Skip to main content
PMC home page
Clinical Orthopaedics and Related Research logoLink to Clinical Orthopaedics and Related Research
. 2014 Mar 22;472(7):2290–2300. doi: 10.1007/s11999-014-3567-0

The Effect of Resident Participation on Short-term Outcomes After Orthopaedic Surgery

Andrew J Pugely 1,✉, Yubo Gao 1, Christopher T Martin 1, John J Callaghan 1, Stuart L Weinstein 1, J Lawrence Marsh 1
PMCID: PMC4048420  PMID: 24658902

Abstract

Background

The influence of resident involvement on short-term outcomes after orthopaedic surgery is mostly unknown.

Questions/purposes

The purposes of our study were to examine the effects of resident involvement in surgical cases on short-term morbidity, mortality, operating time, hospital length of stay, and reoperation rate and to analyze these parameters by level of training.

Methods

The 2005–2011 American College of Surgeons National Surgical Quality Improvement Program data set was queried using Current Procedural Terminology codes for 66,817 cases across six orthopaedic procedural domains: 28,686 primary total joint arthroplasties (TJAs), 2412 revision TJAs, 16,832 basic and 5916 advanced arthroscopies, 8221 lower extremity traumas, and 4750 spine arthrodeses (fusions). Bivariate and multivariate logistic regression and propensity scores were used to build models of risk adjustment. We compared the morbidity and mortality rates, length of operating time, hospital length of stay, and reoperation rate for cases with or without resident involvement. For cases with resident participation, we analyzed the same parameters by training level.

Results

Resident participation was associated with higher morbidity in TJAs (odds ratio [OR], 1.6; range, 1.4–1.9), lower extremity trauma (OR, 1.3; range, 1.2–1.5), and fusion (OR, 1.4; range, 1.2–1.7) after adjustment. However, resident involvement was not associated with increased mortality. Operative time was greater (all p < 0.001) with resident involvement in all procedural domains. Longer hospital length of stay was associated with resident participation in lower extremity trauma (p < 0.001) and fusion cases (p = 0.003), but resident participation did not affect length of stay in other domains. Resident involvement was associated with greater 30-day reoperation rates for cases of lower extremity trauma (p = 0.041) and fusion (p < 0.001). Level of resident training did not consistently influence surgical outcomes.

Conclusions

Results of our study suggest resident involvement in surgical procedures is not associated with increased short-term major morbidity and mortality after select cases in orthopaedic surgery. Findings of longer operating times and differences in minor morbidity should lead to future initiatives to provide resident surgical skills training and improve perioperative efficiency in the academic setting.

Level of Evidence

Level II, prognostic study. See the Instructions for Authors for a complete description of levels of evidence.

Introduction

Resident education has long been a system of graduated responsibility under the guidance of an attending physician. This model is in widespread use, and teaching hospitals are recognized as providing excellent levels of care [5]. In surgical graduate medical education, surgical skills are acquired predominantly in the operating room during supervised participation in actual surgical cases. The current Accreditation Council for Graduate Medical Education (ACGME) guidelines place renewed emphasis on reducing perioperative morbidity and mortality [27], and the methods by which residents acquire surgical skills have been specifically identified as a potential area for improvement [9]. Some studies have shown that a significant learning curve exists for orthopaedic procedures and that surgeons reduce their operative times with increasing experience [7, 36].

In the general surgery literature, several studies have directly evaluated the effect of resident involvement in surgical procedures [1, 11, 14, 18, 25, 31]. Some of these studies have used the National Surgical Quality Improvement Program (NSQIP), a quality assessment tool used by the American College of Surgeons (ACS) [34]. The ACS NSQIP prospectively collects more than 250 surgical preoperative patient demographics, comorbidities, laboratory values, operative variables, and 30-day outcome variables from 483 participating institutions split into nearly a 50/50 mix of private and academic hospitals. Postoperative 30-day outcomes across 21 categories of morbidity and mortality are recorded by specially trained surgical clinical reviewers. Data collection continues prospectively for 30 days regardless of inpatient status. High data fidelity is ensured by routine auditing with a disagreement rate less than 1.8% [34]. Given these advantages, the use of the ACS NSQIP database has been accepted for use in short-term surgical outcomes in various surgical specialties, including general surgery, vascular surgery, and orthopaedic surgery [6, 13, 24].

In the general surgery literature, numerous studies have evaluated the effect of resident involvement in surgical procedures [1, 11, 14, 16, 18, 25, 31]. These studies showed there were minor differences in patient morbidity with resident surgical participation without any differences in mortality [1, 11, 18, 25, 31] and that the level of resident supervision in the operating room did not consistently influence complications [16]. Some studies also showed significantly longer operative times with simple and complex procedures [1, 11, 14]. To the best of our knowledge, no study has shown a negative effect of resident participation on short-term patient morbidity or mortality in orthopaedic procedures [4, 37].

The purpose of our study was to use the ACS NSQIP to evaluate whether resident participation in surgery affected (1) morbidity rate; (2) mortality rate; (3) length of operation; (4) hospital length of stay; and (5) reoperation rate across six major domains of orthopaedic surgery (primary total joint arthroplasty [TJA], revision TJA, basic and advanced arthroscopy, lower extremity trauma, and spine arthrodesis). A secondary goal was to subanalyze the influence of resident level of training (junior, senior, fellow) on the same outcomes.

Patients and Methods

The ACS NSQIP database was queried for patients undergoing orthopaedic surgery between 2005 and 2011. The cases were assigned to one of six procedural domains: primary TKA/THA, revision TKA/THA, basic arthroscopy (shoulder and knee), advanced arthroscopy (ACL reconstruction and rotator cuff repair), lower extremity fracture treatment (hip fracture open reduction and internal fixation [ORIF], femur/tibia intramedullary rodding, ankle fracture ORIF), and spine arthrodesis (cervical and lumbar). These domains were based on those defined by the ACGME for residency training competency. Surgical cases were selected based on Current Procedural Terminology (CPT) codes and then grouped by domain (Appendix 1). Emergency cases (except in lower extremity trauma), preoperative infections, sepsis, and multilevel complex spine fusions (levels greater than three and osteotomies) were excluded. The NSQIP database records whether a resident was present (logged as an assistant) for a case and specifies their year and level of training. In total, 66,817 patients undergoing orthopaedic surgery were identified from the NSQIP data files. In the defined surgical domains, there were 28,686 primary TJAs, 2412 revision TJAs, 16,832 basic arthroscopies, 5916 advanced arthroscopies, 8221 cases of lower extremity trauma, and 4750 spine arthrodeses. Overall, residents participated in 25.7% or 17,011 of these NSQIP cases. In the subgroups, resident involvement ranged from 20% in basic arthroscopies to 44% in revision TJAs (25% in TJAs, 24% in advanced arthroscopies, 31% in lower extremity trauma, 34% in spine arthrodeses). All data were collected prospectively by a trained, onsite, nurse abstractor. Thirty-day followup data are collected regardless of inpatient status, with disagreement rates less than 2% [2]. Multiple methods, including chart review, surgeon queries, and direct patient (or family) contact are used to ensure high data fidelity. Furthermore, the reliability administrative claims data have been questioned [8, 19], especially when directly compared with the NSQIP [21]. Overall, the use of the ACS NSQIP database has been accepted for use in short-term surgical outcomes by various surgical specialties, including general surgery, vascular surgery, and orthopaedic surgery [6, 13, 24].

We analyzed orthopaedic-relevant NSQIP patient and case variables. These included demographic data (age, sex, and race), adverse health habits, medical comorbidities, preoperative laboratory, and operative variables (Table 1) [2]. For most domains, patient demographics differed between residents and nonresidents. Patient age varied inconsistently between cohorts. In the primary TJA group, for example, patient age was slightly older in the nonresident group (66.9 versus 65.5 years, p < 0.001), whereas in the spine arthrodesis group, patient age was older in the resident group (56.2 versus 54.4 years, p < 0.001). Total case Relative Value Units were used to assess case complexity and later entered in the adjustment models. Unadjusted mean case Relative Value Units were greater for residents for spine cases (38.2 versus 30.1 units, p < 0.001). When Relative Value Units were entered in the multivariate model, they narrowed the morbidity and mortality odds ratio (OR) differences between residents and nonresidents. Each of the six orthopaedic procedural domains was analyzed for 30-day morbidity and mortality. Morbidity was defined as the occurrence of any one of the 30-day complications and mortality was defined as death within 30 days. The NSQIP reports on more than 25 short-term complications (Table 2) in the following categories: local infections, systemic infections, cardiac, hematologic, respiratory, renal, neurologic, and reoperation. Complications were subcategorized into major and minor groups [22, 28, 32]. Minor complications included superficial wound infection, pneumonia, urinary tract infection, deep vein thrombosis, blood transfusion, and renal insufficiency.

Table 1.

Patient demographics across six orthopaedic surgery domains, resident vs nonresident

Characteristic Resident involvement
No Yes p value No Yes p value No Yes p value
Primary TJA Revision TJA Basic arthroscopy
Patient demographics
 Age (years), mean 66.9* 65.5* < 0.001 66.9* 65.3* 0.002 49.7* 48.0* < 0.001
 Gender (female %) 61.6* 60.1* 0.029 56.8* 52.8* 0.056 45.0* 47.3* 0.018
 Race (white %) 80.0* 73.8* < 0.001 76.7 73.9 0.270 66.4* 66.0* < 0.001
 Emergency procedure (%) 0.4 0.5 0.385 0.4 0.9 < 0.001
Comorbidities
 BMI (kg/m2), mean 31.8 31.7 0.236 30.9 30.8 0.844 30.5* 30.2* 0.029
 Diabetes 16.2* 14.8* 0.004 16.6* 13.9* 0.067 10.4* 8.3* < 0.001
 Smoking 10.3 10.5 0.592 12.9 13.6 0.611 18.1 19.1 0.175
 COPD 4.2* 3.3* 0.001 5.0 3.7 0.149 1.7 1.1 0.007
 Congestive heart failure 0.3 0.4 0.215 0.4 0.2 0.706 0.1 0.1 0.536
 Hypertension 66.5* 63.3* < 0.001 66.0* 60.6* 0.006 34.6* 30.3* < 0.001
 Steroid use (%) 2.6* 3.6* < 0.001 3.9 4.1 0.748 0.9 0.8 0.822
 Dialysis (%) 0.2 0.2 0.931 0.5 0.1 0.149 0.1 0.1 0.745
ASA class (%) 0.539 0.255 < 0.001
 1 or 2 – no or mild disturbance 52.1 52.2 44.2 47.0 78.7* 81.9*
 3 – severe disturbance 46.0 45.7 52.3 50.2 20.7* 17.4*
 4 – life-threatening disturbance 1.9 2.1 3.6 2.8 0.6* 0.7*
Characteristic Resident involvement
No Yes p value No Yes p value No Yes p value
Advanced arthroscopy Lower extremity fracture Spine
Patient demographics
 Age (years), mean 45.4* 47.7* < 0.001 67.1* 63.5* < 0.001 54.5* 56.2* < 0.001
 Gender (female %) 37.9 39.8 0.232 64.9* 61.3* 0.002 51.8 52.1 0.850
 Race (white %) 67.8* 61.6* < 0.001 81.0* 60.0* < 0.001 79.2* 72.0* < 0.001
 Emergency procedure (%) 0.3 0.2 0.772 23.3* 26.9* < 0.001
Comorbidities
 BMI (kg/m2), mean 29.0 28.9 0.831 27.3 27.2 0.823 30.4* 29.6* < 0.001
 Diabetes 7.8 8.5 0.407 15.9 15.2 0.448 14.0 12.7 0.221
 Smoking 17.6 15.5 0.073 18.2* 20.2* 0.033 26.5* 21.6* < 0.001
 COPD 1.3 1.0 0.447 8.2 8.5 0.649 2.9 3.6 0.208
 Congestive heart failure 0.1 0.1 1.000 2.2 2.1 0.686 0.2 0.3 0.481
 Hypertension 27.2* 30.7* 0.013 54.9* 52.0* 0.013 47.4 47.6 0.905
 Steroid use (%) 0.8 0.7 0.793 3.3 4.0 0.133 2.5 3.0 0.322
 Dialysis (%) 0.1 0.0 1.000 1.3* 2.5* < 0.001 0.1 0.1 1.000
ASA class (%) 0.650 0.009 0.650
 1 or 2 – no or mild disturbance 83.3 82.3 40.5* 43.8* 65.2 64.6
 3 – severe disturbance 16.2 17.3 47.6* 44.1* 33.4 34.2
 4 – life-threatening disturbance 0.5 0.4 11.9* 12.1* 1.5 1.2
Open in a new tab

* Statistically significant difference, p < 0.05, when comparing resident vs nonresident; TJA = total joint arthroplasty; COPD = chronic obstructive pulmonary disease; ASA = American Society of Anesthesiology.

Table 2.

National Surgical Quality Improvement Program reported 30-day perioperative complications

Mortality
Morbidity
 Major complication
  Organ space infection
  Sepsis
  Septic shock
  Deep surgical site infection
  Wound dehiscence
  Pulmonary embolism
  Ventilator > 48 hours
  Unplanned intubation
  Acute renal failure
  Cardiac arrest requiring cardiopulmonary resuscitation
  Myocardial infarction
  Stroke
  Coma > 24 hours
  Graft/prosthesis/flap failure
  Return to operating room
 Minor complication
  Superficial surgical site infection
  Pneumonia
  Urinary tract infection
  Deep vein thrombosis
  Blood transfusions
  Peripheral nerve injury
  Renal insufficiency
Open in a new tab

Overall and domain-specific rates of morbidity, mortality, operative time, hospital length of stay, and rate of reoperation within 30 days were calculated.

Bivariate, multivariate logistic regression, and propensity scores were used to build models of risk adjustment. Ultimately, our goal was to build six independent propensity score-adjusted models to correct for the inherent selection bias and compare these results with the crude unadjusted resident versus nonresident complication rates. For each domain, bivariate analysis identified differences between patient characteristics and comorbidities. Student’s t-tests and chi-square analysis were used and significance was defined as p less than 0.05. SAS 9.3 for Windows (SAS Institute, Cary, NC, USA) was used to perform the statistical analysis.

Propensity scores were introduced as a method to control for selection bias between resident and nonresident groups. The propensity score is defined as the conditional probability of receiving one treatment (resident case) over another (nonresident) based on the inherent patient characteristic of covariants. This score is reported as a continuous variable between zero and one. Historically, teaching hospitals see sicker, more complex patients. The potential for selection bias is high. Although three methods of propensity score analysis have been described, matching and logistic regression [10] were used in this study. The propensity scores were created by identifying any imbalance between cohorts. First, univariate analysis was used to compare resident versus nonresident variables for each of the six domains. Any variable with a p value less than 0.1 was considered for inclusion in the propensity score.

Additionally, case complexity was controlled by including aggregate case Relative Value Units in the propensity scores. The NSQIP captures up to 11 procedural CPT codes and their corresponding Relative Value Units. For certain domains like primary TJA, where usually one CPT code was claimed, there was relatively little imbalance, but for more complex procedures such as spine arthrodesis, this imbalance was significant. Surgical operative time also was included in the propensity score. Finally, the propensity score was incorporated in a multivariate logistic regression model. These adjustments were performed independently for morbidity and mortality in each of the six surgical domains. Standard OR and 95% CIs were calculated and reported.

In a separate analysis, all resident cases from each of the six domains were subdivided into three groups of increasing resident experience: postgraduate years 1 to 3 (junior), years 4 to 5 (senior), and years greater than 5 (fellow). For each domain, these three resident levels were compared for morbidity, mortality, operative time, hospital length of stay, and 30-day rates of return to the operating room using chi-square and ANOVA statistical methods for categorical and continuous variables.

Results

Patient morbidity was greater when residents were involved in cases of revision TJA, lower extremity trauma, and spine arthrodesis (Table 3). The increased morbidity OR of resident participation in these domains decreased but still remained after propensity adjustment (Table 4). These morbidity differences were from minor complications in the revision TJA and lower extremity trauma cohorts but major and minor morbidities in spine fusion procedures.

Table 3.

Unadjusted postoperative outcomes after orthopaedic surgery

 Variable  Procedure Resident involvement p value
No Yes
Morbidity, % Primary TJA 13.45 (95% CI,12.99–13.90) 13.88 (95% CI, 13.08–14.68) 0.3534
Revision TJA 19.58 (95% CI,17.46–21.66)* 28.5 (95% CI, 25.76–31.24)* < 0.0001
Basic arthroscopy 1.61 (95% CI,1.40–1.82) 1.36 (95% CI, 0.97–1.76) 0.301
Advanced arthroscopy 1.37 (95% CI,1.00–1.74) 1.22 (95% CI, 0.64–1.79) 0.6783
Lower extremity trauma 18.53 (95% CI,17.51–19.54)* 24.57 (95% CI, 22.91–26.23)* < 0.0001
Spine fusions 13.19 (95% CI,12.02–14.36)* 23.3 (95% CI, 21.18–25.42)* < 0.0001
Mortality, % Primary TJA 0.22 (95% CI, 0.16–0.30) 0.21 (95% CI, 0.10–0.32) 0.8317
Revision TJA 0.44 (95% CI, 0.09–0.79) 0.29 (95% CI, 0.00–0.61) 0.7401
Basic arthroscopy 0.03 (95% CI, 0.00–0.06) 0.12 (95% CI, 0.00–0.24) 0.0528
Advanced arthroscopy 0 0
Lower extremity trauma 3.44 (95% CI, 2.96–3.95) 3.02 (95% CI, 2.36–3.68) 0.328
Spine fusions 0.22 (95% CI, 0.06–0.38) 0.46 (95% CI, 0.12–0.80) 0.1608
Operative time, minutes Primary TJA 93.13 (95% CI, 92.63–93.64)* 109.4 (95% CI, 108.3–110.4)* < 0.0001
Revision TJA 137.5 (95% CI, 134.0–141.0)* 158.7 (95% CI, 154.1–163.3)* < 0.0001
Basic arthroscopy 44.31 (95% CI, 43.69–44.92)* 51.47 (95% CI, 50.31–52.63)* < 0.0001
Advanced arthroscopy 95.79 (95% CI, 94.32–97.26)* 105.9 (95% CI, 103.0 –108.9)* < 0.0001
Lower extremity trauma 65.47 (95% CI, 64.40–66.53)* 92.45 (95% CI, 90.58–94.32)* < 0.0001
Spine fusions 146.6 (95% CI, 143.2–150.0)* 187.5 (95% CI, 181.0–194.0)* < 0.0001
Length of stay, days Primary TJA 3.58 (95% CI, 3.54–3.62) 3.59 (95% CI, 3.51–3.67) 0.5727
Revision TJA 4.24 (95% CI, 4.07–4.41) 4.35 (95% CI, 4.11–4.58) 0.4764
Basic arthroscopy 0.15 (95% CI, 0.11–0.19) 0.23 (95% CI, 0.15–0.30) 0.1057
Advanced arthroscopy 0.29 (95% CI, 0.19–0.39) 0.26 (95% CI, 0.20–0.31) 0.5278
Lower extremity trauma 4.74 (95% CI, 4.56–4.91)* 5.81 (95% CI, 5.41–6.19)* < 0.0001
Spine fusions 3.17 (95% CI, 3.04–3.30)* 3.69 (95% CI, 3.38–4.00)* 0.0026
Reoperation, % Primary TJA 1.63 (95% CI, 1.46–1.80) 1.7 (95% CI, 1.4–2.0) 0.6756
Revision TJA 4.96 (95% CI, 3.81–6.11) 4.7 (95% CI, 3.42–5.99) 0.7676
Basic arthroscopy 0.69 (95% CI, 0.55–0.83) 0.54 (95% CI, 0.29–0.80) 0.3645
Advanced arthroscopy 0.5 (95% CI, 0.28–0.72) 0.29 (95% CI, 0.01–0.57) 0.3567
Lower extremity trauma 2.5 (95% CI, 2.09–2.91)* 3.29 (95% CI, 2.61–3.98)* 0.0408
Spine fusions 2.61 (95% CI, 2.06–3.16)* 5.76 (95% CI, 4.59–6.93)* < 0.0001
Open in a new tab

* Statistically significant difference, p < 0.05; TJA = total joint arthroplasty.

Table 4.

Crude and propensity score-adjusted odds ratios of 30-day morbidity and mortality

Procedure Mortality Any morbidity Major morbidity Minor morbidity
Crude OR Adjusted OR Crude OR Adjusted OR Crude OR Adjusted OR Crude OR Adjusted OR
Primary TJA 0.94 (0.53–1.68) 0.98 (0.50–1.91) 1.04 (0.96–1.12) 1.05 (0.96–1.14)
Revision TJA 0.66(0.16–2.63) 1.01(0.24–4.29) 1.64 (1.36–1.98)* 1.55 (1.27–1.88)* 0.99 (0.73–1.35) 1.00 (0.73–1.39) 1.92 (1.54–2.39)* 1.77 (1.41–2.22)*
Basic arthroscopy 4.10 (1.03–16.40)* 3.41 (0.75–15.46) 0.84 (0.61–1.17) 0.84 (0.60–1.17)
Arthroscopy advanced N/A N/A 0.89 (0.51–1.54) 0.78 (0.44–1.34)
Lower extremity trauma 0.88 (0.67–1.14) 0.89 (0.66–1.20) 1.43 (1.28–1.60)* 1.34 (1.17–1.53)* 1.22 (1.04–1.43) 1.13 (0.94–1.35) 1.39 (1.21–1.60)* 1.32 (1.13–1.56)*
Spine 2.11 (0.74–6.04) 1.98 (0.63–6.18) 2.00 (1.71–2.34)* 1.44 (1.21–1.73)* 1.92 (1.50–2.46) 1.71 (1.30–2.25)* 1.94 (1.61–2.34)* 1.26 (1.01–1.57)*
Open in a new tab

* Indicate statistically significant differences, p < 0.05; OR = odds ratio; TJA = total joint arthroplasty; N/A = no calculable zero deaths.

Overall patient mortality was not greater with resident involvement. After propensity score risk adjustment, there were no differences in mortality (Table 4): primary TJA (OR, 0.98; 95% CI, 0.50–1.91), revision TJA (OR, 0.54; 95% CI, 0.17–1.71), basic arthroscopy (OR, 3.41; 95% CI, 0.75–15.46), advanced arthroscopy (no deaths), lower extremity trauma (OR, 0.84; 95% CI, 0.61–1.14), and spine arthrodesis (OR, 1.94; 95% CI, 0.68–5.40).

Operative time was greater (all p < 0.001) with resident involvement in all procedural domains (Table 3).

Longer hospital length of stay was associated with resident participation for lower extremity trauma (p < 0.001) and spine arthrodesis (p = 0.003); resident participation did not affect length of stay in other domains (Table 3).

Similar to length of stay, resident participation was associated with greater 30-day reoperation rate for lower extremity trauma (p = 0.041) and spine arthrodesis (p < 0.001).

Level of resident training did not consistently influence operative time, hospital length of stay, or 30-day surgical outcomes (Table 5). In the primary TJA domain, for example, no differences were noted among junior, senior, and fellow levels for operative time (p = 0.28), return to the operating room (p = 0.80), morbidity (p = 0.11), or mortality (p = 0.96). Additionally, in 99% of cases the attending surgeon was physically present in the operating room. In the other 36 of 17,011 cases, an attending was not present but was immediately available.

Table 5.

Resident sublevel analysis

Procedure Resident level Morbidity, (%) Mortality (%) Length of stay (days) Operative time (minutes) Return to operating room (%)
Primary TJA (n = 7162) Junior (2376) 13.85 0.21 3.42 110.46 2.19
Senior (3155) 14.64 0.22 3.6 108.49 1.43
Fellow (1631) 12.45 0.18 3.83 109.41 1.53
p value 0.114 1.000 0.001 0.278 0.079
Revision TJA (n = 1042) Junior (348) 20.11 0.57 4.06 151.46 4.02
Senior (374) 32.09 0 4.57 158.9 6.68
Fellow (320) 33.44 0.31 4.4 166.24 3.13
p value < 0.0001 0.411 0.191 0.041 0.067
Basic arthroscopy (n = 3303) Junior (1361) 1.32 0.15 0.2 50.25 0.44
Senior (1469) 1.57 0.14 0.19 51.99 0.75
Fellow (473) 0.85 0 0.41 53.35 0.21
p value 0.495 1.000 0.157 0.172 0.306
Advanced arthroscopy (n = 1396) Junior (485) 1.86 0 0.29 107.3 0.21
Senior (607) 0.82 0 0.18 107.3 0.16
Fellow (304) 0.99 0 0.35 100.9 0.66
p value 0.278 0.041 0.211 0.439
Lower extremity trauma (n = 2580) Junior (837) 21.27 3.23 4.94 87.8 3.35
Senior (1449) 25.81 2.97 5.9 94.98 3.59
Fellow (294) 27.89 2.72 7.79 93.25 1.7
p value 0.019 0.894 0.001 0.003 0.254
Spine (n = 1528) Junior (610) 22.62 0.49 3.17 199.92 4.1
Senior (220) 23.18 0.45 4.12 188.93 7.73
Fellow (698) 23.93 0.43 4.01 176.21 6.59
p value 0.856 1.000 0.030 0.004 0.062
Open in a new tab

TJA = total joint arthroplasty.

Discussion

Resident education is pivotal to providing healthcare providers in the future. All current surgeons learned their skills in a residency program. In this study, the ACS NSQIP data were examined to assess the effect resident involvement in procedures had on short-term complications after six major orthopaedic surgery domains of variable technical complexity. Small differences in minor morbidity, after complex procedures, with no associated differences in patient mortality were found. Operative times generally were longer when residents were involved in procedures. Different resident training levels did not affect the hospital length of stay, return to the operating room, morbidity, or mortality.

This study has several limitations. Although our findings show few differences in complications between resident and nonresident cohorts, the adequacy of risk adjustment is a limitation. It is known that academic hospitals often treat patients with a greater medical disease burden, a characteristic we attempted to adjust for with propensity score modeling. In an attempt to adjust for other inequalities in orthopaedic disease severity, we took several steps including excluding complex CPT codes such as spine osteotomies. In addition, aggregate case Relative Value Units were calculated and entered in our risk adjustment modeling. This technique for case mix adjustment has been described for use with the NSQIP [12, 13]. Although these techniques may help reduce some of bias, they cannot account for the wide range of orthopaedic disease severity as, for example, in patients having revision TJAs. In addition, the NSQIP does not specifically differentiate academic versus private hospitals. A previous NSQIP study [31], however, showed that the hospital type ratio is nearly identical to the resident versus nonresident ratio, implying that nearly every academic case had a resident. Other limitations included short-term, 30-day data collection; some orthopaedic complications occur after this time. In addition, the NSQIP does not capture orthopaedic-specific outcomes such as pain level and functional status. Despite these limitations, we believe the NSQIP provides clinical data of the highest integrity and is worthy of exposure to the orthopaedic community. Programs like the ACS NSQIP have been recognized nationally for decreasing morbidity and mortality for participating general surgery departments [15].

In our study, resident involvement had minimal effect on morbidity and no effect on 30-day mortality. These results are largely consistent with those of studies published in the general surgery literature using the ACS NSQIP data set [1, 11, 14, 18, 25, 31]. In these retrospective reviews, the authors concluded that resident involvement in general surgical cases was associated with a slightly higher risk of morbidity but a lower risk of mortality [31, 35]. The evidence that resident involvement in surgical cases increases errors is mixed [3, 4, 20, 37]. Although few studies in the orthopaedic literature have examined this issue, they concluded that resident involvement did not increase the incidence of complications in TJA [37] or in scoliosis surgery [4]. In the general surgery literature, one study showed a 5% increase in the incidence of complications with junior resident involvement in cholecystectomy [17]. A prospective patient-controlled trial of patients undergoing mastectomy showed no difference in complications with resident involvement [26]. In another recent study, Schoenfeld et al. [33] reviewed 43,343 orthopaedic cases and reported a small increase in morbidity for arthroplasty procedures with resident involvement but no increased morbidity in spine, hand, amputations, or sports cases. When stratified for severity of morbidity, only differences in minor complications were observed. Multiple factors may explain our findings of greater morbidity with resident involvement.

Resident participation alone should not be blamed for greater minor morbidity; the observed differences must be interpreted with caution. First, this increase in morbidity may have low clinical relevance. Second, increased morbidity with resident involvement may represent more widespread inefficacies associated with the training environment. In these hospitals, trainees pervade all aspects of care delivery from the anesthesia to nursing teams. Inexperienced anesthesia personnel may be more likely to cause airway trauma or increase time to intubation. Novice nursing staff may slow operating room setup, lack familiarity with surgeon instrumentation and preferences, and cause greater breaches in the sterile field. These inefficiencies in the academic setting will compound throughout all phases of care, increasing the risk of an adverse event. Third, academic hospitals often treat patients with a greater disease burden, a characteristic we attempted to adjust for with propensity score modeling. Ultimately, efficiency throughout all phases of care should remain an important goal, especially in the academic setting.

There were significant differences in operative time between resident and nonresident cases in all six of the surgical domains. These differences ranged from 7 minutes in basic arthroscopy to 41 minutes with spine fusions. Although statistical differences were detected, the influence of disease severity and the clinical significance of these findings are not known. In the general surgery and orthopaedic literature, resident participation leads to prolonged operative times [1, 11, 33]. It is possible that the slight increases in morbidity we found are primarily the result of the increased operative time necessary for teaching in the operating room. Some studies have associated prolonged operative times with greater morbidity, particularly surgical site infections [23, 29, 30]. Either way, preventing prolonged operative time should be an important goal when residents are involved with surgical procedures.

To our knowledge, the influence of resident participation on patients’ hospital length of stay has not been evaluated before. According to our findings, length of stay did not differ for arthroplasty and arthroscopy cases. Patients undergoing these procedures typically have more routine, streamlined postoperative protocols that do not differ much between patients. Length of stay, however, was approximately 1 day longer for patients with lower extremity trauma and ½ day longer for patients having spine arthrodesis. We suspect that the more heterogeneous nature of these types of injuries causes more variability in postoperative discharge protocols. Although we excluded patients with polytrauma and complex spinal procedures, as aforementioned, we cannot guarantee the resident and nonresident groups to be completely equal. As more residents typically work at tertiary referral centers, this may explain some differences in length of stay. Either way, hospitals, payers, and policymakers should recognize this trend and maintain higher funding for graduate medical education.

Our findings regarding reoperation rates showed a similar trend as LOS. Thirty-day reoperation rates were not different with resident participation for arthroplasty and arthroscopy. Both of these numbers were low and similar to previously reported rates [22, 30]. Reoperation rates were slightly greater (less than 1%) in cases with resident involvement for lower extremity trauma. Reoperation rates, however, were almost double for cases with resident involvement. It is known that academic centers treat some complex patients with previously failed surgeries. We also suspect that some of these reoperations may be planned or two-stage surgeries. Although the NSQIP has begun differentiating between planned and unplanned reoperations in their data collection methods, they were not separated in 2005 to 2011. As pay-for-performance programs become more widespread, policymakers should proceed cautiously as the etiology of these differences in reoperation rates are not fully understood.

Interestingly, the experience level of resident participation (junior, senior, or fellow) and supervision did not consistently influence short-term patient complications. In general, variations in patients’ length of stay, operative time, reoperation, morbidity, and mortality were not associated with higher or lower resident year in training. For example, in the primary TJA domain, no clinically significant differences were observed; operative time varied by 2 minutes between junior and senior residents, morbidity by less than 2%, and mortality by less than 0.1%. An exception to this, however, was greater morbidity, length of stay, and operative time with fellow participation in lower extremity trauma cases. We suspect these cases were of greater difficulty and more attending autonomy was given. These findings were largely consistent with those of two NSQIP studies in the general surgery literature [1, 11]. These studies reported no consistent variation in morbidity with resident level, but longer operative times with senior resident or fellow participation for select procedures. Nevertheless, we found faculty supervision over residents in the operating room was performed consistently, with the faculty member physically present in the operating room in more than 99% of cases. These results are similar to previously published data from the Veterans Affairs NSQIP [16]. We suspect that these results imply appropriate graduated responsibility among resident surgeons.

Resident involvement in orthopaedic surgery procedures was not associated with increases in major morbidity or mortality across all domains. Operative times were longer for orthopaedic surgery cases with residents involved across the range of operative complexity. These findings highlight the importance of teaching efficient healthcare delivery in the academic setting.

Appendix 1. Inclusion CPT codes and frequency

Variable CPT code Frequency Percentage
Primary total joint arthroplasty
27447 10,207 35.58
27130 18,479 64.42
Revision total joint arthroplasty
27134 843 34.95
27137 373 15.46
27138 142 5.89
27487 1054 43.70
Basic arthroscopy
29806 618 3.67
29807 790 4.69
29819 31 0.18
29820 28 0.17
29821 50 0.30
29824 415 2.47
29825 123 0.73
29826 2620 15.57
29828 78 0.46
29873 261 1.55
29874 167 0.99
29875 492 2.92
29876 358 2.13
29877 999 5.94
29879 477 2.83
29880 2299 13.66
29881 6572 39.04
29882 261 1.55
29883 57 0.34
29884 98 0.58
29885 4 0.02
29886 9 0.05
29887 25 0.15
Advanced arthroscopy
27403 1 0.02
27405 1 0.02
27407 1 0.02
27418 1 0.02
27427 1 0.02
27428 4 0.07
27429 1 0.02
27447 1 0.02
27599 1 0.02
27881 1 0.02
29827 2674 45.20
29868 2 0.03
29876 1 0.02
29877 2 0.03
29879 2 0.03
29880 35 0.59
29881 83 1.40
29882 16 0.27
29883 3 0.05
29888 3077 52.01
29889 8 0.14
Lower extremity trauma
27235 615 7.481
27236 1655 20.131
27244 772 9.391
27245 1534 18.660
27506 340 4.136
27759 327 3.978
27766 204 2.481
27769 15 0.182
27792 775 9.427
27814 888 10.802
27822 506 6.155
27823 131 1.593
27826 21 0.255
27827 146 1.776
27828 152 1.849
27829 140 1.703
Spine
22551 116 2.44
22554 530 11.16
22558 381 8.02
22590 10 0.21
22595 25 0.53
22600 108 2.27
22610 76 1.60
22612 1232 25.94
22614 117 2.46
22630 481 10.13
22800 22 0.46
22842 34 0.72
22851 16 0.34
63001 37 0.78
63003 12 0.25
63012 33 0.69
63015 21 0.44
63020 51 1.07
63030 1366 28.76
63040 11 0.23
63045 71 1.49
Open in a new tab

CPT = Current Procedural Terminology.

Footnotes

Each author certifies that he or she, or a member of his or her immediate family, has no funding or commercial associations (eg, consultancies, stock ownership, equity interest, patent/licensing arrangements, etc) that might pose a conflict of interest in connection with the submitted article. The authors’ institution and one author (AJP) received funding for this study from the Orthopaedic Trauma Association.

All ICMJE Conflict of Interest Forms for authors and Clinical Orthopaedics and Related Research editors and board members are on file with the publication and can be viewed on request.

Each author certifies that his or her institution approved or waived approval for the reporting of this case and that all investigations were conducted in conformity with ethical principles of research.

The American College of Surgeons National Surgical Quality Improvement Program and the hospitals participating in the ACS NSQIP are the source of the data used herein. They have not verified and are not responsible for the statistical validity of the data analysis or the conclusions derived by the authors.

References

  • 1.Advani V, Ahad S, Gonczy C, Markwell S, Hassan I. Does resident involvement effect surgical times and complication rates during laparoscopic appendectomy for uncomplicated appendicitis? An analysis of 16,849 cases from the ACS-NSQIP. Am J Surg. 2012;203:347–351. doi: 10.1016/j.amjsurg.2011.08.015. [DOI] [PubMed] [Google Scholar]
  • 2.American College of Surgeons National Surgical Quality Improvement Program. Program specifics: ACS NSQIP data collection overview. Available at: http://acsnsqip.org/main/programspecs/program_data_collection.jsp. Accessed September 25, 2012.
  • 3.Anderson KL, Koval KJ, Spratt KF. Hip fracture outcome: is there a “July effect”? Am J Orthop (Belle Mead NJ). 2009;38:606–611. [PubMed] [Google Scholar]
  • 4.Auerbach JD, Lonner BS, Antonacci MD, Kean KE. Perioperative outcomes and complications related to teaching residents and fellows in scoliosis surgery. Spine (Phila Pa 1976). 2008;33:1113–1118. [DOI] [PubMed]
  • 5.Ayanian JZ, Weissman JS. Teaching hospitals and quality of care: a review of the literature. Milbank Q. 2002;80:569–593, v. [DOI] [PMC free article] [PubMed]
  • 6.Belmont PJ, Jr, Davey S, Orr JD, Ochoa LM, Bader JO, Schoenfeld AJ. Risk factors for 30-day postoperative complications and mortality after below-knee amputation: a study of 2,911 patients from the National Surgical Quality Improvement Program. J Am Coll Surg. 2011;213:370–378. doi: 10.1016/j.jamcollsurg.2011.05.019. [DOI] [PubMed] [Google Scholar]
  • 7.Bjorgul K, Novicoff WM, Saleh KJ. Learning curves in hip fracture surgery. Int Orthop. 2011;35:113–119. doi: 10.1007/s00264-010-0950-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8.Cram P, Ibrahim SA, Lu X, Wolf BR. Impact of alternative coding schemes on incidence rates of key complications after total hip arthroplasty: a risk-adjusted analysis of a national data set. Geriatr Orthop Surg Rehabil. 2012;3:17–26. doi: 10.1177/2151458511435723. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9.Curet MJ. Resident work hour restrictions: where are we now? J Am Coll Surg. 2008;207:767–776. doi: 10.1016/j.jamcollsurg.2008.07.010. [DOI] [PubMed] [Google Scholar]
  • 10.D’Agostino RB., Jr Propensity score methods for bias reduction in the comparison of a treatment to a non-randomized control group. Stat Med. 1998;17:2265–2281. doi: 10.1002/(SICI)1097-0258(19981015)17:19&#x0003c;2265::AID-SIM918&#x0003e;3.0.CO;2-B. [DOI] [PubMed] [Google Scholar]
  • 11.Davis SS, Jr, Husain FA, Lin E, Nandipati KC, Perez S, Sweeney JF. Resident participation in index laparoscopic general surgical cases: impact of the learning environment on surgical outcomes. J Am Coll Surg. 2013;216:96–104. doi: 10.1016/j.jamcollsurg.2012.08.014. [DOI] [PubMed] [Google Scholar]
  • 12.Greenblatt DY, Kelly KJ, Rajamanickam V, Wan Y, Hanson T, Rettammel R, Winslow ER, Cho CS, Weber SM. Preoperative factors predict perioperative morbidity and mortality after pancreaticoduodenectomy. Ann Surg Oncol. 2011;18:2126–2135. doi: 10.1245/s10434-011-1594-6. [DOI] [PubMed] [Google Scholar]
  • 13.Greenblatt DY, Rajamanickam V, Pugely AJ, Heise CP, Foley EF, Kennedy GD. Short-term outcomes after laparoscopic-assisted proctectomy for rectal cancer: results from the ACS NSQIP. J Am Coll Surg. 2011;212:844–854. doi: 10.1016/j.jamcollsurg.2011.01.005. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14.Hernandez-Irizarry R, Zendejas B, Ali SM, Lohse CM, Farley DR. Impact of resident participation on laparoscopic inguinal hernia repairs: are residents slowing us down? J Surg Educ. 2012;69:746–752. doi: 10.1016/j.jsurg.2012.08.013. [DOI] [PubMed] [Google Scholar]
  • 15.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. doi: 10.1016/j.yasu.2010.05.003. [DOI] [PubMed] [Google Scholar]
  • 16.Itani KM, DePalma RG, Schifftner T, Sanders KM, Chang BK, Henderson WG, Khuri SF. Surgical resident supervision in the operating room and outcomes of care in Veterans Affairs hospitals. Am J Surg. 2005;190:725–731. doi: 10.1016/j.amjsurg.2005.06.042. [DOI] [PubMed] [Google Scholar]
  • 17.Kauvar DS, Braswell A, Brown BD, Harnisch M. Influence of resident and attending surgeon seniority on operative performance in laparoscopic cholecystectomy. J Surg Res. 2006;132:159–163. doi: 10.1016/j.jss.2005.11.578. [DOI] [PubMed] [Google Scholar]
  • 18.Kazaure HS, Roman SA, Sosa JA. The resident as surgeon: an analysis of ACS-NSQIP. J Surg Res. 2012;178:126–132. doi: 10.1016/j.jss.2011.12.033. [DOI] [PubMed] [Google Scholar]
  • 19.Keeney JA, Adelani MA, Nunley RM, Clohisy JC, Barrack RL. Assessing readmission databases: how reliable is the information? J Arthroplasty. 2012;27(8 suppl):72–76. e1–2. [DOI] [PubMed]
  • 20.Koval KJ, Rust CL, Spratt KF. The effect of hospital setting and teaching status on outcomes after hip fracture. Am J Orthop (Belle Mead NJ). 2011;40:19–28. [PubMed] [Google Scholar]
  • 21.Lawson EH, Louie R, Zingmond DS, Brook RH, Hall BL, Han L, Rapp M, Ko CY. A comparison of clinical registry versus administrative claims data for reporting of 30-day surgical complications. Ann Surg. 2012;256:973–981. doi: 10.1097/SLA.0b013e31826b4c4f. [DOI] [PubMed] [Google Scholar]
  • 22.Martin CT, Pugely AJ, Gao Y, Wolf BR. Risk factors for thirty-day morbidity and mortality following knee arthroscopy: a review of 12,271 patients from the National Surgical Quality Improvement Program database. J Bone Joint Surg Am. 2013;95(e98):1–10. doi: 10.2106/JBJS.L.01440. [DOI] [PubMed] [Google Scholar]
  • 23.Moucha CS, Clyburn TA, Evans RP, Prokuski L. Modifiable risk factors for surgical site infection. Instr Course Lect. 2011;60:557–564. [PubMed] [Google Scholar]
  • 24.Ortega G, Rhee DS, Papandria DJ, Yang J, Ibrahim AM, Shore AD, Makary MA, Abdullah F. An evaluation of surgical site infections by wound classification system using the ACS-NSQIP. J Surg Res. 2012;74:33–38. doi: 10.1016/j.jss.2011.05.056. [DOI] [PubMed] [Google Scholar]
  • 25.Papandria D, Rhee D, Ortega G, Zhang Y, Gorgy A, Makary MA, Abdullah F. Assessing trainee impact on operative time for common general surgical procedures in ACS-NSQIP. J Surg Educ. 2012;69:149–155. doi: 10.1016/j.jsurg.2011.08.003. [DOI] [PubMed] [Google Scholar]
  • 26.Patel SP, Gauger PG, Brown DL, Englesbe MJ, Cederna PS. Resident participation does not affect surgical outcomes, despite introduction of new techniques. J Am Coll Surg. 2010;211:540–545. doi: 10.1016/j.jamcollsurg.2010.06.008. [DOI] [PubMed] [Google Scholar]
  • 27.Privette AR, Shackford SR, Osler T, Ratliff J, Sartorelli K, Hebert JC. Implementation of resident work hour restrictions is associated with a reduction in mortality and provider-related complications on the surgical service: a concurrent analysis of 14,610 patients. Ann Surg. 2009;250:316–321. doi: 10.1097/SLA.0b013e3181ae332a. [DOI] [PubMed] [Google Scholar]
  • 28.Pugely AJ, Martin CT, Gao Y, Klocke NF, Callaghan JJ, Marsh JL. A risk calculator for short-term morbidity and mortality following hip fracture surgery. J Orthop Trauma. 2014;28:63–69. doi: 10.1097/BOT.0b013e3182a22744. [DOI] [PubMed] [Google Scholar]
  • 29.Pugely AJ, Martin CT, Gao Y, Mendoza-Lattes SA. Outpatient surgery reduces short-term complications in lumbar discectomy: an analysis of 4310 patients from the ACS-NSQIP database. Spine (Phila Pa 1976). 2013;38:264–271. [DOI] [PubMed]
  • 30.Pugely AJ, Martin CT, Gao Y, Mendoza-Lattes S, Callaghan JJ. Differences in short-term complications between spinal and general anesthesia for primary total knee arthroplasty. J Bone Joint Surg Am. 2013;95:193–199. doi: 10.2106/JBJS.K.01682. [DOI] [PubMed] [Google Scholar]
  • 31.Raval MV, Wang X, Cohen ME, Ingraham AM, Bentrem DJ, Dimick JB, Flynn T, Hall BL, Ko CY. The influence of resident involvement on surgical outcomes. J Am Coll Surg. 2011;212:889–898. doi: 10.1016/j.jamcollsurg.2010.12.029. [DOI] [PubMed] [Google Scholar]
  • 32.Schoenfeld AJ, Ochoa LM, Bader JO, Belmont PJ., Jr Risk factors for immediate postoperative complications and mortality following spine surgery: a study of 3475 patients from the National Surgical Quality Improvement Program. J Bone Joint Surg Am. 2011;93:1577–1582. doi: 10.2106/JBJS.J.01048. [DOI] [PubMed] [Google Scholar]
  • 33.Schoenfeld AJ, Serrano JA, Waterman BR, Bader JO, Belmont PJ., Jr The impact of resident involvement on post-operative morbidity and mortality following orthopaedic procedures: a study of 43,343 cases. Arch Orthop Trauma Surg. 2013;133:1483–1491. doi: 10.1007/s00402-013-1841-3. [DOI] [PubMed] [Google Scholar]
  • 34.SooHoo NF, Zingmond DS, Ko CY. Comparison of reoperation rates following ankle arthrodesis and total ankle arthroplasty. J Bone Joint Surg Am. 2007;89:2143–2149. doi: 10.2106/JBJS.F.01611. [DOI] [PubMed] [Google Scholar]
  • 35.Tseng WH, Jin L, Canter RJ, Martinez SR, Khatri VP, Gauvin J, Bold RJ, Wisner D, Taylor S, Chen SL. Surgical resident involvement is safe for common elective general surgery procedures. J Am Coll Surg. 2011;213:19–26; discussion 26–28. [DOI] [PubMed]
  • 36.Witjes S, Smolders JM, Beaule PE, Pasker P, Van Susante JL. Learning from the learning curve in total hip resurfacing: a radiographic analysis. Arch Orthop Trauma Surg. 2009;129:1293–1299. doi: 10.1007/s00402-009-0875-z. [DOI] [PubMed] [Google Scholar]
  • 37.Woolson ST, Kang MN. A comparison of the results of total hip and knee arthroplasty performed on a teaching service or a private practice service. J Bone Joint Surg Am. 2007;89:601–607. doi: 10.2106/JBJS.F.00584. [DOI] [PubMed] [Google Scholar]

Articles from Clinical Orthopaedics and Related Research are provided here courtesy of The Association of Bone and Joint Surgeons

RESOURCES