Identifying potentially modifiable risk factors associated with racially disparate postoperative outcomes following benign hysterectomy.

In this retrospective cohort study, Black patients were twice as likely to experience a major postoperative complication following hysterectomy compared with their White counterparts. After adjusting for patient- and surgery-level risk factors, Black patients had a 39% higher likelihood of experiencing a major postoperative complication following hysterectomy. Although some risk factors are nonmodifiable, such as ASA class and uterine weight, this analysis highlights several factors that can be modified, with the goal of reducing the racially disparate outcomes following hysterectomy. In a multivariable analysis controlling for nonmodifiable factors including BMI, ASA class, diabetes, and uterine weight, patients who underwent a laparotomy surgical approach were 39% more likely to have a major postoperative complication than those who underwent an MIS approach. Each additional hour of operative time was associated with 13% increased likelihood of major complications. Use of nonpreferred preoperative antibiotic regimens and hemostatic agents was associated with 50% and 55% higher likelihood of major postoperative complications, respectively. Compared with patients who underwent hysterectomy performed by a high-volume surgeon, patients who underwent surgery performed by a low-volume surgeon were 45% more likely to experience a major postoperative complication. It is essential to emphasize that race is a social rather than biological construct. As a recent systematic review and expert commentary on racial disparities thoughtfully concluded, “root causes of racial health disparities are much more complex than genes, individual health behaviors, and access to care.” 15 We acknowledge that race is a factor of interest not because of inherent genetic or biological differences, but because it represents systematic and persistent exposure to chronic physiological and psychological stress associated with navigating personal and structural racism over one’s lifetime, with resulting epigenetic, interpersonal, and broader social implications. Previously published studies have investigated the relationship between racial disparities in hysterectomy outcomes and patient-level risk factors, such as uterine weight or medical comorbidities. Fibroids, which substantially contribute to uterine weight, may be a key factor underlying these disparities. In addition to having a greater lifetime incidence of fibroids, Black women are diagnosed with fibroids at earlier ages, have greater size and growth rates of fibroids, and are more likely to undergo surgical intervention than other racial groups. However, in our analysis, Black women were at greater risk for experiencing a postoperative complication after adjusting for uterine weight. This analysis builds upon existing literature demonstrating that racial disparities in postoperative outcomes persist after accounting for patient-level perioperative risk factors. 8 , 13 , 14 It is essential to further explore the factors that contribute to the persistence of racially disparate outcomes. Herein, we evaluate the impact of modifiable factors that can be incorporated into clinical practice with the goal of reducing increased perioperative morbidity experienced by Black patients. Our findings are consistent with well-established evidence supporting the use of a minimally invasive surgical approach to hysterectomy whenever feasible. Laparotomy is associated with substantially higher complication rates compared with minimally invasive hysterectomy, even in cases with markedly enlarged uteri. 24 Although minimally invasive hysterectomy is associated with longer operative times compared with laparotomy, prior literature indicates that there is no operative duration at which the risk for laparotomy is lower than that for laparoscopy. 25 As noted above, the potential impact of laparotomy on major postoperative complications is considerably greater than that of multiple additional hours of operative time. Although Black patients tend to have larger average uterine weight than White patients, studies show that racial disparities in minimally invasive hysterectomy rates can be meaningfully reduced—or even eliminated—through quality improvement efforts. These include implementing best clinical and surgical practices, changing referral patterns, enhancing surgeon training, and reducing the number of low-volume surgeons. 26 In our analysis, low surgeon volume tertile was independently associated with higher postoperative complications, even after adjusting for MIS approach and other perioperative best practices. These findings are consistent with extensive data demonstrating lower complication rates for hysterectomies and other gynecologic surgical procedures completed by higher-volume surgeons compared with low- or very low-volume surgeons. 16 , 17 , 27 – 29 There are also extensive data demonstrating that high-volume surgeons or minimally invasive gynecologic surgery fellowship—trained surgeons are more likely to complete hysterectomy via a minimally invasive approach. 17 , 21 , 28 , 30 , 31 Similar to existing data, our findings demonstrate a meaningful decrease in postoperative complications with the use of preferred antibiotic prophylaxis regimens—such as cefazolin, cefazolin plus metronidazole, or cefoxitin—compared with non— β -lactam or other regimens. 19 , 20 , 32 Furthermore, recent data suggest that many people with history of allergy to penicillin or β -lactam antibiotics may receive cephalosporins with very low risk for serious allergic reaction. 33 This is particularly notable given that the major postoperative complication rate primarily comprised surgical site infection and sepsis. However, it is important to note that Black patients were more likely to receive preferred antibiotic prophylaxis in this cohort and still had higher rates of surgical site infection and sepsis compared with White patients. Use of hemostatic agents was associated with an independent increase in perioperative complications, even after adjusting for uterine weight or surgical approach. This is consistent with prior studies indicating increase in postoperative pelvic abscesses, readmission, and reoperation with the use of hemostatic agents. 34 It is difficult to assess whether the increased risk may be due to the agent itself or its use for managing blood loss that was greater than expected. However, there are data demonstrating that use of hemostatic agents is strongly driven by surgeon preference and hospital-level factors rather than patient-level factors. 35 We do not intend to suggest that hemostatic agents should not be used when appropriate; rather, their use should be individualized rather than routine. The most critical research implication of this study is that racial disparities in postoperative outcomes persisted in this cohort, even after adjusting for a broad range of potentially modifiable and nonmodifiable risk factors. Additional research is needed to identify pragmatic and actionable strategies that can help to eliminate these persistent disparities. Strengths of this analysis include the use of a statewide database, enabling a large sample of validated clinical data from nearly 70 hospitals, including variation in geographical location (urban vs rural), hospital size, and teaching status, which contributes to the generalizability of the study. Limitations include the observational design and likely confounding factors that could not be assessed in this database. Many of the examined factors were strongly correlated, such as EBL and uterine size, and we had to decide which collinear factors to include in the model on the basis of clinical assumptions in addition to statistical considerations. Additional patient-level factors that could be related to risk for major postoperative complications or differential access to care, such as education level or socioeconomic status, are not collected in this database. Another important limitation is that data collected after December 2018 could not be included because of changes in variables collected by the MSQC database, preventing presentation of more recent data from the past 7 years. It is important to emphasize that assessment of surgeon volume is inherently limited by the sampling methodology used in this database. Because the MSQC database uses sampling rather than complete case capture, total annual case volume per surgeon cannot be determined. Therefore, we categorized surgeon volume by tertiles according to cases in the database that were attributed to an individual surgeon per year, which is intended to serve as a proxy for total case volume. In MSQC, categorization by comparative volume tertiles has been associated with quality and safety metrics such as lower rates of conversion to laparotomy and higher rates of minimally invasive hysterectomy for large uteri. 21 , 22 Maximizing the use of the MIS approach whenever appropriate, increasing access to high-volume gynecologic surgeons, using preferred antibiotic prophylaxis, and minimizing operative time may help reduce racially disparate outcomes following hysterectomy. However, Black patients continue to experience significantly greater risk for postoperative complications following hysterectomy, even after adjusting for patient-level risk factors and clinical and surgical best practices. It is essential to continue to explore factors that contribute to racial disparities in gynecologic care.

This was a retrospective cohort study of patients who underwent hysterectomy for benign indications between January 2015 and December 2018 using the Michigan Surgical Quality Collaborative (MSQC) database. MSQC is a collaboration of 69 hospitals and is funded by Blue Cross Blue Shield of Michigan/Blue Care Network. Participating hospitals have trained nurse data extractors who collect information from medical records using a standardized protocol. Data collected include patient demographic and clinical characteristics, intraoperative data, and 30-day postoperative outcomes. Each month, cases of hysterectomy performed at each site during rotating 8-day cycles are randomly selected for sampling and included in the database for abstraction. We did not include data from hysterectomies performed after 2018 because the collaborative made substantive changes to clinical variables that were abstracted after this time, a number of which we considered to be important modifiable factors that we wished to evaluate in this analysis. Hysterectomy cases were included in this analysis if surgery was performed for benign, nonobstetrical indications and if patients identified their race as White or Black. Patient race was self-reported and obtained from medical record review. We limited this analysis to patients who reported either White or Black race because patients reporting Asian, Hawaiian/Pacific Islander, Native American/Alaskan, or Hispanic ethnicity individually comprised <2% of the study sample. The primary outcome was major postoperative complications, defined as a composite of any of the following events: surgical site infection (deep incision/organ space), sepsis, septic shock, major cardiovascular complications, pneumonia, and venous thromboembolism (VTE). Secondary outcomes included presentation to the emergency department or urgent care, return to the operating room, blood transfusions within the first 72 hours postoperatively, and hospital readmission. Rates of bladder/ureteral injury, bowel injury, and vaginal cuff dehiscence were very low (<0.01% of cases). Therefore, we reported rates of return to the operating room because this composite variable was considered to capture most of these rare complications. We examined a number of risk factors related to major complications. The primary predictor of interest was patient-reported race. Additional patient-level factors included demographics (age, insurance type) and clinical characteristics (body mass index [BMI], American Society of Anesthesiologists [ASA] class, preoperative anemia, coronary artery disease, diabetes, obstructive sleep apnea, hypertension, tobacco use, history of VTE). We aimed to assess the most comprehensive set of clinical and surgical best practices feasible using a large multisystem database. We were able to evaluate a wide range of preoperative clinical best practices, including preoperative workup (documentation of pelvic imaging, Papanicolaou test, and endometrial biopsy) and documentation of prior trials for symptom management (hormonal suppression, progestin intrauterine device, myomectomy, uterine artery embolization, endometrial ablation). For perioperative best practices, we were able to evaluate utilization of surgical skin preparation, mechanical VTE prophylaxis, intraoperative warming, and same-day discharge. We classified cefazolin, cefazolin plus metronidazole, or cefoxitin preoperative antibiotic regimens as “preferred” on the basis of the American College of Obstetricians and Gynecologists recommendations and recent data indicating superior prophylaxis. 18 – 20 We assessed surgical factors including operative time, minimally invasive surgical (MIS) approach (including vaginal, traditional laparoscopy, and robotic-assisted laparoscopy), conversion from minimally invasive surgery to laparotomy, uterine weight, estimated blood loss (EBL), and use of hemostatic agents. Although most variables had complete data, several variables had missingness >10%, which is noted in the tables. Regarding hospital and surgeon characteristics, we were able to evaluate hospital teaching status, hospital bed size, same-day discharge, and surgeon volume. Surgeon volume was assessed by the number of hysterectomy cases in the database attributed to each surgeon per year. Individual surgeons were categorized into tertiles of high, intermediate, and low volume based on their relative annual case contribution. Surgeon volume was categorized on an annual basis, independent of categorization in previous years. This approach accounted for variation in surgeon and hospital participation across the study period, annual differences in case sampling, and year-to-year changes in individual surgeon volume. Given that the MSQC uses a sampling methodology and does not capture all surgical procedures performed at each site, absolute case counts were not meaningful. This method of surgeon volume categorization, using proxy estimates, was comparable to that used in previously published studies from the MSQC. 21 , 22 We began with evaluation of the relationship between major postoperative complications and various risk factors. Rates of primary and secondary outcomes were then reported according to patient-reported race. All factors that were significantly associated ( P <.05) with postoperative complications were included in this analysis, with the addition of several factors considered to be clinically relevant. Patient, clinical, surgical, and hospital/surgeon factors related to major postoperative complications were classified as potentially modifiable or nonmodifiable, and these data were presented according to patient-reported race. Wilcoxon rank-sum, Pearson chi-squared, and Fisher exact tests were used as indicated for the above analyses. The relationship between selected risk factors and major postoperative complications was then evaluated using univariable logistic regression. We evaluated the independent relationship between various risk factors and major postoperative complications using multivariable logistic regression. Selection of covariates was based on factors that were significant ( P <.05) in univariable regression analysis, with particular emphasis on modifiable factors. We performed backwards elimination in a stepwise manner using a significance threshold of P <.05, evaluated for collinearity or interactions, and assessed model goodness of fit using the Hosmer—Lemeshow test. All analyses were performed using R, version 4.3.0 (R Core Team, Vienna, Austria). Reporting is consistent with STROBE (STrengthening the Reporting of OBservational studies in Epidemiology) guidelines. This study met our institutional review board exemption criteria because it used deidentified data.

A total of 18,395 patients underwent hysterectomy for benign indications from January 2015 to December 2018 and met the inclusion criteria for the study. A total of 303 (1.6%) patients experienced a major postoperative complication. Patient demographics, clinical characteristics, preoperative clinical best practices, perioperative characteristics, and hospital and surgeon characteristics were compared according to whether patients had experienced a major postoperative complication ( Table 1 ). Patient-level factors associated with major complications included Black race ( P <.001), public insurance ( P <.001), obesity ( P <.001), ASA class ≥3 ( P <.001), preoperative anemia ( P =.003), diabetes ( P =.008), hypertension ( P =.001), and history of VTE ( P <.001). Patients who had major postoperative complications were less likely to have received preferred antibiotic regimens ( P <.001) or to have had same-day discharge ( P <.001). Major postoperative complications were associated with longer operative time, higher uterine weight, laparotomy approach, EBL ≥400 mL, use of hemostatic agents, and low-volume surgeons (all P <.001). Rates of major postoperative complications were higher among Black patients (n=90; 2.8%) than White patients (n=213; 1.4%; P <.001) ( Table 2 ). Black patients also had higher rates of secondary outcomes, including presentation to the emergency department and urgent care (11.2% vs 9.6%; P =.005), blood transfusion (3.7% vs 0.9%; P <.001), and readmission (4.8% vs 2.7%; P <.001). We classified factors as potentially modifiable or nonmodifiable and evaluated differences in these risk factors between Black and White patients ( Table 3 ). Regarding potentially modifiable factors, Black patients were less likely to undergo a minimally invasive surgical approach (57.2% vs 83.7%; P <.001) or have a high-volume surgeon (59.5% vs 73.6%; P <.001). Black patients were more likely to undergo conversion to laparotomy (3.7% vs 2.0%; P <.001), have hemostatic agents used (39.8% vs 28.3%; P <.001), have preoperative anemia (40.7% vs 13.0%; P <.001), or have EBL ≥400 mL (15.9% vs 5.2%; P <.001). However, Black patients had higher rates of documented preoperative imaging (83.3% vs 72.9%; P <.001) and use of preferred antibiotics (84.8% vs 80.5%; P <.001). Among nonmodifiable risk factors, Black patients had higher rates of obesity and morbid obesity, ASA class ≥3, diabetes, obstructive sleep apnea, hypertension, prior myomectomy, and higher uterine weight (all P <.001). In unadjusted analysis, Black race was associated with a 2-fold increase in risk for major postoperative complications (unadjusted odds ratio [OR], 2.01; 95% confidence interval [CI], 1.56—2.57; P <.001) ( Table 4 ). Additional factors associated with major postoperative complications in univariable logistic regression analysis included public insurance, morbid obesity, ASA class ≥3, preoperative anemia, diabetes, uterine weight ≥1000 g, EBL ≥400 mL, longer operative time, laparotomy, use of hemostatic agents, use of nonpreferred preoperative antibiotic regimens, and intermediate or low surgeon volume tertile. Black race remained independently associated with higher risk for major postoperative complications after adjusting for additional risk factors in multivariable logistic regression (adjusted OR [aOR], 1.39; 95% CI, 1.04—1.85; P =.026) ( Table 5 ). Potentially modifiable risk factors that remained independently associated with higher risk for major postoperative complications included operative time (aOR, 1.13; 95% CI, 1.01—1.25; P =.033), laparotomy surgical approach (aOR, 1.39; 95% CI, 1.03—1.84; P =.026), use of hemostatic agents (aOR, 1.55; 95% CI, 1.22—1.96; P <.001), use of nonpreferred preoperative antibiotic regimen (aOR, 1.50; 95% CI, 1.15—1.94; P =.002), and low surgeon volume tertile (aOR, 1.45; 95% CI, 1.00—2.04; P =.041). We performed a sensitivity analysis to further explore how surgical history might impact the association between race and postoperative complications. Of note, rates of prior cesarean delivery and prior myomectomy did not differ between those who experienced postoperative complications and those who did not in bivariate analysis ( Table 1 ). However, prior myomectomy was more common among Black patients than White patients in bivariate analysis ( Table 3 ). There was high missingness in surgical history variables in this database, including 12% missingness for prior myomectomy, and sensitivity analysis was limited to patients with complete data. Black race remained independently associated with higher risk for major postoperative complications after the addition of prior myomectomy to the multivariable logistic regression model (aOR, 1.51; 95% CI, 1.02—2.22; P =.037) ( Supplemental Table ).

Racial disparities in healthcare are pervasive and have been extensively documented within nearly every medical specialty, including obstetrics and gynecology. In obstetrics, Black patients experience disproportionally worse outcomes regarding preterm birth, 1 preeclampsia, 2 pain management, and complications following cesarean delivery. 3 , 4 In gynecology, Black patients experience disproportionally worse outcomes in diagnosis and surgical management of endometriosis, 5 , 6 uterine fibroids, 7 , 8 ectopic pregnancy, 9 prolapse repair, 10 endometrial cancer, 11 and cervical cancer. 12 Among patients undergoing hysterectomy for benign indications, Black patients experience higher rates of perioperative complications across both abdominal and minimally invasive surgical routes, and this disparity persists after adjusting for factors such as uterine weight, medical comorbidities, and other patient-level perioperative risk factors. 8 , 13 , 14 Existing literature related to disparities in benign hysterectomy outcomes has primarily focused on the persistence of disparate outcomes after adjusting for patient-level perioperative risk factors. However, literature regarding strategies to reduce racially disparate outcomes is scant. Given the persistence of racial disparities after accounting for these risk factors, it is essential to identify potentially modifiable surgeon- and hospital-level factors that may be contributing to the disproportionate morbidity experienced by Black patients. Maximizing utilization of the minimally invasive surgical route has been proposed as a potential strategy to mitigate racially disparate perioperative outcomes. 15 Surgeon volume and hospital volume have been proposed as potentially modifiable factors. 16 , 17 However, disparities in the application of other perioperative clinical and surgical best practices have not yet been explored. The objective of this study was to explore factors associated with racial disparities in major complications following hysterectomy for benign indications, with attention to potentially modifiable clinical and surgical practices that may represent targets for reduction of disparate postoperative outcomes. We were particularly interested in exploring the potential impact of surgeon- and hospital-level decisions, including the use of perioperative medications and surgical techniques.