  %  P value  Any complication  247  5.5%  85  5.3%  0.798  438  11.8%  179  12.4%  0.566  Surgical complication  198  4.4%  70  4.3%  0.944  335  9.0%  137  9.5%  0.591  Medical complication  71  1.6%  23  1.4%  0.725  155  4.2%  67  4.6%  0.446  Return to OR  93  2.1%  33  2.0%  1.000  203  5.5%  63  4.4%  0.123  Death  4  0.1%  1  0.1%  1.000  1  0.0%  2  0.1%  0.191  Total operative time (minutes)  148.96  (89.66)  153.10  (101.94)  0.149  182.21  (150.31)  190.56  (155.57)  0.076  Length of hospital stay (days)  0.89  (3.68)  0.97  (3.87)  0.488  2.08  (8.99)  2.03  (6.01)  0.830  *denotes statistical significance at P < 0.05. Continuous variables are reported as their means (standard deviation). View Large Subanalyses by postgraduate year were performed on 624 cases involving interns (PGY1) and 2791 cases involving senior residents (PGY4-6). Neither univariate nor multivariate analyses revealed any significant differences between AQ1 and AQ2-4 with respect to surgical or medical complications, total operative time, or length of hospital stay. In 1907 cases with intraoperative resident involvement performed in the inpatient setting, there were no differences between AQ1 and AQ2-4 in terms of surgical or medical complications or length of hospital stay (Table 4). Total operative time in this subcohort was longer in AQ1 by multivariate (marginal effect, +16.897 minutes; P = 0.013), but not univariate analysis (291.37 ± 193.74 vs 282.30 ± 187.37 minutes; P = 0.346). Table 4. Subgroup Analyses for Inpatient Surgeries and Senior Resident Involvement Subgroup  Outcome  AQ1  AQ2-4  Univariate P value  Adjusted OR  Marginal Effect  95% CI  Multivariate P-value  Inpatient surgeries only  Any complication  22.2%  21.9%  0.902  1.056  —  0.815-1.369  0.681  Surgical complication  14.9%  15.3%  0.887  0.991  —  0.742-1.324  0.952  Medical complication  11.0%  10.1%  0.617  1.122  —  0.783-1.608  0.529  Total operative time (minutes)  291.37 (193.74)  282.30 (187.37)  0.346  —  16.897  3.538-30.256  0.013*  Length of hospital stay (days)  4.71 (7.87)  5.14 (14.27)  0.507  —  -0.280  -1.472-0.911  0.645  Senior residents (PGY 4-6) only  Any complication  12.2%  11.7%  0.744  1.090  —  0.828-1.435  0.540  Surgical complication  9.5%  9.0%  0.662  1.099  —  0.818-1.478  0.530  Medical complication  4.1%  4.6%  0.683  0.904  —  0.564-1.447  0.673  Total operative time (minutes)  185.44 (145.44)  181.43 (149.80)  0.522  —  6.607  -1.323-14.536  0.102  Length of hospital stay (days)  2.07 (6.16)  2.33 (11.66)  0.559  —  -0.282  -1.068-0.504  0.482  Subgroup  Outcome  AQ1  AQ2-4  Univariate P value  Adjusted OR  Marginal Effect  95% CI  Multivariate P-value  Inpatient surgeries only  Any complication  22.2%  21.9%  0.902  1.056  —  0.815-1.369  0.681  Surgical complication  14.9%  15.3%  0.887  0.991  —  0.742-1.324  0.952  Medical complication  11.0%  10.1%  0.617  1.122  —  0.783-1.608  0.529  Total operative time (minutes)  291.37 (193.74)  282.30 (187.37)  0.346  —  16.897  3.538-30.256  0.013*  Length of hospital stay (days)  4.71 (7.87)  5.14 (14.27)  0.507  —  -0.280  -1.472-0.911  0.645  Senior residents (PGY 4-6) only  Any complication  12.2%  11.7%  0.744  1.090  —  0.828-1.435  0.540  Surgical complication  9.5%  9.0%  0.662  1.099  —  0.818-1.478  0.530  Medical complication  4.1%  4.6%  0.683  0.904  —  0.564-1.447  0.673  Total operative time (minutes)  185.44 (145.44)  181.43 (149.80)  0.522  —  6.607  -1.323-14.536  0.102  Length of hospital stay (days)  2.07 (6.16)  2.33 (11.66)  0.559  —  -0.282  -1.068-0.504  0.482  *denotes statistical significance at P < 0.05. Data are from the matched cohort involving residents. Odds ratios and marginal effects indicate the independent influence of AQ1. Continuous variables are reported by their means (standard deviation). View Large Characteristics of Cohort Without Resident Involvement Initially, 10,623 plastic surgery cases without resident surgical involvement were identified. Approximately 70% (7453) of these met inclusion criteria by their primary CPT codes, 1675 (22.5%) from AQ1 and 5778 (77.5%) from AQ2-4. Supplementary Table 1 details the make up of this cohort by procedure type. Unmatched patients differed statistically in terms of several baseline characteristics: BMI, CCI, diabetes, inpatient procedures, total RVUs, breast revisions, and ventral hernia repair/components separation procedures (Supplementary Table 1). After propensity score matching, 6118 patients remained, 1616 (26.4%) from AQ1 and 4502 (73.6%) from AQ2-4. Mean age was 46.75 years (range, 16-89 years), and 2.7% were men. Matching successfully eliminated all statistically significant differences in patient demographics, comorbidities, operative factors and number of each procedure performed (Table 2). There were no statistically significant differences in unadjusted postoperative outcomes for matched cases performed without resident involvement (Table 3; Supplementary Table 2 for individual complications). The overall rate for complications in AQ1 was 5.3% vs 5.5% in AQ2-4 (P = 0.798), mortality (0.1% vs 0.1%; P = 1.000), surgical complication (4.3% vs 4.4%; P = 0.944), medical complication (1.4% vs 1.6%; P = 0.725), total operative time (153.10 ± 101.94 vs 148.96 ± 89.66 minutes; P = 0.149), and length of hospital stay (0.97 ± 3.87 vs 0.89 ± 3.68 days; P = 0.488). Multivariate Regression Analysis for AQ1 as a Risk Factor Table 5 displays the multivariate results for matched cohorts. With residents involved, it was found that operations occurring in AQ1 were statistically significantly longer, albeit not clinically so (marginal effect, +9.937 minutes; P = 0.001). All other outcomes for cases involving residents were insignificant for AQ1 as a risk factor, including any complication (odds ratio [OR], 1.073; 95% confidence interval [95%CI] 0.879-1.311; P = 0.488), surgical complication (OR 1.068; 95%CI 0.861-1.326; P = 0.550), medical complication (OR 1.140; 95%CI 0.821-1.584; P = 0.434), and length of hospital stay (marginal effect, −0.074 days; P = 0.749). Table 5. Multivariate Adjustment for Complications by Academic Quarter for Matched Cohorts Complication  Without residents    With residents    Adjusted OR  95% CI  P value  Adjusted OR  95% CI  P value  Any complication  1.110  0.851-1.448  0.440  1.073  0.879-1.311  0.488  Surgical complication  1.040  0.781-1.384  0.791  1.068  0.861-1.326  0.550  Medical complication  1.278  0.761-2.146  0.354  1.140  0.821-1.584  0.434  Return to OR  1.066  0.706-1.609  0.762  0.790  0.586-1.066  0.123  UTI  1.609  0.522-4.959  0.408  0.948  0.401-2.241  0.904  Pneumonia  2.195  0.227-21.251  0.497  1.002  0.185-5.436  0.999  Blood transfusion  1.159  0.485-2.771  0.740  1.279  0.841-1.947  0.250  Any sepsis  0.849  0.282-2.555  0.770  0.950  0.439-2.057  0.896   Sepsis  0.856  0.284-2.578  0.782  0.875  0.390-1.959  0.745  VTE  3.172  0.569-16.882  0.176  0.935  0.377-2.317  0.884   DVT  4.062  0.353-46.734  0.261  0.491  0.131-1.841  0.291  Any SSI  1.067  0.728-1.563  0.740  1.257  0.946-1.670  0.114   Superficial SSI  1.037  0.661-1.627  0.875  1.210  0.858-1.704  0.277   Deep SSI  1.193  0.513-2.775  0.682  1.227  0.665-2.263  0.513   Organ/space SSI  0.659  0.191-2.271  0.509  1.376  0.649-2.921  0.405  Dehiscence  1.838  0.806-4.188  0.148  1.737  0.912-3.308  0.093  Graft/prosthesis/flap failure  0.850  0.152-4.749  0.853  0.930  0.481-1.800  0.830  AQ1 Marginal effect  Marginal effect  95% CI  P value  Marginal effect  95% CI  P value  Total operative time (minutes)  3.432  −0.937-7.801  0.124  9.937  3.950-15.923  0.001*  Length of hospital stay (days)  −0.026  −0.202-0.149  0.767  −0.074  −0.527-0.379  0.749  Complication  Without residents    With residents    Adjusted OR  95% CI  P value  Adjusted OR  95% CI  P value  Any complication  1.110  0.851-1.448  0.440  1.073  0.879-1.311  0.488  Surgical complication  1.040  0.781-1.384  0.791  1.068  0.861-1.326  0.550  Medical complication  1.278  0.761-2.146  0.354  1.140  0.821-1.584  0.434  Return to OR  1.066  0.706-1.609  0.762  0.790  0.586-1.066  0.123  UTI  1.609  0.522-4.959  0.408  0.948  0.401-2.241  0.904  Pneumonia  2.195  0.227-21.251  0.497  1.002  0.185-5.436  0.999  Blood transfusion  1.159  0.485-2.771  0.740  1.279  0.841-1.947  0.250  Any sepsis  0.849  0.282-2.555  0.770  0.950  0.439-2.057  0.896   Sepsis  0.856  0.284-2.578  0.782  0.875  0.390-1.959  0.745  VTE  3.172  0.569-16.882  0.176  0.935  0.377-2.317  0.884   DVT  4.062  0.353-46.734  0.261  0.491  0.131-1.841  0.291  Any SSI  1.067  0.728-1.563  0.740  1.257  0.946-1.670  0.114   Superficial SSI  1.037  0.661-1.627  0.875  1.210  0.858-1.704  0.277   Deep SSI  1.193  0.513-2.775  0.682  1.227  0.665-2.263  0.513   Organ/space SSI  0.659  0.191-2.271  0.509  1.376  0.649-2.921  0.405  Dehiscence  1.838  0.806-4.188  0.148  1.737  0.912-3.308  0.093  Graft/prosthesis/flap failure  0.850  0.152-4.749  0.853  0.930  0.481-1.800  0.830  AQ1 Marginal effect  Marginal effect  95% CI  P value  Marginal effect  95% CI  P value  Total operative time (minutes)  3.432  −0.937-7.801  0.124  9.937  3.950-15.923  0.001*  Length of hospital stay (days)  −0.026  −0.202-0.149  0.767  −0.074  −0.527-0.379  0.749  *denotes statistical significance at P < 0.05. Odds ratios and marginal effects indicate the independent influence of AQ1. View Large In the comparison cohort lacking resident involvement, AQ1 did not significantly increase the risk for any outcome, including any complication (OR 1.110; 95%CI 0.851-1.448; P = 0.440), surgical complication (OR 1.040; 95%CI 0.781-1.384; P = 0.791), medical complication (OR 1.278; 95%CI 0.761-2.146; P = 0.354), total operative time (marginal effect, +3.432 minutes; P = 0.124), or length of hospital stay (marginal effect, −0.026 days; P = 0.767). Multivariate Comparison of Complications by Individual Academic Quarter Successful one-to-one pairwise matching by individual quarter yielded 1337 cases each for AQ1 and AQ2 in the AQ2 comparison, 1154 each for the AQ3 comparison, and 1226 each for the AQ4 comparison. On multivariate analysis, AQ1 did not differ in terms of risk for experiencing any complication overall, surgical complications, or medical complications (Table 6). Length of hospital stay was statistically significantly prolonged in AQ1 when compared to AQ4 (marginal effect, +0.429 days; P = 0.018), but not AQ2 (marginal effect, +0.042 days; P = 0.802) or AQ3 (marginal effect, +0.113 days; P = 0.621). AQ1 was also associated with increased total operative time in comparison to AQ2 (marginal effect, +9.882 minutes; P = 0.007) and AQ4 (marginal effect, +11.723 minutes; P = 0.004), but not AQ3 (marginal effect, +7.433 minutes; P = 0.070). Table 6. Multivariate Comparison of Complications by Individual Academic Quarter With Residents Academic quarter  Total cases  Any complication  Surgical complication  Medical complication  Total operative time (minutes)  Length of hospitalization (days)  Incidence  Adjusted OR  P value  Incidence  Adjusted OR  P value  Incidence  Adjusted OR  P value  Mean (SD)  Marginal effect  P value  Mean (SD)  Marginal effect  P value  AQ1  1337  12.6%  0.984  0.901  9.6%  1.022  0.875  4.8%  1.047  0.825  190.28 (154.87)  9.882  0.007*  2.05 (5.92)  0.042  0.802  AQ2  1337  12.6%  —  —  9.4%  —  —  4.4%  —  —  180.63 (141.97)  —  —  1.97 (5.17)  —  —  AQ1  1154  12.2%  1.236  0.136  9.2%  1.193  0.253  4.9%  1.557  0.068  193.66 (158.83)  7.433  0.070  2.12 (6.42)  0.113  0.621  AQ3  1154  10.5%  —  —  8.1%  —  —  3.4%  —  —  186.37 (156.19)  —  —  2.03 (6.92)  —  —  AQ1  1226  13.2%  1.104  0.459  10.2%  1.108  0.476  5.3%  1.273  0.268  194.17 (162.67)  11.723  0.004*  2.43 (7.21)  0.429  0.018*  AQ4  1226  12.2%  —  —  9.2%  —  —  4.4%  —  —  181.69 (148.87)  —  —  2.02 (4.65)  —  —  Academic quarter  Total cases  Any complication  Surgical complication  Medical complication  Total operative time (minutes)  Length of hospitalization (days)  Incidence  Adjusted OR  P value  Incidence  Adjusted OR  P value  Incidence  Adjusted OR  P value  Mean (SD)  Marginal effect  P value  Mean (SD)  Marginal effect  P value  AQ1  1337  12.6%  0.984  0.901  9.6%  1.022  0.875  4.8%  1.047  0.825  190.28 (154.87)  9.882  0.007*  2.05 (5.92)  0.042  0.802  AQ2  1337  12.6%  —  —  9.4%  —  —  4.4%  —  —  180.63 (141.97)  —  —  1.97 (5.17)  —  —  AQ1  1154  12.2%  1.236  0.136  9.2%  1.193  0.253  4.9%  1.557  0.068  193.66 (158.83)  7.433  0.070  2.12 (6.42)  0.113  0.621  AQ3  1154  10.5%  —  —  8.1%  —  —  3.4%  —  —  186.37 (156.19)  —  —  2.03 (6.92)  —  —  AQ1  1226  13.2%  1.104  0.459  10.2%  1.108  0.476  5.3%  1.273  0.268  194.17 (162.67)  11.723  0.004*  2.43 (7.21)  0.429  0.018*  AQ4  1226  12.2%  —  —  9.2%  —  —  4.4%  —  —  181.69 (148.87)  —  —  2.02 (4.65)  —  —  *denotes statistical significance at P < 0.05. Data are derived from individual academic quarters (AQs) matched head-to-head with AQ1 for procedures involving residents. Odds ratios and marginal effects indicate the independent influence of AQ1. View Large Sensitivity Analysis for Total Complications and Operative Time Stratified by Surgical Complexity Of the 5156 matched cases involving residents, 4717 had greater than 10 total RVU, 2707 greater than 20 total RVU, 1987 greater than 30 total RVU, 1011 greater than 40 total RVU, 577 greater than 50 total RVU, and 388 greater than 60 total RVU. At each RVU interval, there was no difference between AQ1 and AQ2-4 in the overall complication rate (Table 7). Total operative time was significantly longer during AQ1 for all RVU intervals, and the difference increased as case complexity increased: all procedures (+9.937 minutes; P = 0.001), >10 RVUs (+10.821 minutes; P = 0.001), >20 RVUs (+12.941 minutes; 0.010), >30 RVUs (+18.708 minutes; P = 0.003), >40 RVUs (+35.947 minutes; P = 0.001), >50 RVUs (+34.316 minutes; P = 0.030), and >60 RVUs (+48.486 minutes; P = 0.014). Table 7. Sensitivity Analysis for Total Complications and Operative Time Stratified by Surgical Complexity     Any complication    Total operative time (minutes)    Total RVUs  Total cases  AQ1 adjusted OR  95% CI  P value  AQ1 marginal effect  95% CI  P value  All  5156  1.073  0.879-1.311  0.488  9.937  3.950-15.923  0.001*  >10  4717  1.025  0.835-1.258  0.816  10.821  4.401-17.242  0.001*  >20  2707  1.161  0.907-1.488  0.236  12.941  3.105-22.776  0.010*  >30  1987  1.053  0.790-1.405  0.723  18.708  6.387-31.030  0.003*  >40  1011  1.191  0.832-1.703  0.339  35.947  15.036-56.858  0.001*  >50  577  1.129  0.704-1.810  0.614  34.316  3.371-65.261  0.030*  >60  388  1.183  0.677-2.067  0.555  48.486  9.750-87.222  0.014*      Any complication    Total operative time (minutes)    Total RVUs  Total cases  AQ1 adjusted OR  95% CI  P value  AQ1 marginal effect  95% CI  P value  All  5156  1.073  0.879-1.311  0.488  9.937  3.950-15.923  0.001*  >10  4717  1.025  0.835-1.258  0.816  10.821  4.401-17.242  0.001*  >20  2707  1.161  0.907-1.488  0.236  12.941  3.105-22.776  0.010*  >30  1987  1.053  0.790-1.405  0.723  18.708  6.387-31.030  0.003*  >40  1011  1.191  0.832-1.703  0.339  35.947  15.036-56.858  0.001*  >50  577  1.129  0.704-1.810  0.614  34.316  3.371-65.261  0.030*  >60  388  1.183  0.677-2.067  0.555  48.486  9.750-87.222  0.014*  *denotes statistical significance at P < 0.05. Data are from the matched cohort involving residents. Odds ratios and marginal effects indicate the independent influence of AQ1. Total RVUs indicates the combined total of relative value units for the procedure. View Large DISCUSSION As part of the ongoing scrutiny of resident training and patient safety, the popular news, social media and the scientific literature have continued to consider the compelling idea that the annual turnover of new residents and other healthcare workers amid the summer months results in a theoretical increase in morbidity and mortality.1,2 Findings of such a “July phenomenon” would lay the groundwork for implementing added precautionary measures and oversight aimed to support transitioning house staff. In 2003, the Accreditation Council of Graduate Medical Education (ACGME) issued a mandate to restrict resident work hours in an effort to bolster patient safety while new and inexperienced healthcare workers are trained.45 Furthermore in 2011, the ACGME introduced new supervision standards, which included definitions for “direct supervision,” “indirect supervision,” and “oversight.”46 Nonetheless, there is no substitute for experience, and thus, by nature there remain unavoidable factors in the teaching hospital setting, which may represent added risk to the patient. Prior to the present investigation, this “July Effect” has never been examined in the field of plastic surgery. It is possible that plastic surgery may be especially vulnerable due to the meticulous surgical techniques and postoperative assessment it entails. Further, the elective nature of the specialty could have implications on patient scheduling during the academic year. Using robust statistical techniques to match procedures by quarter of year, we have provided the first analysis in plastic surgery through query of the validated, multi-institutional NSQIP registry. By and large, we found insufficient evidence to support conjectures of such a phenomenon. Based on our findings, patients can be reassured that undergoing a procedure during the time of the so-called July Effect will not have an adverse impact on their outcomes. Propensity score matching provided control over differences in patient/operative factors and procedure composition between AQ1 and AQ2-4 groups. Groups were thus selected to be equivalent in every measurable aspect except for timing during the year. In both cases involving residents and those lacking resident involvement, there were no significant differences between AQ1 and the remainder of the year in terms of mortality, overall incidence of complications, length of hospital stay, or any individual complication captured by NSQIP (Tables 3 and 7). When parsed out by individual AQ, these findings were also largely recapitulated, even for the most theoretically dramatic comparison against AQ4. Likewise, sensitivity analysis for increasing procedural complexity demonstrated no differences, despite varying the cohort’s minimum RVU to as high as 60, suggesting that lower risk procedures were not masking or diluting a true effect (Table 7). Together, these results imply that even for the most demanding plastic surgery cases, the graduate medical education system is robust against increases in 30 day complications during the early academic year. Our data did demonstrate that in cases involving residents, surgical duration was statistically significantly longer in AQ1 by nearly ten minutes per operation on average compared to AQ2-4. While increased operative time has been demonstrated to correspond to a higher rate of morbidity including VTE and UTI,47-49 ten minutes on average is unlikely to be a clinically significant prolongation in this regard. This disparity became more pronounced as procedural complexity was increased in our sensitivity analysis; for example, operations took 35 minutes longer on average during AQ1 for procedures over 40 RVUs, and almost 50 minutes longer for procedures over 60 RVUs (Table 7). Although the increase in surgical duration for these most complex procedures no doubt represents a rise that is significant to the attending and medical team, most importantly we did not observe a parallel rise in 30-day complications. Other factors that we were unable to control for may be partially contributing to this observation, such as the turnover of ancillary staff in the summer. We believe that this increase in operative duration, which is marginal for the average procedure, is but a necessary and modest sacrifice to train the next generation of physicians, and that it is not feasible to eliminate the operating room learning curve altogether. Three separate subanalyses were performed to further focus our investigation on three theoretically vulnerable types of cases: those involving new residents (interns), those involving senior-level residents (PGY 4-6), and those performed in the inpatient setting. An analysis for intern participation was conducted because they possess the least clinical experience of all residents and have newly embarked on their residencies in July. Our next subanalysis was for cases involving senior-level residents because of the significantly greater autonomy both in and outside of the operating room that they gain upon graduation from junior ranks. The third analysis was performed exclusively for inpatient procedures because these procedures are more likely to involve residents in postoperative care in comparison to outpatient, providing additional opportunity for residents to impact patient outcomes. In cases involving either intern participation, senior-level residents, or inpatient cases, AQ1 again did not yield higher risk for complications or length of hospital stay. Procedures performed with interns present during AQ1 were longer by nearly 17 minutes on average. While this is a further extension of operative time beyond that observed when all resident levels are combined, it again represents only a diminution in efficiency rather than added risk to the patient, as evidenced by stable complication rates. The results in both of these subanalyses further boost our confidence in the findings of no July Effect. It is important to note that this study did not aim to gauge the impact of resident involvement itself on complication rate, which has previously been reported.50 Direct comparisons between resident-involved cases and those without residents in the present study are inappropriate because these 2 cohorts have been selected to be matched by AQ1 vs AQ2-4, and thus have considerable differences including procedure composition and preoperative patient health. This study stratified cohorts of patients based on whether or not a resident was present in the operating room in order to address the possibility that a mixed cohort could mask a July Effect theoretically driven by resident participation. Nonetheless, no such effect could be attributed to residents upon isolation of these cases only. Separation of cohorts also allowed us to investigate for the possibility of a July Effect due to other factors such as physician vacation time or the promotion of new attending surgeons and other health professionals including nurses, operating room staff, therapists, and pharmacists. Notably, without resident involvement, there was no difference in total operative time by academic quarter, as seen in the cohort with residents. It can reasonably be extrapolated that residents assisted more slowly and required additional guidance in the operating room during the beginning of a new academic year, which marginally diminished efficiency. From large multicenter databases like NSQIP and the National Inpatient Sample (NIS) to retrospective intra-institutional data, the literature has reported mixed findings on the presence or absence of a July Effect.4-15,23-33 Since 2010, when Philips and Barker reported a 10% peak in fatal medication errors during the month of July, numerous reports have refuted the existence of the phenomenon.3,5,6,8-11,23,24,27-33 Within the surgical literature, Bohl et al found no evidence of a July Effect in examination of total arthroplasty and spine surgery using NSQIP.27,28 Ehlert et al found no July Effect for the most common inpatient procedures in NSQIP, although the study did not propensity match, separate cases by resident involvement or report individual complications.5 A 2011 systematic review of 39 studies acknowledged the heterogeneity of the literature, but did conclude that mortality increases and efficiency drops during the academic turnover period.15 While increased mortality was not observed in the low risk plastic surgery procedures we examined, a modest prolongation in operative time could be appreciated. Furthermore, the 2011 review could not draw firm conclusions on morbidity, and rather, reported that many of the available studies were either limited or suboptimally analyzed. For instance, 41% of studies in the review did not adjust for confounders, many were not able to isolate cases in which residents were specifically involved and many did not use rigorous or appropriate statistical methodologies.15 Although the NIS dataset allows case selection by month rather than quarter, investigations utilizing the NIS are limited by the lack of surgical complications collected (hematoma only), categorization of teaching vs nonteaching hospital setting only and exclusion of outpatient procedures, and thus are not applicable to plastic surgery.6,7,23,24,29-32,51 Also noteworthy, many studies did not utilize propensity matching or use other methods to account for seasonal differences in case mix (eg, complex trauma vs elective).4,5,12,13,27,28 It is not out of question that certain fields or procedures actually do experience a July Effect. Thus, pointed specialty-specific investigations are warranted to avoid masking important individual findings within large all-encompassing studies. Our study has been conducted using an optimized statistical approach performed with rigor, including multivariate analysis subsequent to robustly controlled propensity score matching. In this light, we find it reasonable to conclude that our findings of no July Effect in plastic surgery are reliable and limited only by our data source. While feasibly due to other factors, the July Effect is traditionally thought to be attributable to the involvement of new interns or rising residents. There are a number of interpretations that could explain the absence of the phenomenon. First, the support network surrounding inexperienced workers may mitigate such deficits. Other members of the team, including senior residents, fellows, attending surgeons, and experienced nurses likely lend greater oversight during this delicate period.6 In the operating room, greater proportions of the surgery may be performed by the attending due to the technical immaturity of newer residents.6,11 On the floor, efficacious management and oversight by attendings and more senior-level residents may act to combat a new staff member’s lack of proficiency, both pre- and postoperatively.11 Additionally, chief residents and attending surgeons must always provide final approval of decisions, thereby limiting autonomy of the new residents.5,11 Further, one might expect some outcomes like VTE and SSI to be limited by standardized clinical management protocols.9,32 However, considering residents are responsible for executing such protocols and that protocols vary substantially by hospital, the simple presence of standardized measures may compensate for certain vulnerabilities in the early academic year. New residents may also be more cautious and diligent themselves, and importantly, more likely to solicit assistance rather than attempting to problem solve alone.6 As the year progresses and these residents gain experience and confidence, however, the aforementioned safeguards may be reduced and senior members may be less vigilant, thereby “diluting” or “buffering” added early risk.8 While our study is reassuring in demonstrating that the safety of plastic surgery patients does not change seasonally, the skill level, confidence, and independence of each resident at the individual level no doubt progress through the year. As graduate medical education has moved toward competency − rather than volume − based curricula, tools available to surgical residency programs such as the Objective Structured Assessment of Technical Skills (OSATS) are increasingly being utilized to establish learning curves and permit adequate monitoring of residents’ progression.52 In this vein, plastic surgery-specific tools, such as the microsurgical skill assessment methodologies developed by Starkes et al, in addition to surveys could be further studied in plastic surgery to illuminate how technical skills and confidence improve over the course of each year.53 Although the present study was well controlled and conducted with statistical rigor, it should be interpreted in the context of its methodology and limitations. Our analyses were limited to 30 day complications captured by NSQIP, which does not track several key variables. First, while robust in recording safety data, NSQIP fails to capture qualitative cosmetic outcomes. A multi-institutional prospective study determining whether cosmetic outcomes vary across the year would be an important extension of our findings. Other variables important to plastic surgeons but omitted by NSQIP include complications such as seroma or hematoma, or those occurring beyond 30 days postoperative. It was also impossible to assess readmissions while controlling for resident involvement due to lack of power. Large, retrospective intra-institutional studies, although not likely to be as statistically powerful or rigorous, have the potential to address these and other safety metrics lacking from NSQIP. Next, NSQIP does not record by operative month, but rather only by quarter of the year. Although unlikely, it is possible that the purported July Effect is confined to the month of July or is more short lived, and that enough experience is gained to make it undetectable within a single quarter. Importantly, minor errors and “near misses” on the floor are difficult to capture by any metric and may diminish the efficiency of the medical team in ways which did not ultimately impact outcomes or length of stay. Another aspect of this study that we could not control was involvement of other new members of the healthcare team in the procedure—new attending surgeons and supporting staff. Other important aspects we could not control included the aforementioned buffering/diluting factors that are taken to ensure consistently safe care as the academic year transitions in July; however, in light of our findings we affirm that regardless of the mechanism by which safety is maintained, the current system in place is one that is resilient. Further studies would be necessary to determine which of these individual buffering/diluting factors are truly substantial. Finally, it was impossible to determine whether a resident was involved in the case on the floor pre- or postoperatively since only intraoperative participation is recorded. CONCLUSIONS At this time, there is no evidence to suggest that patients undergoing plastic surgery in the first academic quarter—during the time of the “July Effect”—are at additional risk for morbidity and mortality. There is a statistically significant prolongation of operative time when residents are involved in cases during the first academic quarter in comparison to the remaining quarters; however, this is not clinically significant and represents a necessary learning curve for young surgeons. Based on the present study, patients undergoing plastic surgery procedures can be reassured that operations occurring during the beginning months of the academic calendar are not riskier, conceivably due to proper supervision and infrastructural safeguards. Graduate medical education has successfully utilized a structure of graded and progressive responsibility to foster an environment which is both effective for mentees and safe for patients. Supplementary Material This article contains supplementary material located online at www.aestheticsurgeryjournal.com. Disclosures The authors declared no potential conflicts of interest with respect to the research, authorship, and publication of this article. Funding The authors received no financial support for the research, authorship, and publication of this article. REFERENCES 1. Crane K. Headed to the hospital? Beware the ‘July Effect’. U.S. News & World Report. July 21, 2014. http://health.usnews.com/health-news/patient-advice/articles/2014/07/21/headed-to-the-hospital-beware-the-july-effect. Accessed July 2, 2015. 2. Gray M. ‘July Effect:’ Does the medical legend have a pulse? WAFF48. July 1, 2015. http://www.waff.com/story/29455649/july-effect-does-the-medical-legend-have-a-pulse. Accessed July 2, 2015. 3. Phillips DP, Barker GE. A July spike in fatal medication errors: a possible effect of new medical residents. J Gen Intern Med . 2010; 25( 8): 774- 779. Google Scholar CrossRef Search ADS PubMed  4. Englesbe MJ, Pelletier SJ, Magee JCet al.   Seasonal variation in surgical outcomes as measured by the American College of Surgeons-National Surgical Quality Improvement Program (ACS-NSQIP). Ann Surg . 2007; 246( 3): 456- 462; discussion 463-455. Google Scholar CrossRef Search ADS PubMed  5. Ehlert BA, Nelson JT, Goettler CEet al.   Examining the myth of the “July Phenomenon” in surgical patients. Surgery . 2011; 150( 2): 332- 338. Google Scholar CrossRef Search ADS PubMed  6. Ravi P, Trinh VQ, Sun Met al.   Is there any evidence of a “July effect” in patients undergoing major cancer surgery? Can J Surg . 2014; 57( 2): 82- 88. Google Scholar CrossRef Search ADS PubMed  7. Alshekhlee A, Walbert T, DeGeorgia M, Preston DC, Furlan AJ. The impact of accreditation council for graduate medical education duty hours, the July phenomenon, and hospital teaching status on stroke outcomes. J Stroke Cerebrovasc Dis . 2009; 18( 3): 232- 238. Google Scholar CrossRef Search ADS PubMed  8. Averbukh Y, Southern W. A “reverse july effect”: association between timing of admission, medical team workload, and 30-day readmission rate. J Grad Med Educ . 2014; 6( 1): 65- 70. Google Scholar CrossRef Search ADS PubMed  9. Hoashi JS, Samdani AF, Betz RR, Bastrom TP, Cahill PJ; Harms Study Group. Is there a “July effect” in surgery for adolescent idiopathic scoliosis? J Bone Joint Surg Am . 2014; 96( 7): e55. Google Scholar CrossRef Search ADS PubMed  10. Pang JH, Karipineni F, Panchal H, Campos S, Ortiz J. Seasonal variations in outcomes after kidney transplantation: UNOS review of 336,330 transplants. J Surg Educ . 2013; 70( 3): 357- 367. Google Scholar CrossRef Search ADS PubMed  11. Yaghoubian A, de Virgilio C, Chiu V, Lee SL. “July effect” and appendicitis. J Surg Educ . 2010; 67( 3): 157- 160. Google Scholar CrossRef Search ADS PubMed  12. Highstead RG, Johnson LS, Johnson LCet al.   July-as good a time as any to be injured. J Trauma . 2009; 67( 5): 1087- 1090. Google Scholar CrossRef Search ADS PubMed  13. Schroeppel TJ, Fischer PE, Magnotti LJ, Croce MA, Fabian TC. The “July phenomenon”: is trauma the exception? J Am Coll Surg . 2009; 209( 3): 378- 384. Google Scholar CrossRef Search ADS PubMed  14. Dasenbrock HH, Clarke MJ, Thompson RE, Gokaslan ZL, Bydon A. The impact of July hospital admission on outcome after surgery for spinal metastases at academic medical centers in the United States, 2005 to 2008. Cancer . 2012; 118( 5): 1429- 1438. Google Scholar CrossRef Search ADS PubMed  15. Young JQ, Ranji SR, Wachter RM, Lee CM, Niehaus B, Auerbach AD. “July effect”: impact of the academic year-end changeover on patient outcomes: a systematic review. Ann Intern Med . 2011; 155( 5): 309- 315. Google Scholar CrossRef Search ADS PubMed  16. Kim SC, Song KB, Jung YSet al.   Short-term clinical outcomes for 100 consecutive cases of laparoscopic pylorus- preserving pancreatoduodenectomy: improvement with surgical experience. Surg Endosc . 2013; 27( 1): 95- 103. Google Scholar CrossRef Search ADS PubMed  17. Budäus L, Sun M, Abdollah Fet al.   Impact of surgical experience on in-hospital complication rates in patients undergoing minimally invasive prostatectomy: a population-based study. Ann Surg Oncol . 2011; 18( 3): 839- 847. Google Scholar CrossRef Search ADS PubMed  18. Padman M, Warwick AM, Fernandes JA, Flowers MJ, Davies AG, Bell MJ. Closed reduction and stabilization of supracondylar fractures of the humerus in children: the crucial factor of surgical experience. J Pediatr Orthop B . 2010; 19( 4): 298- 303. Google Scholar CrossRef Search ADS PubMed  19. Riml S, Larcher L, Kompatscher P. Complete excision of nonmelanotic skin cancer: a matter of surgical experience. Ann Plast Surg . 2013; 70( 1): 66- 69. Google Scholar CrossRef Search ADS PubMed  20. Netsch C, Knipper AS, Orywal AK, Tiburtius C, Gross AJ. Impact of surgical experience on stone-free rates of ureteroscopy for single urinary calculi of the upper urinary tract: a matched-paired analysis of 600 patients. J Endourol . 2015; 29( 1): 78- 83. Google Scholar CrossRef Search ADS PubMed  21. Cunningham EJ, Debar S, Bell BA. Association between surgeon seniority and outcome in intracranial aneurysm surgery. Br J Neurosurg . 2003; 17( 2): 124- 128; discussion 129. Google Scholar CrossRef Search ADS PubMed  22. Ramsay CR, Grant AM, Wallace SA, Garthwaite PH, Monk AF, Russell IT. Statistical assessment of the learning curves of health technologies. Health Technol Assess . 2001; 5( 12): 1- 79. Google Scholar CrossRef Search ADS PubMed  23. Smith ER, Butler WE, Barker FG2nd. Is there a “July phenomenon” in pediatric neurosurgery at teaching hospitals? J Neurosurg . 2006; 105( 3 Suppl): 169- 176. Google Scholar PubMed  24. Gopaldas RR, Overbey DM, Dao TK, Markley JG. The impact of academic calendar cycle on coronary artery bypass outcomes: a comparison of teaching and non-teaching hospitals. J Cardiothorac Surg . 2013; 8: 191. Google Scholar CrossRef Search ADS PubMed  25. Bakaeen FG, Huh J, LeMaire SAet al.   The July effect: impact of the beginning of the academic cycle on cardiac surgical outcomes in a cohort of 70,616 patients. Ann Thorac Surg . 2009; 88( 1): 70- 75. Google Scholar CrossRef Search ADS PubMed  26. Dhaliwal AS, Chu D, Deswal Aet al.   The July effect and cardiac surgery: the effect of the beginning of the academic cycle on outcomes. Am J Surg . 2008; 196( 5): 720- 725. Google Scholar CrossRef Search ADS PubMed  27. Bohl DD, Fu MC, Golinvaux NS, Basques BA, Gruskay JA, Grauer JN. The “July effect” in primary total hip and knee arthroplasty: analysis of 21,434 cases from the ACS-NSQIP database. J Arthroplasty . 2014; 29( 7): 1332- 1338. Google Scholar CrossRef Search ADS PubMed  28. Bohl DD, Fu MC, Gruskay JA, Basques BA, Golinvaux NS, Grauer JN. “July effect” in elective spine surgery: analysis of the American College of Surgeons National Surgical Quality Improvement Program database. Spine (Phila Pa 1976) . 2014; 39( 7): 603- 611. Google Scholar CrossRef Search ADS PubMed  29. Nandyala SV, Marquez-Lara A, Fineberg SJ, Singh K. Perioperative characteristics and outcomes of patients undergoing anterior cervical fusion in July: analysis of the “July effect”. Spine (Phila Pa 1976) . 2014; 39( 7): 612- 617. Google Scholar CrossRef Search ADS PubMed  30. McDonald JS, Clarke MJ, Helm GA, Kallmes DF. The effect of July admission on inpatient outcomes following spinal surgery. J Neurosurg Spine . 2013; 18( 3): 280- 288. Google Scholar CrossRef Search ADS PubMed  31. Weaver KJ, Neal D, Hoh DJ, Mocco J, Barker FG2nd, Hoh BL. The “July phenomenon” for neurosurgical mortality and complications in teaching hospitals: an analysis of more than 850,000 neurosurgical patients in the nationwide inpatient sample database, 1998 to 2008. Neurosurgery . 2012; 71( 3): 562- 571; discussion 571. Google Scholar CrossRef Search ADS PubMed  32. McDonald RJ, Cloft HJ, Kallmes DF. Impact of admission month and hospital teaching status on outcomes in subarachnoid hemorrhage: evidence against the July effect. J Neurosurg . 2012; 116( 1): 157- 163. Google Scholar CrossRef Search ADS PubMed  33. Kim HS, Park CW, Yoo CJ, Kim EY, Kim YB, Kim WK. Impact of admission month on outcomes in spontaneous subarachnoid hemorrhage: evidence against the march effect. J Cerebrovasc Endovasc Neurosurg . 2013; 15( 2): 67- 75. Google Scholar CrossRef Search ADS PubMed  34. Birkmeyer JD, Shahian DM, Dimick JBet al.   Blueprint for a new American College of Surgeons: National Surgical Quality Improvement Program. J Am Coll Surg . 2008; 207( 5): 777- 782. Google Scholar CrossRef Search ADS PubMed  35. Ingraham AM, Richards KE, Hall BL, Ko CY. Quality improvement in surgery: the American College of Surgeons National Surgical Quality Improvement Program approach. Adv Surg . 2010; 44: 251- 267. Google Scholar CrossRef Search ADS PubMed  36. Harring TR, Nguyen NT, Liu H, Goss JA, O’Mahony CA. Liver transplant fellowship and resident training is not a part of the “July effect”. J Surg Res . 2013; 182( 1): 1- 5. Google Scholar CrossRef Search ADS PubMed  37. Bilimoria KY, Liu Y, Paruch JLet al.   Development and evaluation of the universal ACS NSQIP surgical risk calculator: a decision aid and informed consent tool for patients and surgeons. J Am Coll Surg . 2013; 217( 5): 833- 842.e1. Google Scholar CrossRef Search ADS PubMed  38. Charlson ME, Pompei P, Ales KL, MacKenzie CR. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis . 1987; 40( 5): 373- 383. Google Scholar CrossRef Search ADS PubMed  39. D’Hoore W, Bouckaert A, Tilquin C. Practical considerations on the use of the Charlson comorbidity index with administrative data bases. J Clin Epidemiol . 1996; 49( 12): 1429- 1433. Google Scholar CrossRef Search ADS PubMed  40. Sundararajan V, Henderson T, Perry C, Muggivan A, Quan H, Ghali WA. New ICD-10 version of the Charlson comorbidity index predicted in-hospital mortality. J Clin Epidemiol . 2004; 57( 12): 1288- 1294. Google Scholar CrossRef Search ADS PubMed  41. Austin PC. Some methods of propensity-score matching had superior performance to others: results of an empirical investigation and Monte Carlo simulations. Biom J . 2009; 51( 1): 171- 184. Google Scholar CrossRef Search ADS PubMed  42. Austin PC. Optimal caliper widths for propensity-score matching when estimating differences in means and differences in proportions in observational studies. Pharm Stat . 2011; 10( 2): 150- 161. Google Scholar CrossRef Search ADS PubMed  43. Austin PC. A comparison of 12 algorithms for matching on the propensity score. Stat Med . 2014; 33( 6): 1057- 1069. Google Scholar CrossRef Search ADS PubMed  44. Thoemmes F. Propensity score matching in SPSS. arXiv.org. January 30, 2012; 1201(6385v1). https://arxiv.org/abs/1201.6385. Accessed July 27, 2015. 45. Hutter MM, Kellogg KC, Ferguson CM, Abbott WM, Warshaw AL. The impact of the 80-hour resident workweek on surgical residents and attending surgeons. Ann Surg . 2006; 243( 6): 864- 871; discussion 871. Google Scholar CrossRef Search ADS PubMed  46. Accreditation Council for Graduate Medical Education. Common Program Requirements. July 1, 2011. https://www.acgme.org/Portals/0/PDFs/Common_Program_Requirements_07012011[2].pdf. Accessed April 2, 2017. 47. Kim JY, Khavanin N, Rambachan Aet al.   Surgical duration and risk of venous thromboembolism. JAMA Surg . 2015; 150( 2): 110- 117. Google Scholar CrossRef Search ADS PubMed  48. Mlodinow AS, Khavanin N, Ver Halen JP, Rambachan A, Gutowski KA, Kim JY. Increased anaesthesia duration increases venous thromboembolism risk in plastic surgery: A 6-year analysis of over 19,000 cases using the NSQIP dataset. J Plast Surg Hand Surg . 2015; 49( 4): 191- 197. Google Scholar CrossRef Search ADS PubMed  49. Qin C, de Oliveira G, Hackett N, Kim JY. Surgical duration and risk of urinary tract infection: an analysis of 1,452,369 patients using the National Surgical Quality Improvement Program (NSQIP). Int J Surg . 2015; 20: 107- 112. Google Scholar CrossRef Search ADS PubMed  50. Jordan SW, Mioton LM, Smetona Jet al.   Resident involvement and plastic surgery outcomes: an analysis of 10,356 patients from the American College of Surgeons National Surgical Quality Improvement Program database. Plast Reconstr Surg . 2013; 131( 4): 763- 773. Google Scholar CrossRef Search ADS PubMed  51. Hennessey PT, Francis HW, Gourin CG. Is there a “July effect” for head and neck cancer surgery? Laryngoscope . 2013; 123( 8): 1889- 1895. Google Scholar CrossRef Search ADS PubMed  52. Reznick RK, MacRae H. Teaching surgical skills-changes in the wind. N Engl J Med . 2006; 355( 25): 2664- 2669. Google Scholar CrossRef Search ADS PubMed  53. Starkes JL, Payk I, Hodges NJ. Developing a standardized test for the assessment of suturing skill in novice microsurgeons. Microsurgery . 1998; 18( 1): 19- 22. Google Scholar CrossRef Search ADS PubMed  © 2017 The American Society for Aesthetic Plastic Surgery, Inc. Reprints and permission: journals.permissions@oup.com http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Aesthetic Surgery Journal Oxford University Press

Demystifying the “July Effect” in Plastic Surgery: A Multi-Institutional Study

Loading next page...
 
/lp/ou_press/demystifying-the-july-effect-in-plastic-surgery-a-multi-institutional-PGZSM6BHGm
Publisher
Oxford University Press
Copyright
© 2017 The American Society for Aesthetic Plastic Surgery, Inc. Reprints and permission: journals.permissions@oup.com
ISSN
1090-820X
eISSN
1527-330X
D.O.I.
10.1093/asj/sjx099
Publisher site
See Article on Publisher Site

Abstract

Abstract Background The “July Effect” refers to a theoretical increase in complications that may occur with the influx of inexperienced interns and residents at the beginning of each academic year in July. Objectives We endeavored to determine if a July Effect occurs in plastic surgery. Methods Plastic surgery procedures were isolated from the National Surgical Quality Improvement Program registry. Cases involving residents were grouped as either having occurred within the first academic quarter (AQ1) or remaining year (AQ2-4). Groups were propensity matched using patient/operative factors and procedure type to account for baseline differences. Univariate and multivariate regression analyses assessed differences in overall complications, surgical and medical complications, individual complications, length of hospital stay, and operative time. A comparison group comprised of procedures without resident involvement was also analyzed. Results There were 5967 cases with resident involvement, 5156 of which successfully matched. Both univariate and multivariate regression analyses revealed no significant differences between AQ1 and AQ2-4 in terms of overall, surgical, medical and individual complications, or length of hospital stay. There was a statistically significant, albeit not clinically significant, increase in operative time by 10 minutes per procedure during AQ1 in comparison to AQ2-4 (P = 0.001). For procedures lacking resident participation, there were no differences between AQ1 and AQ2-4 in terms of these outcomes. Conclusions A July Effect was not observed for plastic surgery procedures in our study, conceivably due to enhanced resident oversight and infrastructural safeguards. Patients electing to undergo plastic surgery early in the academic year can be reassured of their safety during this period. Level of Evidence: 2 In the United States, the new academic year begins every July and brings anticipation and a spirit of renewal as newly minted doctors and current residents and fellows are promoted to the next postgraduate level. At the peak of trauma season, house staff may find themselves on unfamiliar services, perhaps at different hospitals and universally in new roles with new responsibilities. Interns learn under rapid fire at the first call for lists of patients. Exemplary junior residents gain senior resident status, not only with increased operative autonomy, but also with the added responsibility for mentorship and oversight of their juniors and coordination of care with the rest of the healthcare team. Meanwhile, attending surgeons are developing familiarity with the new resident class, attempting to gauge their base knowledge and skills. Experienced academic surgeons, nurses, and ancillary staff recognize the challenges that the yearly July ritual brings. In recent years, popular media, as well as the scientific literature, have called into question the hypothetical increase in medical morbidity and mortality that purportedly occurs during this turnover of healthcare staff, which has come to be known as the “July Effect” or the “July Phenomenon.”1,2 One widely referenced study reported a 10% spike in fatal medication errors during the month of July.3 Since this 2010 study, investigations across a variety of specialties have attempted to examine this phenomenon, resulting in a controversial mix of reports in favor or against existence of a July Effect.4-22 Explanatory mechanisms for the July Effect have yet to be demonstrated; however, an assortment of etiologies have been posited. Most attention is placed on inexperienced interns and residents who could adversely affect the patient during preoperative evaluation, intraoperatively, and in postoperative care and discharge. In surgery, there is a learning curve for developing technical expertise, which may affect a surgeon’s complication rate.16-22 Each progressive year, residents and new attending physicians take on increased levels of responsibility, opening up the possibility for added morbidity and mortality. Notably, this is not a transition unique to residents. Other health professionals including nurses, operating room staff, respiratory therapists, and pharmacists may also be starting their career or beginning at a new institution with foreign protocols during this transition period.14,23 Although several surgical specialties, including cardiac, orthopaedic, and neurosurgery have examined the so-called July Effect, there are currently no publications detailing this phenomenon in plastic surgery.9,23-33 As a field defined by technically diverse and complex skills in the operating room and reliant upon nuanced flap assessment and wound care on the floor, plastic surgery may be especially vulnerable to the influx of new staff who have yet to reach mastery of these skills. Further, detecting a spike in complications during the early academic year could have considerable implications for patients undergoing elective procedures, who may choose to avoid operations during this time of year. In the present investigation, the authors sought to determine what extent the first academic quarter impacts outcomes in plastic surgery by providing a statistically powerful analysis using the National Surgical Quality Improvement Program clinical registry. METHODS Institutional Review Board The Northwestern University Institutional Review Board (IRB) has declared this retrospective investigation of de-identified patient data exempt from IRB review. Data Acquisition and Patient Selection Clinical data from the 2005 to 2013 American College of Surgeon’s National Surgical Quality Improvement Program (ACS-NSQIP) patient registry were accessed in July 2015 (J.B. and M.V.). The NSQIP registry has previously been described in detail.34,35 Briefly, variables collected include patient demographics, comorbidities, operative details, and 30-day complications. The registry comprises data which were independently abstracted prospectively by trained surgical reviewers and subjected to random auditing. These measures have ensured NSQIP data to be high quality and well standardized, as supported by an overall inter-rater disagreement rate of less than 1.6%.36 Initially, all operations performed by plastic surgeons in the data set were isolated. Only procedures which accounted for >1% of plastic surgery cases were included so that procedure type could be controlled for. Procedures were selected by primary Current Procedural Terminology (CPT) codes, which are displayed in Table 1. Next, because the July Effect is most often attributed to resident involvement, cases were divided into 2 cohorts—those which involved residents and those which did not. Patients for whom intraoperative resident involvement was unknown were excluded. A separate analysis of cases which did not involve residents enabled a comparative group. Patients were stratified by academic quarter in which the operation occurred. One group in each cohort assessed for the July Effect and included operations performed in the third quarter of the year (July-September), corresponding to the first academic quarter (AQ1). This is consistent with previous literature which has used AQ1 to approximate the time frame during which the July Effect is thought to occur.5,8,13,27,28 Prior studies have used the term “July Effect” to describe varying periods from one month to three months beginning in July when the graduate medical education cycle resets.5,8,9,13,23-33 Operations performed outside the window for the July Effect were those which occurred during the remaining three academic quarters (October-June, AQ2-4). These three academic quarters were grouped and analyzed together in order to achieve the statistical power and granularity necessary to perform a tightly controlled and informative propensity match against our primary index group, AQ1. Table 1. Procedure Listing by Current Procedural Terminology (CPT) Code Procedure  CPT code(s)  Breast reduction  19318  Prosthetic reconstruction  19357  Autologous reconstruction  19361, 19364, 19367, 19368, 19369  Mastopexy  19316  Augmentation  19324, 19325  Abdominoplasty/panniculectomy  15830, 15847  Delayed/immediate insertion of prosthesis  19340, 19342  Implant removal  19328, 19330  Capsulotomy/capsulectomy  19370, 19371  Breast Revision  19380  Nipple-areola reconstruction  19350  Free flap  15756, 15757, 15758, 15842, 20969, 20970, 20972, 20973, 43496, 49906  VHR/components separation  15734, 49560, 49561, 49565, 49566, 49568  Ganglion cyst removal  25111  Procedure  CPT code(s)  Breast reduction  19318  Prosthetic reconstruction  19357  Autologous reconstruction  19361, 19364, 19367, 19368, 19369  Mastopexy  19316  Augmentation  19324, 19325  Abdominoplasty/panniculectomy  15830, 15847  Delayed/immediate insertion of prosthesis  19340, 19342  Implant removal  19328, 19330  Capsulotomy/capsulectomy  19370, 19371  Breast Revision  19380  Nipple-areola reconstruction  19350  Free flap  15756, 15757, 15758, 15842, 20969, 20970, 20972, 20973, 43496, 49906  VHR/components separation  15734, 49560, 49561, 49565, 49566, 49568  Ganglion cyst removal  25111  VHR, ventral hernia repair View Large Risk Adjustment Variables and Outcomes of Interest Patient demographic and operative characteristics were used to compare baseline differences between groups and in propensity score matching as covariates. These variables included age, body mass index (BMI), sex, a modified Charlson comorbidity index (CCI), American Society of Anesthesiologists’ (ASA) status, surgical wound classification, smoking status, diabetes, dyspnea, hypertension, functional status (dependent vs independent), bleeding disorder, setting (inpatient vs outpatient), recent chemotherapy or radiotherapy (chemo or XRT), total relative value units (RVU) for the procedure, highest postgraduate year resident involved, and primary CPT (Table 1). NSQIP-specific definitions for these variables can be found in the user manual.37 While primary CPT allowed for significant control over procedure type, a sum of total RVUs allowed for a more robust and standardized measure of procedure intensity/complexity by accounting for concurrent procedures and their added risks. The modified CCI, which caters to the comorbidities made available through databases like NSQIP, has been demonstrated to perform similarly to the original CCI and has been employed in earlier NSQIP analyses.5,27,38-40 The modified CCI variable is not captured by NSQIP directly, and was computed as previously described. Briefly, each of the following comorbidities was assigned a point value: 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.5,27,39,40 Outcomes of interest included mortality, overall incidence of complication, surgical complications, medical complications, total operative time, and length of hospital stay. Surgical complications included surgical-site infection (SSI; superficial, deep, organ/space), wound dehiscence, graft/prosthesis/flap failure, and unplanned return to the operating room. Medical complications were coma >24 hours, cerebrovascular accident (CVA)/stroke, ventilator dependence >48 hours, unplanned intubation, progressive or acute renal failure, urinary tract infection (UTI), peripheral nerve injury, pneumonia, cardiac arrest, myocardial infarction (MI), bleeding requiring transfusion, sepsis or septic shock, and venous thromboembolism (VTE; DVT, deep venous thrombosis or PE, pulmonary embolism). Any complication included both surgical and medical complications. Readmissions were not included in this study due to insufficient cases available for which both readmission status and resident involvement status were reported. Propensity Score Matching Propensity score matching using the aforementioned patient/operative characteristics and procedure types as covariates was used to eliminate statistically significant differences in baseline characteristics between operations performed in AQ1 vs the remaining year (AQ2-4). Specifically, nearest-neighbor matching without replacement in a 3:1 ratio was used, as previously described.41-43 Essentially, propensity score matching ensured that any potentially confounding differences between the patient demographics, operative factors, or procedural makeup in AQ1 vs AQ2-4 were reduced in order to more precisely determine the extent to which AQ1 influenced complications. Subgroup Analyses In the first subanalysis, three separate subgroups were isolated from the matched cohort to explore whether an effect exists for a specific population. Univariate and multivariate analyses were conducted for: (1) cases involving interns (defined as a postgraduate year 1 resident being the most senior resident present); (2) those involving senior-level residents (defined as a postgraduate year 4-6 resident being the most senior resident present); and (3) those performed in the inpatient setting only, as these patients often receive postoperative care by residents. Pairwise comparisons between AQ1 and each individual quarter were also performed. AQ2, AQ3, and AQ4 were each matched separately 1:1 with AQ1 using the original unmatched cohort involving residents. Multivariate analyses were performed for incidence of any complication and total operative time. Conceivably, the difference between AQ1 and AQ4 would be expected to be the most dramatic, as residents fine tune both their operative skills and floor management as the year progresses. Finally, to investigate whether differences between AQ1 and AQ2-4 become more pronounced as procedural complexity increases, we performed a sensitivity analysis based on total case RVU. Using matched cases including residents, multivariate analyses of overall complications and total operative time were conducted for cohorts with increasing minimum total RVUs. The analysis was done at increments of 10 RVU, ranging from >10 RVU to >60 RVU. Statistical Analysis Univariate analysis (Fischer’s exact test or χ2 for categorical variables and independent t test for continuous variables) was used to assess for differences in baseline patient/operative characteristics and postoperative outcomes between groups. To adjust risk and determine the independent effect of AQ1, we utilized binary logistic regression for 30-day complications and linear regression for total operative time and length of hospital stay. Multivariate analyses were generated for matched cohorts, and limited to complications of >30 events. In all analyses, significance was considered as two-tailed at the α = 0.05 level. All statistics were performed using SPSS version 22 (IBM Corp., Armonk, NY) and R version 2.15.3 with the PS Matching package (R Fdn. for Statistical Computing, Vienna, Austria).44 RESULTS Characteristics of Cohort With Resident Involvement There were 9728 plastic surgery procedures in the data set with intraoperative resident involvement. After eliminating primary CPT codes comprising <1% of the cohort, 5967 remained, 1493 (25.0%) from AQ1 and 4474 (75.0%) from AQ2-4. Unmatched patients did not differ significantly in terms of baseline patient/operative characteristics by academic quarter; however, they did differ by number of breast reductions (27.9% AQ1 vs 24.0% AQ2-4; P = 0.003) and capsulotomy/capsulectomies (4.8% vs 6.7%; P = 0.008) performed (Supplementary Table 1). Propensity score matching successfully eliminated these differences and maintained similarity in all other patient/operative factors between AQ1 and AQ2-4 (Table 2). Mean patient age was 48.31 years (range, 16-89 years), and 5.1% were men. After matching, there were 5156 cases, 1443 (28.0%) from AQ1 and 3713 (72.0%) from AQ2-4. Table 2. Patient/Operative Characteristics by Academic Quarter for Matched Cohorts   Without residents (n = 6118)    With residents (n = 5156)    Characteristic  AQ1  AQ 2-4  P value  AQ1  AQ 2-4  P value  No. of patients  1616 (26.4%)  4502 (73.6%)  —  1443 (28.0%)  3713 (72.0%)  —  Demographic  Age (years)  47.00 (13.44)  46.66 (13.37)  0.389  48.34 (13.29)  48.30 (13.12)  0.922  BMI (kg/m2)  28.38 (6.86)  28.09 (6.79)  0.148  28.82 (6.81)  28.65 (6.84)  0.420  Male  3.0  2.6  0.323  5.2  5.1  0.833  CCI      0.558      0.929   0  54.7  56.2  —  48.7  49.1  —   1  23.3  23.2  —  25.7  26.3  —   2  13.1  12.8  —  15.2  14.1  —   3  5.6  5.3  —  6.7  6.7  —   4  2.7  2.2  —  2.4  2.6  —   5 or greater  0.7  0.4  —  1.3  1.2  —  ASA status      0.605      0.666   1  19.2  19.9  —  16.2  16.5  —   2  66.0  66.7  —  61.2  61.2  —   3 or greater  14.7  13.4  —  22.6  22.3  —  Wound classification      0.936      0.940   II -Clean-contaminated  3.6  3.4  —  3.9  3.8  —   III-Contaminated  1.1  1.1  —  1.7  1.7  —   IV-Dirty  1.5  1.3  —  2.3  2.0  —  Smoking  14.1  13.6  0.613  12.5  12.1  0.705  Diabetes  5.4  4.8  0.349  6.4  6.4  1.000  Dyspnea  2.2  1.9  0.533  3.0  2.6  0.389  Hypertension  23.1  21.7  0.235  24.3  23.6  0.636  Dependent  0.5  0.3  0.350  1.2  1.2  0.775  Bleeding disorder  0.6  0.7  0.861  1.3  1.2  0.780  Operative factor  Inpatient  19.4  17.3  0.063  37.2  36.9  0.847  Chemo or XRT  2.5  2.2  0.495  3.1  3.3  0.793  Total RVUs  19.31 (14.85)  19.33 (14.31)  0.955  27.80 (19.74)  28.15 (20.16)  0.576  Highest PGY  —  —  —  4.68 (2.17)  4.69 (2.20)  0.873  Procedure type  Breast reduction  27.6  28.3  0.628  28.3  26.1  0.115  Prosthetic reconstruction  5.6  5.6  0.950  7.8  7.9  0.954  Autologous reconstruction  4.6  4.5  0.835  10.7  11.2  0.692  Mastopexy  7.1  6.3  0.291  3.7  3.4  0.554  Augmentation  9.8  10.7  0.368  4.4  4.5  0.881  Abdominoplasty/panniculectomy  12.3  12.3  0.965  9.6  10.2  0.605  Delayed/immediate insertion of prosthesis  5.9  5.6  0.708  8.5  9.1  0.515  Implant removal  2.7  2.7  0.928  2.3  1.8  0.212  Capsulotomy/capsulectomy  7.3  7.2  0.866  5.0  5.5  0.535  Breast revision  7.7  8.0  —  8.0  8.3  0.821  Nipple-areola reconstruction  4.8  5.3  0.513  3.7  3.8  0.871  Free flap  0.5  0.4  0.830  2.2  2.5  0.542  VHR/components separation  2.2  1.8  0.287  4.5  4.4  0.940  Ganglion cyst removal  1.8  1.3  0.143  1.2  1.3  0.783    Without residents (n = 6118)    With residents (n = 5156)    Characteristic  AQ1  AQ 2-4  P value  AQ1  AQ 2-4  P value  No. of patients  1616 (26.4%)  4502 (73.6%)  —  1443 (28.0%)  3713 (72.0%)  —  Demographic  Age (years)  47.00 (13.44)  46.66 (13.37)  0.389  48.34 (13.29)  48.30 (13.12)  0.922  BMI (kg/m2)  28.38 (6.86)  28.09 (6.79)  0.148  28.82 (6.81)  28.65 (6.84)  0.420  Male  3.0  2.6  0.323  5.2  5.1  0.833  CCI      0.558      0.929   0  54.7  56.2  —  48.7  49.1  —   1  23.3  23.2  —  25.7  26.3  —   2  13.1  12.8  —  15.2  14.1  —   3  5.6  5.3  —  6.7  6.7  —   4  2.7  2.2  —  2.4  2.6  —   5 or greater  0.7  0.4  —  1.3  1.2  —  ASA status      0.605      0.666   1  19.2  19.9  —  16.2  16.5  —   2  66.0  66.7  —  61.2  61.2  —   3 or greater  14.7  13.4  —  22.6  22.3  —  Wound classification      0.936      0.940   II -Clean-contaminated  3.6  3.4  —  3.9  3.8  —   III-Contaminated  1.1  1.1  —  1.7  1.7  —   IV-Dirty  1.5  1.3  —  2.3  2.0  —  Smoking  14.1  13.6  0.613  12.5  12.1  0.705  Diabetes  5.4  4.8  0.349  6.4  6.4  1.000  Dyspnea  2.2  1.9  0.533  3.0  2.6  0.389  Hypertension  23.1  21.7  0.235  24.3  23.6  0.636  Dependent  0.5  0.3  0.350  1.2  1.2  0.775  Bleeding disorder  0.6  0.7  0.861  1.3  1.2  0.780  Operative factor  Inpatient  19.4  17.3  0.063  37.2  36.9  0.847  Chemo or XRT  2.5  2.2  0.495  3.1  3.3  0.793  Total RVUs  19.31 (14.85)  19.33 (14.31)  0.955  27.80 (19.74)  28.15 (20.16)  0.576  Highest PGY  —  —  —  4.68 (2.17)  4.69 (2.20)  0.873  Procedure type  Breast reduction  27.6  28.3  0.628  28.3  26.1  0.115  Prosthetic reconstruction  5.6  5.6  0.950  7.8  7.9  0.954  Autologous reconstruction  4.6  4.5  0.835  10.7  11.2  0.692  Mastopexy  7.1  6.3  0.291  3.7  3.4  0.554  Augmentation  9.8  10.7  0.368  4.4  4.5  0.881  Abdominoplasty/panniculectomy  12.3  12.3  0.965  9.6  10.2  0.605  Delayed/immediate insertion of prosthesis  5.9  5.6  0.708  8.5  9.1  0.515  Implant removal  2.7  2.7  0.928  2.3  1.8  0.212  Capsulotomy/capsulectomy  7.3  7.2  0.866  5.0  5.5  0.535  Breast revision  7.7  8.0  —  8.0  8.3  0.821  Nipple-areola reconstruction  4.8  5.3  0.513  3.7  3.8  0.871  Free flap  0.5  0.4  0.830  2.2  2.5  0.542  VHR/components separation  2.2  1.8  0.287  4.5  4.4  0.940  Ganglion cyst removal  1.8  1.3  0.143  1.2  1.3  0.783  *denotes statistical significance at P < 0.05. Categorical variables are reported as percentages. Continuous variables are reported as their means (standard deviation). View Large There were no differences in any unadjusted outcomes after matching (Table 3; Supplementary Table 2 for individual complications). Postmatch rate of any complication in AQ1 was 12.4% vs 11.8% in AQ2-4 (P = 0.566), mortality (0.1% vs 0.0%; P = 0.191), surgical complication (9.5% vs 9.0%; P = 0.591), medical complication (4.6% vs 4.2%; P = 0.446), total operative time (190.56 ± 155.57 vs 182.21 ± 150.31 minutes; P = 0.076), and length of hospital stay (2.03 ± 6.01 vs 2.08 ± 8.99 days; P = 0.830). Table 3. Unadjusted Complication Rates by Academic Quarter for Matched Cohorts Complication  Without residents    With residents    AQ 2-4  AQ1    AQ 2-4  AQ1    Count  %  Count  %  P value  Count  %  Count  %  P value  Any complication  247  5.5%  85  5.3%  0.798  438  11.8%  179  12.4%  0.566  Surgical complication  198  4.4%  70  4.3%  0.944  335  9.0%  137  9.5%  0.591  Medical complication  71  1.6%  23  1.4%  0.725  155  4.2%  67  4.6%  0.446  Return to OR  93  2.1%  33  2.0%  1.000  203  5.5%  63  4.4%  0.123  Death  4  0.1%  1  0.1%  1.000  1  0.0%  2  0.1%  0.191  Total operative time (minutes)  148.96  (89.66)  153.10  (101.94)  0.149  182.21  (150.31)  190.56  (155.57)  0.076  Length of hospital stay (days)  0.89  (3.68)  0.97  (3.87)  0.488  2.08  (8.99)  2.03  (6.01)  0.830  Complication  Without residents    With residents    AQ 2-4  AQ1    AQ 2-4  AQ1    Count  %  Count  %  P value  Count  %  Count  %  P value  Any complication  247  5.5%  85  5.3%  0.798  438  11.8%  179  12.4%  0.566  Surgical complication  198  4.4%  70  4.3%  0.944  335  9.0%  137  9.5%  0.591  Medical complication  71  1.6%  23  1.4%  0.725  155  4.2%  67  4.6%  0.446  Return to OR  93  2.1%  33  2.0%  1.000  203  5.5%  63  4.4%  0.123  Death  4  0.1%  1  0.1%  1.000  1  0.0%  2  0.1%  0.191  Total operative time (minutes)  148.96  (89.66)  153.10  (101.94)  0.149  182.21  (150.31)  190.56  (155.57)  0.076  Length of hospital stay (days)  0.89  (3.68)  0.97  (3.87)  0.488  2.08  (8.99)  2.03  (6.01)  0.830  *denotes statistical significance at P < 0.05. Continuous variables are reported as their means (standard deviation). View Large Subanalyses by postgraduate year were performed on 624 cases involving interns (PGY1) and 2791 cases involving senior residents (PGY4-6). Neither univariate nor multivariate analyses revealed any significant differences between AQ1 and AQ2-4 with respect to surgical or medical complications, total operative time, or length of hospital stay. In 1907 cases with intraoperative resident involvement performed in the inpatient setting, there were no differences between AQ1 and AQ2-4 in terms of surgical or medical complications or length of hospital stay (Table 4). Total operative time in this subcohort was longer in AQ1 by multivariate (marginal effect, +16.897 minutes; P = 0.013), but not univariate analysis (291.37 ± 193.74 vs 282.30 ± 187.37 minutes; P = 0.346). Table 4. Subgroup Analyses for Inpatient Surgeries and Senior Resident Involvement Subgroup  Outcome  AQ1  AQ2-4  Univariate P value  Adjusted OR  Marginal Effect  95% CI  Multivariate P-value  Inpatient surgeries only  Any complication  22.2%  21.9%  0.902  1.056  —  0.815-1.369  0.681  Surgical complication  14.9%  15.3%  0.887  0.991  —  0.742-1.324  0.952  Medical complication  11.0%  10.1%  0.617  1.122  —  0.783-1.608  0.529  Total operative time (minutes)  291.37 (193.74)  282.30 (187.37)  0.346  —  16.897  3.538-30.256  0.013*  Length of hospital stay (days)  4.71 (7.87)  5.14 (14.27)  0.507  —  -0.280  -1.472-0.911  0.645  Senior residents (PGY 4-6) only  Any complication  12.2%  11.7%  0.744  1.090  —  0.828-1.435  0.540  Surgical complication  9.5%  9.0%  0.662  1.099  —  0.818-1.478  0.530  Medical complication  4.1%  4.6%  0.683  0.904  —  0.564-1.447  0.673  Total operative time (minutes)  185.44 (145.44)  181.43 (149.80)  0.522  —  6.607  -1.323-14.536  0.102  Length of hospital stay (days)  2.07 (6.16)  2.33 (11.66)  0.559  —  -0.282  -1.068-0.504  0.482  Subgroup  Outcome  AQ1  AQ2-4  Univariate P value  Adjusted OR  Marginal Effect  95% CI  Multivariate P-value  Inpatient surgeries only  Any complication  22.2%  21.9%  0.902  1.056  —  0.815-1.369  0.681  Surgical complication  14.9%  15.3%  0.887  0.991  —  0.742-1.324  0.952  Medical complication  11.0%  10.1%  0.617  1.122  —  0.783-1.608  0.529  Total operative time (minutes)  291.37 (193.74)  282.30 (187.37)  0.346  —  16.897  3.538-30.256  0.013*  Length of hospital stay (days)  4.71 (7.87)  5.14 (14.27)  0.507  —  -0.280  -1.472-0.911  0.645  Senior residents (PGY 4-6) only  Any complication  12.2%  11.7%  0.744  1.090  —  0.828-1.435  0.540  Surgical complication  9.5%  9.0%  0.662  1.099  —  0.818-1.478  0.530  Medical complication  4.1%  4.6%  0.683  0.904  —  0.564-1.447  0.673  Total operative time (minutes)  185.44 (145.44)  181.43 (149.80)  0.522  —  6.607  -1.323-14.536  0.102  Length of hospital stay (days)  2.07 (6.16)  2.33 (11.66)  0.559  —  -0.282  -1.068-0.504  0.482  *denotes statistical significance at P < 0.05. Data are from the matched cohort involving residents. Odds ratios and marginal effects indicate the independent influence of AQ1. Continuous variables are reported by their means (standard deviation). View Large Characteristics of Cohort Without Resident Involvement Initially, 10,623 plastic surgery cases without resident surgical involvement were identified. Approximately 70% (7453) of these met inclusion criteria by their primary CPT codes, 1675 (22.5%) from AQ1 and 5778 (77.5%) from AQ2-4. Supplementary Table 1 details the make up of this cohort by procedure type. Unmatched patients differed statistically in terms of several baseline characteristics: BMI, CCI, diabetes, inpatient procedures, total RVUs, breast revisions, and ventral hernia repair/components separation procedures (Supplementary Table 1). After propensity score matching, 6118 patients remained, 1616 (26.4%) from AQ1 and 4502 (73.6%) from AQ2-4. Mean age was 46.75 years (range, 16-89 years), and 2.7% were men. Matching successfully eliminated all statistically significant differences in patient demographics, comorbidities, operative factors and number of each procedure performed (Table 2). There were no statistically significant differences in unadjusted postoperative outcomes for matched cases performed without resident involvement (Table 3; Supplementary Table 2 for individual complications). The overall rate for complications in AQ1 was 5.3% vs 5.5% in AQ2-4 (P = 0.798), mortality (0.1% vs 0.1%; P = 1.000), surgical complication (4.3% vs 4.4%; P = 0.944), medical complication (1.4% vs 1.6%; P = 0.725), total operative time (153.10 ± 101.94 vs 148.96 ± 89.66 minutes; P = 0.149), and length of hospital stay (0.97 ± 3.87 vs 0.89 ± 3.68 days; P = 0.488). Multivariate Regression Analysis for AQ1 as a Risk Factor Table 5 displays the multivariate results for matched cohorts. With residents involved, it was found that operations occurring in AQ1 were statistically significantly longer, albeit not clinically so (marginal effect, +9.937 minutes; P = 0.001). All other outcomes for cases involving residents were insignificant for AQ1 as a risk factor, including any complication (odds ratio [OR], 1.073; 95% confidence interval [95%CI] 0.879-1.311; P = 0.488), surgical complication (OR 1.068; 95%CI 0.861-1.326; P = 0.550), medical complication (OR 1.140; 95%CI 0.821-1.584; P = 0.434), and length of hospital stay (marginal effect, −0.074 days; P = 0.749). Table 5. Multivariate Adjustment for Complications by Academic Quarter for Matched Cohorts Complication  Without residents    With residents    Adjusted OR  95% CI  P value  Adjusted OR  95% CI  P value  Any complication  1.110  0.851-1.448  0.440  1.073  0.879-1.311  0.488  Surgical complication  1.040  0.781-1.384  0.791  1.068  0.861-1.326  0.550  Medical complication  1.278  0.761-2.146  0.354  1.140  0.821-1.584  0.434  Return to OR  1.066  0.706-1.609  0.762  0.790  0.586-1.066  0.123  UTI  1.609  0.522-4.959  0.408  0.948  0.401-2.241  0.904  Pneumonia  2.195  0.227-21.251  0.497  1.002  0.185-5.436  0.999  Blood transfusion  1.159  0.485-2.771  0.740  1.279  0.841-1.947  0.250  Any sepsis  0.849  0.282-2.555  0.770  0.950  0.439-2.057  0.896   Sepsis  0.856  0.284-2.578  0.782  0.875  0.390-1.959  0.745  VTE  3.172  0.569-16.882  0.176  0.935  0.377-2.317  0.884   DVT  4.062  0.353-46.734  0.261  0.491  0.131-1.841  0.291  Any SSI  1.067  0.728-1.563  0.740  1.257  0.946-1.670  0.114   Superficial SSI  1.037  0.661-1.627  0.875  1.210  0.858-1.704  0.277   Deep SSI  1.193  0.513-2.775  0.682  1.227  0.665-2.263  0.513   Organ/space SSI  0.659  0.191-2.271  0.509  1.376  0.649-2.921  0.405  Dehiscence  1.838  0.806-4.188  0.148  1.737  0.912-3.308  0.093  Graft/prosthesis/flap failure  0.850  0.152-4.749  0.853  0.930  0.481-1.800  0.830  AQ1 Marginal effect  Marginal effect  95% CI  P value  Marginal effect  95% CI  P value  Total operative time (minutes)  3.432  −0.937-7.801  0.124  9.937  3.950-15.923  0.001*  Length of hospital stay (days)  −0.026  −0.202-0.149  0.767  −0.074  −0.527-0.379  0.749  Complication  Without residents    With residents    Adjusted OR  95% CI  P value  Adjusted OR  95% CI  P value  Any complication  1.110  0.851-1.448  0.440  1.073  0.879-1.311  0.488  Surgical complication  1.040  0.781-1.384  0.791  1.068  0.861-1.326  0.550  Medical complication  1.278  0.761-2.146  0.354  1.140  0.821-1.584  0.434  Return to OR  1.066  0.706-1.609  0.762  0.790  0.586-1.066  0.123  UTI  1.609  0.522-4.959  0.408  0.948  0.401-2.241  0.904  Pneumonia  2.195  0.227-21.251  0.497  1.002  0.185-5.436  0.999  Blood transfusion  1.159  0.485-2.771  0.740  1.279  0.841-1.947  0.250  Any sepsis  0.849  0.282-2.555  0.770  0.950  0.439-2.057  0.896   Sepsis  0.856  0.284-2.578  0.782  0.875  0.390-1.959  0.745  VTE  3.172  0.569-16.882  0.176  0.935  0.377-2.317  0.884   DVT  4.062  0.353-46.734  0.261  0.491  0.131-1.841  0.291  Any SSI  1.067  0.728-1.563  0.740  1.257  0.946-1.670  0.114   Superficial SSI  1.037  0.661-1.627  0.875  1.210  0.858-1.704  0.277   Deep SSI  1.193  0.513-2.775  0.682  1.227  0.665-2.263  0.513   Organ/space SSI  0.659  0.191-2.271  0.509  1.376  0.649-2.921  0.405  Dehiscence  1.838  0.806-4.188  0.148  1.737  0.912-3.308  0.093  Graft/prosthesis/flap failure  0.850  0.152-4.749  0.853  0.930  0.481-1.800  0.830  AQ1 Marginal effect  Marginal effect  95% CI  P value  Marginal effect  95% CI  P value  Total operative time (minutes)  3.432  −0.937-7.801  0.124  9.937  3.950-15.923  0.001*  Length of hospital stay (days)  −0.026  −0.202-0.149  0.767  −0.074  −0.527-0.379  0.749  *denotes statistical significance at P < 0.05. Odds ratios and marginal effects indicate the independent influence of AQ1. View Large In the comparison cohort lacking resident involvement, AQ1 did not significantly increase the risk for any outcome, including any complication (OR 1.110; 95%CI 0.851-1.448; P = 0.440), surgical complication (OR 1.040; 95%CI 0.781-1.384; P = 0.791), medical complication (OR 1.278; 95%CI 0.761-2.146; P = 0.354), total operative time (marginal effect, +3.432 minutes; P = 0.124), or length of hospital stay (marginal effect, −0.026 days; P = 0.767). Multivariate Comparison of Complications by Individual Academic Quarter Successful one-to-one pairwise matching by individual quarter yielded 1337 cases each for AQ1 and AQ2 in the AQ2 comparison, 1154 each for the AQ3 comparison, and 1226 each for the AQ4 comparison. On multivariate analysis, AQ1 did not differ in terms of risk for experiencing any complication overall, surgical complications, or medical complications (Table 6). Length of hospital stay was statistically significantly prolonged in AQ1 when compared to AQ4 (marginal effect, +0.429 days; P = 0.018), but not AQ2 (marginal effect, +0.042 days; P = 0.802) or AQ3 (marginal effect, +0.113 days; P = 0.621). AQ1 was also associated with increased total operative time in comparison to AQ2 (marginal effect, +9.882 minutes; P = 0.007) and AQ4 (marginal effect, +11.723 minutes; P = 0.004), but not AQ3 (marginal effect, +7.433 minutes; P = 0.070). Table 6. Multivariate Comparison of Complications by Individual Academic Quarter With Residents Academic quarter  Total cases  Any complication  Surgical complication  Medical complication  Total operative time (minutes)  Length of hospitalization (days)  Incidence  Adjusted OR  P value  Incidence  Adjusted OR  P value  Incidence  Adjusted OR  P value  Mean (SD)  Marginal effect  P value  Mean (SD)  Marginal effect  P value  AQ1  1337  12.6%  0.984  0.901  9.6%  1.022  0.875  4.8%  1.047  0.825  190.28 (154.87)  9.882  0.007*  2.05 (5.92)  0.042  0.802  AQ2  1337  12.6%  —  —  9.4%  —  —  4.4%  —  —  180.63 (141.97)  —  —  1.97 (5.17)  —  —  AQ1  1154  12.2%  1.236  0.136  9.2%  1.193  0.253  4.9%  1.557  0.068  193.66 (158.83)  7.433  0.070  2.12 (6.42)  0.113  0.621  AQ3  1154  10.5%  —  —  8.1%  —  —  3.4%  —  —  186.37 (156.19)  —  —  2.03 (6.92)  —  —  AQ1  1226  13.2%  1.104  0.459  10.2%  1.108  0.476  5.3%  1.273  0.268  194.17 (162.67)  11.723  0.004*  2.43 (7.21)  0.429  0.018*  AQ4  1226  12.2%  —  —  9.2%  —  —  4.4%  —  —  181.69 (148.87)  —  —  2.02 (4.65)  —  —  Academic quarter  Total cases  Any complication  Surgical complication  Medical complication  Total operative time (minutes)  Length of hospitalization (days)  Incidence  Adjusted OR  P value  Incidence  Adjusted OR  P value  Incidence  Adjusted OR  P value  Mean (SD)  Marginal effect  P value  Mean (SD)  Marginal effect  P value  AQ1  1337  12.6%  0.984  0.901  9.6%  1.022  0.875  4.8%  1.047  0.825  190.28 (154.87)  9.882  0.007*  2.05 (5.92)  0.042  0.802  AQ2  1337  12.6%  —  —  9.4%  —  —  4.4%  —  —  180.63 (141.97)  —  —  1.97 (5.17)  —  —  AQ1  1154  12.2%  1.236  0.136  9.2%  1.193  0.253  4.9%  1.557  0.068  193.66 (158.83)  7.433  0.070  2.12 (6.42)  0.113  0.621  AQ3  1154  10.5%  —  —  8.1%  —  —  3.4%  —  —  186.37 (156.19)  —  —  2.03 (6.92)  —  —  AQ1  1226  13.2%  1.104  0.459  10.2%  1.108  0.476  5.3%  1.273  0.268  194.17 (162.67)  11.723  0.004*  2.43 (7.21)  0.429  0.018*  AQ4  1226  12.2%  —  —  9.2%  —  —  4.4%  —  —  181.69 (148.87)  —  —  2.02 (4.65)  —  —  *denotes statistical significance at P < 0.05. Data are derived from individual academic quarters (AQs) matched head-to-head with AQ1 for procedures involving residents. Odds ratios and marginal effects indicate the independent influence of AQ1. View Large Sensitivity Analysis for Total Complications and Operative Time Stratified by Surgical Complexity Of the 5156 matched cases involving residents, 4717 had greater than 10 total RVU, 2707 greater than 20 total RVU, 1987 greater than 30 total RVU, 1011 greater than 40 total RVU, 577 greater than 50 total RVU, and 388 greater than 60 total RVU. At each RVU interval, there was no difference between AQ1 and AQ2-4 in the overall complication rate (Table 7). Total operative time was significantly longer during AQ1 for all RVU intervals, and the difference increased as case complexity increased: all procedures (+9.937 minutes; P = 0.001), >10 RVUs (+10.821 minutes; P = 0.001), >20 RVUs (+12.941 minutes; 0.010), >30 RVUs (+18.708 minutes; P = 0.003), >40 RVUs (+35.947 minutes; P = 0.001), >50 RVUs (+34.316 minutes; P = 0.030), and >60 RVUs (+48.486 minutes; P = 0.014). Table 7. Sensitivity Analysis for Total Complications and Operative Time Stratified by Surgical Complexity     Any complication    Total operative time (minutes)    Total RVUs  Total cases  AQ1 adjusted OR  95% CI  P value  AQ1 marginal effect  95% CI  P value  All  5156  1.073  0.879-1.311  0.488  9.937  3.950-15.923  0.001*  >10  4717  1.025  0.835-1.258  0.816  10.821  4.401-17.242  0.001*  >20  2707  1.161  0.907-1.488  0.236  12.941  3.105-22.776  0.010*  >30  1987  1.053  0.790-1.405  0.723  18.708  6.387-31.030  0.003*  >40  1011  1.191  0.832-1.703  0.339  35.947  15.036-56.858  0.001*  >50  577  1.129  0.704-1.810  0.614  34.316  3.371-65.261  0.030*  >60  388  1.183  0.677-2.067  0.555  48.486  9.750-87.222  0.014*      Any complication    Total operative time (minutes)    Total RVUs  Total cases  AQ1 adjusted OR  95% CI  P value  AQ1 marginal effect  95% CI  P value  All  5156  1.073  0.879-1.311  0.488  9.937  3.950-15.923  0.001*  >10  4717  1.025  0.835-1.258  0.816  10.821  4.401-17.242  0.001*  >20  2707  1.161  0.907-1.488  0.236  12.941  3.105-22.776  0.010*  >30  1987  1.053  0.790-1.405  0.723  18.708  6.387-31.030  0.003*  >40  1011  1.191  0.832-1.703  0.339  35.947  15.036-56.858  0.001*  >50  577  1.129  0.704-1.810  0.614  34.316  3.371-65.261  0.030*  >60  388  1.183  0.677-2.067  0.555  48.486  9.750-87.222  0.014*  *denotes statistical significance at P < 0.05. Data are from the matched cohort involving residents. Odds ratios and marginal effects indicate the independent influence of AQ1. Total RVUs indicates the combined total of relative value units for the procedure. View Large DISCUSSION As part of the ongoing scrutiny of resident training and patient safety, the popular news, social media and the scientific literature have continued to consider the compelling idea that the annual turnover of new residents and other healthcare workers amid the summer months results in a theoretical increase in morbidity and mortality.1,2 Findings of such a “July phenomenon” would lay the groundwork for implementing added precautionary measures and oversight aimed to support transitioning house staff. In 2003, the Accreditation Council of Graduate Medical Education (ACGME) issued a mandate to restrict resident work hours in an effort to bolster patient safety while new and inexperienced healthcare workers are trained.45 Furthermore in 2011, the ACGME introduced new supervision standards, which included definitions for “direct supervision,” “indirect supervision,” and “oversight.”46 Nonetheless, there is no substitute for experience, and thus, by nature there remain unavoidable factors in the teaching hospital setting, which may represent added risk to the patient. Prior to the present investigation, this “July Effect” has never been examined in the field of plastic surgery. It is possible that plastic surgery may be especially vulnerable due to the meticulous surgical techniques and postoperative assessment it entails. Further, the elective nature of the specialty could have implications on patient scheduling during the academic year. Using robust statistical techniques to match procedures by quarter of year, we have provided the first analysis in plastic surgery through query of the validated, multi-institutional NSQIP registry. By and large, we found insufficient evidence to support conjectures of such a phenomenon. Based on our findings, patients can be reassured that undergoing a procedure during the time of the so-called July Effect will not have an adverse impact on their outcomes. Propensity score matching provided control over differences in patient/operative factors and procedure composition between AQ1 and AQ2-4 groups. Groups were thus selected to be equivalent in every measurable aspect except for timing during the year. In both cases involving residents and those lacking resident involvement, there were no significant differences between AQ1 and the remainder of the year in terms of mortality, overall incidence of complications, length of hospital stay, or any individual complication captured by NSQIP (Tables 3 and 7). When parsed out by individual AQ, these findings were also largely recapitulated, even for the most theoretically dramatic comparison against AQ4. Likewise, sensitivity analysis for increasing procedural complexity demonstrated no differences, despite varying the cohort’s minimum RVU to as high as 60, suggesting that lower risk procedures were not masking or diluting a true effect (Table 7). Together, these results imply that even for the most demanding plastic surgery cases, the graduate medical education system is robust against increases in 30 day complications during the early academic year. Our data did demonstrate that in cases involving residents, surgical duration was statistically significantly longer in AQ1 by nearly ten minutes per operation on average compared to AQ2-4. While increased operative time has been demonstrated to correspond to a higher rate of morbidity including VTE and UTI,47-49 ten minutes on average is unlikely to be a clinically significant prolongation in this regard. This disparity became more pronounced as procedural complexity was increased in our sensitivity analysis; for example, operations took 35 minutes longer on average during AQ1 for procedures over 40 RVUs, and almost 50 minutes longer for procedures over 60 RVUs (Table 7). Although the increase in surgical duration for these most complex procedures no doubt represents a rise that is significant to the attending and medical team, most importantly we did not observe a parallel rise in 30-day complications. Other factors that we were unable to control for may be partially contributing to this observation, such as the turnover of ancillary staff in the summer. We believe that this increase in operative duration, which is marginal for the average procedure, is but a necessary and modest sacrifice to train the next generation of physicians, and that it is not feasible to eliminate the operating room learning curve altogether. Three separate subanalyses were performed to further focus our investigation on three theoretically vulnerable types of cases: those involving new residents (interns), those involving senior-level residents (PGY 4-6), and those performed in the inpatient setting. An analysis for intern participation was conducted because they possess the least clinical experience of all residents and have newly embarked on their residencies in July. Our next subanalysis was for cases involving senior-level residents because of the significantly greater autonomy both in and outside of the operating room that they gain upon graduation from junior ranks. The third analysis was performed exclusively for inpatient procedures because these procedures are more likely to involve residents in postoperative care in comparison to outpatient, providing additional opportunity for residents to impact patient outcomes. In cases involving either intern participation, senior-level residents, or inpatient cases, AQ1 again did not yield higher risk for complications or length of hospital stay. Procedures performed with interns present during AQ1 were longer by nearly 17 minutes on average. While this is a further extension of operative time beyond that observed when all resident levels are combined, it again represents only a diminution in efficiency rather than added risk to the patient, as evidenced by stable complication rates. The results in both of these subanalyses further boost our confidence in the findings of no July Effect. It is important to note that this study did not aim to gauge the impact of resident involvement itself on complication rate, which has previously been reported.50 Direct comparisons between resident-involved cases and those without residents in the present study are inappropriate because these 2 cohorts have been selected to be matched by AQ1 vs AQ2-4, and thus have considerable differences including procedure composition and preoperative patient health. This study stratified cohorts of patients based on whether or not a resident was present in the operating room in order to address the possibility that a mixed cohort could mask a July Effect theoretically driven by resident participation. Nonetheless, no such effect could be attributed to residents upon isolation of these cases only. Separation of cohorts also allowed us to investigate for the possibility of a July Effect due to other factors such as physician vacation time or the promotion of new attending surgeons and other health professionals including nurses, operating room staff, therapists, and pharmacists. Notably, without resident involvement, there was no difference in total operative time by academic quarter, as seen in the cohort with residents. It can reasonably be extrapolated that residents assisted more slowly and required additional guidance in the operating room during the beginning of a new academic year, which marginally diminished efficiency. From large multicenter databases like NSQIP and the National Inpatient Sample (NIS) to retrospective intra-institutional data, the literature has reported mixed findings on the presence or absence of a July Effect.4-15,23-33 Since 2010, when Philips and Barker reported a 10% peak in fatal medication errors during the month of July, numerous reports have refuted the existence of the phenomenon.3,5,6,8-11,23,24,27-33 Within the surgical literature, Bohl et al found no evidence of a July Effect in examination of total arthroplasty and spine surgery using NSQIP.27,28 Ehlert et al found no July Effect for the most common inpatient procedures in NSQIP, although the study did not propensity match, separate cases by resident involvement or report individual complications.5 A 2011 systematic review of 39 studies acknowledged the heterogeneity of the literature, but did conclude that mortality increases and efficiency drops during the academic turnover period.15 While increased mortality was not observed in the low risk plastic surgery procedures we examined, a modest prolongation in operative time could be appreciated. Furthermore, the 2011 review could not draw firm conclusions on morbidity, and rather, reported that many of the available studies were either limited or suboptimally analyzed. For instance, 41% of studies in the review did not adjust for confounders, many were not able to isolate cases in which residents were specifically involved and many did not use rigorous or appropriate statistical methodologies.15 Although the NIS dataset allows case selection by month rather than quarter, investigations utilizing the NIS are limited by the lack of surgical complications collected (hematoma only), categorization of teaching vs nonteaching hospital setting only and exclusion of outpatient procedures, and thus are not applicable to plastic surgery.6,7,23,24,29-32,51 Also noteworthy, many studies did not utilize propensity matching or use other methods to account for seasonal differences in case mix (eg, complex trauma vs elective).4,5,12,13,27,28 It is not out of question that certain fields or procedures actually do experience a July Effect. Thus, pointed specialty-specific investigations are warranted to avoid masking important individual findings within large all-encompassing studies. Our study has been conducted using an optimized statistical approach performed with rigor, including multivariate analysis subsequent to robustly controlled propensity score matching. In this light, we find it reasonable to conclude that our findings of no July Effect in plastic surgery are reliable and limited only by our data source. While feasibly due to other factors, the July Effect is traditionally thought to be attributable to the involvement of new interns or rising residents. There are a number of interpretations that could explain the absence of the phenomenon. First, the support network surrounding inexperienced workers may mitigate such deficits. Other members of the team, including senior residents, fellows, attending surgeons, and experienced nurses likely lend greater oversight during this delicate period.6 In the operating room, greater proportions of the surgery may be performed by the attending due to the technical immaturity of newer residents.6,11 On the floor, efficacious management and oversight by attendings and more senior-level residents may act to combat a new staff member’s lack of proficiency, both pre- and postoperatively.11 Additionally, chief residents and attending surgeons must always provide final approval of decisions, thereby limiting autonomy of the new residents.5,11 Further, one might expect some outcomes like VTE and SSI to be limited by standardized clinical management protocols.9,32 However, considering residents are responsible for executing such protocols and that protocols vary substantially by hospital, the simple presence of standardized measures may compensate for certain vulnerabilities in the early academic year. New residents may also be more cautious and diligent themselves, and importantly, more likely to solicit assistance rather than attempting to problem solve alone.6 As the year progresses and these residents gain experience and confidence, however, the aforementioned safeguards may be reduced and senior members may be less vigilant, thereby “diluting” or “buffering” added early risk.8 While our study is reassuring in demonstrating that the safety of plastic surgery patients does not change seasonally, the skill level, confidence, and independence of each resident at the individual level no doubt progress through the year. As graduate medical education has moved toward competency − rather than volume − based curricula, tools available to surgical residency programs such as the Objective Structured Assessment of Technical Skills (OSATS) are increasingly being utilized to establish learning curves and permit adequate monitoring of residents’ progression.52 In this vein, plastic surgery-specific tools, such as the microsurgical skill assessment methodologies developed by Starkes et al, in addition to surveys could be further studied in plastic surgery to illuminate how technical skills and confidence improve over the course of each year.53 Although the present study was well controlled and conducted with statistical rigor, it should be interpreted in the context of its methodology and limitations. Our analyses were limited to 30 day complications captured by NSQIP, which does not track several key variables. First, while robust in recording safety data, NSQIP fails to capture qualitative cosmetic outcomes. A multi-institutional prospective study determining whether cosmetic outcomes vary across the year would be an important extension of our findings. Other variables important to plastic surgeons but omitted by NSQIP include complications such as seroma or hematoma, or those occurring beyond 30 days postoperative. It was also impossible to assess readmissions while controlling for resident involvement due to lack of power. Large, retrospective intra-institutional studies, although not likely to be as statistically powerful or rigorous, have the potential to address these and other safety metrics lacking from NSQIP. Next, NSQIP does not record by operative month, but rather only by quarter of the year. Although unlikely, it is possible that the purported July Effect is confined to the month of July or is more short lived, and that enough experience is gained to make it undetectable within a single quarter. Importantly, minor errors and “near misses” on the floor are difficult to capture by any metric and may diminish the efficiency of the medical team in ways which did not ultimately impact outcomes or length of stay. Another aspect of this study that we could not control was involvement of other new members of the healthcare team in the procedure—new attending surgeons and supporting staff. Other important aspects we could not control included the aforementioned buffering/diluting factors that are taken to ensure consistently safe care as the academic year transitions in July; however, in light of our findings we affirm that regardless of the mechanism by which safety is maintained, the current system in place is one that is resilient. Further studies would be necessary to determine which of these individual buffering/diluting factors are truly substantial. Finally, it was impossible to determine whether a resident was involved in the case on the floor pre- or postoperatively since only intraoperative participation is recorded. CONCLUSIONS At this time, there is no evidence to suggest that patients undergoing plastic surgery in the first academic quarter—during the time of the “July Effect”—are at additional risk for morbidity and mortality. There is a statistically significant prolongation of operative time when residents are involved in cases during the first academic quarter in comparison to the remaining quarters; however, this is not clinically significant and represents a necessary learning curve for young surgeons. Based on the present study, patients undergoing plastic surgery procedures can be reassured that operations occurring during the beginning months of the academic calendar are not riskier, conceivably due to proper supervision and infrastructural safeguards. Graduate medical education has successfully utilized a structure of graded and progressive responsibility to foster an environment which is both effective for mentees and safe for patients. Supplementary Material This article contains supplementary material located online at www.aestheticsurgeryjournal.com. Disclosures The authors declared no potential conflicts of interest with respect to the research, authorship, and publication of this article. Funding The authors received no financial support for the research, authorship, and publication of this article. REFERENCES 1. Crane K. Headed to the hospital? Beware the ‘July Effect’. U.S. News & World Report. July 21, 2014. http://health.usnews.com/health-news/patient-advice/articles/2014/07/21/headed-to-the-hospital-beware-the-july-effect. Accessed July 2, 2015. 2. Gray M. ‘July Effect:’ Does the medical legend have a pulse? WAFF48. July 1, 2015. http://www.waff.com/story/29455649/july-effect-does-the-medical-legend-have-a-pulse. Accessed July 2, 2015. 3. Phillips DP, Barker GE. A July spike in fatal medication errors: a possible effect of new medical residents. J Gen Intern Med . 2010; 25( 8): 774- 779. Google Scholar CrossRef Search ADS PubMed  4. Englesbe MJ, Pelletier SJ, Magee JCet al.   Seasonal variation in surgical outcomes as measured by the American College of Surgeons-National Surgical Quality Improvement Program (ACS-NSQIP). Ann Surg . 2007; 246( 3): 456- 462; discussion 463-455. Google Scholar CrossRef Search ADS PubMed  5. Ehlert BA, Nelson JT, Goettler CEet al.   Examining the myth of the “July Phenomenon” in surgical patients. Surgery . 2011; 150( 2): 332- 338. Google Scholar CrossRef Search ADS PubMed  6. Ravi P, Trinh VQ, Sun Met al.   Is there any evidence of a “July effect” in patients undergoing major cancer surgery? Can J Surg . 2014; 57( 2): 82- 88. Google Scholar CrossRef Search ADS PubMed  7. Alshekhlee A, Walbert T, DeGeorgia M, Preston DC, Furlan AJ. The impact of accreditation council for graduate medical education duty hours, the July phenomenon, and hospital teaching status on stroke outcomes. J Stroke Cerebrovasc Dis . 2009; 18( 3): 232- 238. Google Scholar CrossRef Search ADS PubMed  8. Averbukh Y, Southern W. A “reverse july effect”: association between timing of admission, medical team workload, and 30-day readmission rate. J Grad Med Educ . 2014; 6( 1): 65- 70. Google Scholar CrossRef Search ADS PubMed  9. Hoashi JS, Samdani AF, Betz RR, Bastrom TP, Cahill PJ; Harms Study Group. Is there a “July effect” in surgery for adolescent idiopathic scoliosis? J Bone Joint Surg Am . 2014; 96( 7): e55. Google Scholar CrossRef Search ADS PubMed  10. Pang JH, Karipineni F, Panchal H, Campos S, Ortiz J. Seasonal variations in outcomes after kidney transplantation: UNOS review of 336,330 transplants. J Surg Educ . 2013; 70( 3): 357- 367. Google Scholar CrossRef Search ADS PubMed  11. Yaghoubian A, de Virgilio C, Chiu V, Lee SL. “July effect” and appendicitis. J Surg Educ . 2010; 67( 3): 157- 160. Google Scholar CrossRef Search ADS PubMed  12. Highstead RG, Johnson LS, Johnson LCet al.   July-as good a time as any to be injured. J Trauma . 2009; 67( 5): 1087- 1090. Google Scholar CrossRef Search ADS PubMed  13. Schroeppel TJ, Fischer PE, Magnotti LJ, Croce MA, Fabian TC. The “July phenomenon”: is trauma the exception? J Am Coll Surg . 2009; 209( 3): 378- 384. Google Scholar CrossRef Search ADS PubMed  14. Dasenbrock HH, Clarke MJ, Thompson RE, Gokaslan ZL, Bydon A. The impact of July hospital admission on outcome after surgery for spinal metastases at academic medical centers in the United States, 2005 to 2008. Cancer . 2012; 118( 5): 1429- 1438. Google Scholar CrossRef Search ADS PubMed  15. Young JQ, Ranji SR, Wachter RM, Lee CM, Niehaus B, Auerbach AD. “July effect”: impact of the academic year-end changeover on patient outcomes: a systematic review. Ann Intern Med . 2011; 155( 5): 309- 315. Google Scholar CrossRef Search ADS PubMed  16. Kim SC, Song KB, Jung YSet al.   Short-term clinical outcomes for 100 consecutive cases of laparoscopic pylorus- preserving pancreatoduodenectomy: improvement with surgical experience. Surg Endosc . 2013; 27( 1): 95- 103. Google Scholar CrossRef Search ADS PubMed  17. Budäus L, Sun M, Abdollah Fet al.   Impact of surgical experience on in-hospital complication rates in patients undergoing minimally invasive prostatectomy: a population-based study. Ann Surg Oncol . 2011; 18( 3): 839- 847. Google Scholar CrossRef Search ADS PubMed  18. Padman M, Warwick AM, Fernandes JA, Flowers MJ, Davies AG, Bell MJ. Closed reduction and stabilization of supracondylar fractures of the humerus in children: the crucial factor of surgical experience. J Pediatr Orthop B . 2010; 19( 4): 298- 303. Google Scholar CrossRef Search ADS PubMed  19. Riml S, Larcher L, Kompatscher P. Complete excision of nonmelanotic skin cancer: a matter of surgical experience. Ann Plast Surg . 2013; 70( 1): 66- 69. Google Scholar CrossRef Search ADS PubMed  20. Netsch C, Knipper AS, Orywal AK, Tiburtius C, Gross AJ. Impact of surgical experience on stone-free rates of ureteroscopy for single urinary calculi of the upper urinary tract: a matched-paired analysis of 600 patients. J Endourol . 2015; 29( 1): 78- 83. Google Scholar CrossRef Search ADS PubMed  21. Cunningham EJ, Debar S, Bell BA. Association between surgeon seniority and outcome in intracranial aneurysm surgery. Br J Neurosurg . 2003; 17( 2): 124- 128; discussion 129. Google Scholar CrossRef Search ADS PubMed  22. Ramsay CR, Grant AM, Wallace SA, Garthwaite PH, Monk AF, Russell IT. Statistical assessment of the learning curves of health technologies. Health Technol Assess . 2001; 5( 12): 1- 79. Google Scholar CrossRef Search ADS PubMed  23. Smith ER, Butler WE, Barker FG2nd. Is there a “July phenomenon” in pediatric neurosurgery at teaching hospitals? J Neurosurg . 2006; 105( 3 Suppl): 169- 176. Google Scholar PubMed  24. Gopaldas RR, Overbey DM, Dao TK, Markley JG. The impact of academic calendar cycle on coronary artery bypass outcomes: a comparison of teaching and non-teaching hospitals. J Cardiothorac Surg . 2013; 8: 191. Google Scholar CrossRef Search ADS PubMed  25. Bakaeen FG, Huh J, LeMaire SAet al.   The July effect: impact of the beginning of the academic cycle on cardiac surgical outcomes in a cohort of 70,616 patients. Ann Thorac Surg . 2009; 88( 1): 70- 75. Google Scholar CrossRef Search ADS PubMed  26. Dhaliwal AS, Chu D, Deswal Aet al.   The July effect and cardiac surgery: the effect of the beginning of the academic cycle on outcomes. Am J Surg . 2008; 196( 5): 720- 725. Google Scholar CrossRef Search ADS PubMed  27. Bohl DD, Fu MC, Golinvaux NS, Basques BA, Gruskay JA, Grauer JN. The “July effect” in primary total hip and knee arthroplasty: analysis of 21,434 cases from the ACS-NSQIP database. J Arthroplasty . 2014; 29( 7): 1332- 1338. Google Scholar CrossRef Search ADS PubMed  28. Bohl DD, Fu MC, Gruskay JA, Basques BA, Golinvaux NS, Grauer JN. “July effect” in elective spine surgery: analysis of the American College of Surgeons National Surgical Quality Improvement Program database. Spine (Phila Pa 1976) . 2014; 39( 7): 603- 611. Google Scholar CrossRef Search ADS PubMed  29. Nandyala SV, Marquez-Lara A, Fineberg SJ, Singh K. Perioperative characteristics and outcomes of patients undergoing anterior cervical fusion in July: analysis of the “July effect”. Spine (Phila Pa 1976) . 2014; 39( 7): 612- 617. Google Scholar CrossRef Search ADS PubMed  30. McDonald JS, Clarke MJ, Helm GA, Kallmes DF. The effect of July admission on inpatient outcomes following spinal surgery. J Neurosurg Spine . 2013; 18( 3): 280- 288. Google Scholar CrossRef Search ADS PubMed  31. Weaver KJ, Neal D, Hoh DJ, Mocco J, Barker FG2nd, Hoh BL. The “July phenomenon” for neurosurgical mortality and complications in teaching hospitals: an analysis of more than 850,000 neurosurgical patients in the nationwide inpatient sample database, 1998 to 2008. Neurosurgery . 2012; 71( 3): 562- 571; discussion 571. Google Scholar CrossRef Search ADS PubMed  32. McDonald RJ, Cloft HJ, Kallmes DF. Impact of admission month and hospital teaching status on outcomes in subarachnoid hemorrhage: evidence against the July effect. J Neurosurg . 2012; 116( 1): 157- 163. Google Scholar CrossRef Search ADS PubMed  33. Kim HS, Park CW, Yoo CJ, Kim EY, Kim YB, Kim WK. Impact of admission month on outcomes in spontaneous subarachnoid hemorrhage: evidence against the march effect. J Cerebrovasc Endovasc Neurosurg . 2013; 15( 2): 67- 75. Google Scholar CrossRef Search ADS PubMed  34. Birkmeyer JD, Shahian DM, Dimick JBet al.   Blueprint for a new American College of Surgeons: National Surgical Quality Improvement Program. J Am Coll Surg . 2008; 207( 5): 777- 782. Google Scholar CrossRef Search ADS PubMed  35. Ingraham AM, Richards KE, Hall BL, Ko CY. Quality improvement in surgery: the American College of Surgeons National Surgical Quality Improvement Program approach. Adv Surg . 2010; 44: 251- 267. Google Scholar CrossRef Search ADS PubMed  36. Harring TR, Nguyen NT, Liu H, Goss JA, O’Mahony CA. Liver transplant fellowship and resident training is not a part of the “July effect”. J Surg Res . 2013; 182( 1): 1- 5. Google Scholar CrossRef Search ADS PubMed  37. Bilimoria KY, Liu Y, Paruch JLet al.   Development and evaluation of the universal ACS NSQIP surgical risk calculator: a decision aid and informed consent tool for patients and surgeons. J Am Coll Surg . 2013; 217( 5): 833- 842.e1. Google Scholar CrossRef Search ADS PubMed  38. Charlson ME, Pompei P, Ales KL, MacKenzie CR. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis . 1987; 40( 5): 373- 383. Google Scholar CrossRef Search ADS PubMed  39. D’Hoore W, Bouckaert A, Tilquin C. Practical considerations on the use of the Charlson comorbidity index with administrative data bases. J Clin Epidemiol . 1996; 49( 12): 1429- 1433. Google Scholar CrossRef Search ADS PubMed  40. Sundararajan V, Henderson T, Perry C, Muggivan A, Quan H, Ghali WA. New ICD-10 version of the Charlson comorbidity index predicted in-hospital mortality. J Clin Epidemiol . 2004; 57( 12): 1288- 1294. Google Scholar CrossRef Search ADS PubMed  41. Austin PC. Some methods of propensity-score matching had superior performance to others: results of an empirical investigation and Monte Carlo simulations. Biom J . 2009; 51( 1): 171- 184. Google Scholar CrossRef Search ADS PubMed  42. Austin PC. Optimal caliper widths for propensity-score matching when estimating differences in means and differences in proportions in observational studies. Pharm Stat . 2011; 10( 2): 150- 161. Google Scholar CrossRef Search ADS PubMed  43. Austin PC. A comparison of 12 algorithms for matching on the propensity score. Stat Med . 2014; 33( 6): 1057- 1069. Google Scholar CrossRef Search ADS PubMed  44. Thoemmes F. Propensity score matching in SPSS. arXiv.org. January 30, 2012; 1201(6385v1). https://arxiv.org/abs/1201.6385. Accessed July 27, 2015. 45. Hutter MM, Kellogg KC, Ferguson CM, Abbott WM, Warshaw AL. The impact of the 80-hour resident workweek on surgical residents and attending surgeons. Ann Surg . 2006; 243( 6): 864- 871; discussion 871. Google Scholar CrossRef Search ADS PubMed  46. Accreditation Council for Graduate Medical Education. Common Program Requirements. July 1, 2011. https://www.acgme.org/Portals/0/PDFs/Common_Program_Requirements_07012011[2].pdf. Accessed April 2, 2017. 47. Kim JY, Khavanin N, Rambachan Aet al.   Surgical duration and risk of venous thromboembolism. JAMA Surg . 2015; 150( 2): 110- 117. Google Scholar CrossRef Search ADS PubMed  48. Mlodinow AS, Khavanin N, Ver Halen JP, Rambachan A, Gutowski KA, Kim JY. Increased anaesthesia duration increases venous thromboembolism risk in plastic surgery: A 6-year analysis of over 19,000 cases using the NSQIP dataset. J Plast Surg Hand Surg . 2015; 49( 4): 191- 197. Google Scholar CrossRef Search ADS PubMed  49. Qin C, de Oliveira G, Hackett N, Kim JY. Surgical duration and risk of urinary tract infection: an analysis of 1,452,369 patients using the National Surgical Quality Improvement Program (NSQIP). Int J Surg . 2015; 20: 107- 112. Google Scholar CrossRef Search ADS PubMed  50. Jordan SW, Mioton LM, Smetona Jet al.   Resident involvement and plastic surgery outcomes: an analysis of 10,356 patients from the American College of Surgeons National Surgical Quality Improvement Program database. Plast Reconstr Surg . 2013; 131( 4): 763- 773. Google Scholar CrossRef Search ADS PubMed  51. Hennessey PT, Francis HW, Gourin CG. Is there a “July effect” for head and neck cancer surgery? Laryngoscope . 2013; 123( 8): 1889- 1895. Google Scholar CrossRef Search ADS PubMed  52. Reznick RK, MacRae H. Teaching surgical skills-changes in the wind. N Engl J Med . 2006; 355( 25): 2664- 2669. Google Scholar CrossRef Search ADS PubMed  53. Starkes JL, Payk I, Hodges NJ. Developing a standardized test for the assessment of suturing skill in novice microsurgeons. Microsurgery . 1998; 18( 1): 19- 22. Google Scholar CrossRef Search ADS PubMed  © 2017 The American Society for Aesthetic Plastic Surgery, Inc. Reprints and permission: journals.permissions@oup.com

Journal

Aesthetic Surgery JournalOxford University Press

Published: Feb 1, 2018

There are no references for this article.

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off