The Association Between Resident Participation and Hysterectomy Outcomes in Morbidly Obese Patients

Introduction

The prevalence of obesity in the United States is among the highest in the world. ¹ An estimated 10% of women in the United States have a body mass index (BMI) ≥40 kg/m², defined as class 3 obesity, and this number is increasing. ² Obesity confers an increased risk of associated medical comorbidities as well as an increase in surgical complications, including surgical site infections and respiratory complications. ³ Obesity-related complications have an impact on both patient outcomes and health care spending; in the United States, the annual medical costs associated with obesity have been estimated to account for >149 billion dollars. ⁴

More than 400,000 hysterectomies are performed in the United States each year as one of the most common surgical procedures. ⁵ The majority of hysterectomies are performed through a minimally invasive route (laparoscopic, laparoscopic-assisted, vaginal, or robotic-assisted). Minimally invasive hysterectomy confers several advantages over open surgery, including shorter recovery time and lower blood loss, and has been shown to be a safe option for obese patients. ⁶ The Accreditation Council for Graduate Medical Education requires that the majority of hysterectomy cases logged by resident physicians in obstetrics and gynecology (OBGYN) training programs be performed by a minimally invasive approach. ⁷ Resident participation in laparoscopic hysterectomy has been associated with an increased frequency of several perioperative complications, including postoperative transfusion, readmission, and reoperation. ⁸ Resident participation in gynecologic surgery for both benign and malignant disease has also been consistently associated with increased mean operative time.^8–11 However, the effect of resident participation during hysterectomy on perioperative complications among patients with class 3 obesity, which are at higher risk for surgical complications, is not well known.

The objective of our study was to compare intra- and postoperative complications within 30 days of hysterectomy for benign indications between patients with a BMI >40 kg/m² who had resident participation during hysterectomy and patients with no resident participation during hysterectomy.

Materials and Methods

All study procedures received approval from the University of North Carolina Institutional Review Board. Adult patients (18–99 years) with BMI >40 kg/m² who underwent hysterectomy by any route at any University of North Carolina Health Care facility between April 2014 and March 2018 were eligible for inclusion in this retrospective cohort study. A total of 225 patients within this time frame were randomly selected for inclusion. Patients with gynecologic malignancy or a history of bariatric surgery before hysterectomy were excluded.

Patient demographics, past medical and surgical history, hysterectomy indication, hysterectomy route, surgical pathology, intraoperative complications, and postoperative complications within 30 days of hysterectomy were collected by chart review. Manual chart review and data entry were performed by two individuals, with each individual responsible for entering half of the study cohort that was divided using random selection. The prevalence of obesity-related comorbidities and nonobesity-related comorbidities in each group was identified. Obesity-related comorbidities included cardiac disease (coronary artery disease and congestive heart failure), chronic kidney disease, hyperlipidemia, obstructive sleep apnea, type 2 diabetes mellitus, and hypertension. Patients were classified as either having a resident physician participate in their hysterectomy or not based on the operative report. Definitions of individual complications were agreed upon between the two reviewers and allocation conflicts were arbitrated by the senior author. Specifically, surgical site complication was defined as any surgical site infection or wound complication (including seroma and wound dehiscence) documented within 30 days of hysterectomy. The primary outcomes of interest were intraoperative and 30-day postoperative complications, which were categorized as minor (Clavien–Dindo grades 1 and 2) or major (Clavien–Dindo grade 3 and above). Examples of minor complications as defined by the Clavien–Dindo classification included transfusion, urinary tract infection, urinary retention, venous thromboembolism, ileus, and surgical site complications such as incisional cellulitis requiring antibiotics. Major complications included reoperation, visceral injury requiring surgical intervention, and pelvic hematoma or abscess requiring surgical or radiologic intervention. ¹²

Patient demographics, medical comorbidities, and surgical characteristics were compared among those who had resident participation in hysterectomy and those who had not been using chi-square, Fisher's exact, Wilcoxon, and Student's t-tests, where appropriate. A p-value <0.05 was considered statistically significant. Chi-square or Fisher's exact test was used to compare the incidence of intraoperative and 30-day postoperative complications between the two groups. Polytomous logistic regression was used to estimate the odds of major and minor perioperative complications. Binary logistic regression was used to estimate the odds of any intra- or postoperative complications. Adjusted odds ratios (aORs) were calculated, adjusting for patient race (white, black, or other), obesity-related comorbidities (yes/no), prior abdominal surgery (yes/no), additional procedures performed during hysterectomy (yes/no), and specimen weight (continuous). Covariates included in the models were identified as confounders a priori from a directed acyclic graph. All analyses were performed using SAS software version 9.4 (SAS, Inc., Cary, NC).

Results

There were several differences in demographic and surgical characteristics between hysterectomy cases with resident participation and cases without resident participation (Table 1). Despite no significant difference in mean BMI between groups, there was a higher frequency of diabetes mellitus (25.0% vs. 12.0%, p = 0.01) and hypertension (66.7% vs. 45.3%, p < 0.01) among patients in the resident participation group, as well as a higher proportion of both obesity-related medical comorbidities (80.6% vs. 53.9%, p < 0.01) and nonobesity-related medical comorbidities (44.4% vs. 30.8%, p = 0.03) in this group. However, more patients in the group without resident participation had American Society of Anesthesiologists Class 4 (16.2% vs. 2.8%, p < 0.01). Median specimen weight tended to be higher in the resident participation group (214.0 vs. 190.2 g, p = 0.22) and more patients in this group had pelvic mass as the indication for hysterectomy (10.2% vs. 2.6%, p = 0.03). There was a higher proportion of both abdominal hysterectomies (16.7% vs. 5.1%, p < 0.01) and laparoscopic hysterectomies (45.4% vs. 25.6%, p < 0.01) in the resident participation group, whereas the group with no resident participation had a higher frequency of robotic-assisted laparoscopic hysterectomies (54.7% vs. 29.6%, p < 0.01). Operative time was significantly higher in the resident participation group (241.6 minutes compared with 172.9 minutes, p < 0.01).

The frequency of intraoperative complications also differed between patients who had resident participation in hysterectomy and those who did not (Table 2). In terms of individual intraoperative complications, there was a higher frequency of intraoperative hemorrhage or transfusion in the resident participation group, compared with the no resident participation group (5.6% vs. 0%, p = 0.01) and no significant differences in conversion to laparotomy or visceral injury. There was a higher frequency of any intraoperative complication in the group with resident participation (12.0% compared with 3.4%, p = 0.02). There was also a higher frequency of any postoperative complication within 30 days of hysterectomy in the resident participation group (23.2% vs. 11.1%, p = 0.02) (Table 2). Specifically, there was a higher proportion of surgical site complication (7.4% vs. 0.9%, p = 0.02) and a higher frequency of length of hospital stay >1 day, although this finding did not reach statistical significance (92.5% vs. 84.5%, p = 0.06).

Overall, there was a higher odds of any intra- or postoperative complication in the resident participation group (cOR 2.48; 95% confidence interval [CI], 1.28–4.79). After adjusting for demographic and surgical characteristics known to increase case complexity, including race, obesity-related comorbidities, prior abdominal surgery, other procedures performed during hysterectomy, and specimen weight, there remained a higher odds of any intraoperative complication in the resident participation group (aOR 4.08; 95% CI, 1.14–14.56) and minor perioperative complications (Clavien–Dindo grade 1 or 2; aOR 2.56; 95% CI, 1.07–6.13). Considering all perioperative complications, there was no significant difference in major complications (Clavien–Dindo grade 3 or higher) between hysterectomy cases with resident participation and cases without resident participation (aOR 1.88; 95% CI, 0.47–7.53) (Table 3).

Discussion

In our cohort of patients with class 3 obesity undergoing hysterectomy, we found several demographic differences between patients who had a resident physician participate in their surgery and those who did not. Although BMI was similar between the groups, the group with resident participation in hysterectomy had a higher mean age and specimen weight as well as a higher frequency of both obesity- and nonobesity-related medical comorbidities. The resident participation group had a higher frequency of intraoperative hemorrhage or transfusion, surgical site complication, and four times the odds of any intraoperative complication during hysterectomy. Even after adjusting for factors known to increase case complexity, including specimen weight and obesity-related comorbidities, we still found a significantly higher odds of intraoperative and minor perioperative complications in the resident participation group. In addition, resident participation was associated with an increase in operative time of 69 minutes on average. However, we did not find an overall difference in major perioperative complications among patients with resident participation in hysterectomy. The group with resident participation had a higher frequency of both abdominal and laparoscopic hysterectomies, which is attributable to the significantly higher frequency of robotic-assisted hysterectomies in the group without resident participation.

Our findings concur with prior study in both gynecologic and general surgery. In concordance with Igwe et al., we found that resident participation in hysterectomy was associated with longer mean operative time and increased frequency of transfusion without a difference in major complications. ⁸ In contrast to Igwe et al., ⁸ we did not observe a higher frequency of readmission or reoperation in the resident participation group. In the bariatric surgery literature, resident participation in laparoscopic Roux-en-Y gastric bypass has been associated with an increased frequency of superficial site infection and operative time, with a direct linear association with resident training level. ¹³ In our study, the significantly increased operative time in the resident participation group (69 minutes longer on average) may contribute to the increased frequency of surgical site complications and intraoperative hemorrhage or transfusion in this group, as blood loss and risk of surgical site infection are expected to increase in proportion to operative time.

Our study has significant implications for curriculum development in OBGYN residency training programs. Although a correlation with an increased frequency of perioperative complications cannot imply causation, this study highlights the importance of identifying cases that are appropriate for residents of different training levels as well as identifying strategies to maintain resident involvement in complex cases without compromising patient safety. Particularly for early-stage learners, laparoscopic simulation may improve psychomotor skills. ¹⁴ Proficiency-based training programs have been shown to improve the acquisition of minimally invasive skills that transfer favorably to clinical practice. ¹⁵ In comparison with other surgical specialties, OBGYN residency programs would specifically benefit from simulation curricula as OBGYN residents spend considerably less time on surgical rotations and the incidence of hysterectomy is decreasing. Given the rising incidence of class 3 obesity in our population and the increased perioperative risks in this cohort, it would behoove residency programs to implement a combination of simulation-based exercises, preoperative warm-up exercises, and surgical coaching to help residents acquire and maintain surgical skills.

Future study might also examine if patient insurance type has any correlation to perioperative complications among patients undergoing hysterectomy, as socioeconomic status may differ between patients undergoing hysterectomy at a teaching hospital with resident participation and those who pursue surgery with a gynecologic surgeon in private practice and may contribute to disparities in perioperative outcomes.

Although we were able to identify significant differences in individual intra- and postoperative complications, limitations of our study include our lack of sufficient power to detect additional small differences in individual complications between groups. We attempted to address this limitation by combining complications into groups of any complication, major complications, and minor complications based on an objective scale, the Clavien–Dindo classification (Table 4). Patients presenting to an outside facility for a complication related to hysterectomy would not be captured in our chart review, and documentation of intra- and postoperative complications may differ between surgeons at the sites included in our study. We attempted to address this limitation by including all UNC Health Care facilities, which are located across the state of North Carolina, in our data abstraction.

Despite these limitations, our study has many strengths. We performed manual chart review to minimize misclassification errors that can occur with large database studies. We included all UNC Health Care facilities in our data abstraction, which include a heterogeneous population of academic generalists, fellowship-trained surgeons, and private practice gynecologists at facilities across the state of North Carolina. We included all routes of hysterectomy in our study to identify variations in practice patterns between attending physicians who work with resident physicians and those who did not. Future study may focus on laparoscopic or abdominal hysterectomy alone to further isolate the effect of resident participation on hysterectomy outcomes based on route.

In conclusion, we found that resident participation in hysterectomy in patients with class 3 obesity was associated with an increased frequency of intraoperative complications and minor perioperative complications, but no difference in major complications, and a significant increase in operative time. It is crucial that graduates of OBGYN residency training programs feel prepared to perform hysterectomy in morbidly obese patients, a growing population in the United States with increased surgical risk.