The Role of Surgeon Volume in the Relationship Between Race and Surgical Modality for Hysterectomy for Benign Indications.

The Premier Healthcare Database (PHD) comprises Health Insurance Portability and Accountability Act (HIPAA)–compliant hospital-based, service-level, all-payer information on inpatient discharges and outpatient encounters. 14 PHD includes nearly 9 million inpatient admissions per year, representing approximately 25% of annual U.S. inpatient admissions. Outpatient visits to emergency departments, ambulatory surgery centers, and alternate sites of care are also included, totaling more than 86 million visits per year. 14 This observational retrospective cohort study used aggregated, de-identified patient data; thus, IRB approval and patient consent were not required. Hysterectomy was identified using primary International Classification of Diseases, Tenth Revision, Clinical Modification (ICD-10-CM) procedure codes and Current Procedural Terminology (CPT) codes. We included only patients who did not have an indication for a gynecologic malignancy (see Appendix 1, available online at http://links.lww.com/AOG/E550 , for codes defining hysterectomy and gynecologic malignancy). The cohort consisted of women aged 18 years or older who had hysterectomy for benign indications from 2016 to 2023 identified in the PHD database (n=899,657). Inclusion criteria were 1) a self-reported race of non-Hispanic Black or non-Hispanic White, 2) availability of an ICD-10-CM or CPT hysterectomy procedure code, 3) no indication of gynecologic malignancy, and 4) availability of surgeon volume for the previous year. We excluded individuals with concomitant lymphadenectomy and endometrial hyperplasia as well as duplicate encounters (after retaining the encounter of interest). The final cohort consisted of 367,593 patients (Fig. 1 ). Ryntz. Role of surgeon volume in MIS utilization. Obstet Gynecol 2026 . Surgeon volume was defined as the number of hysterectomy (benign and malignant indications) cases by the operating surgeon in the prior calendar year. Surgeons were classified based on their previous year’s case volume, with fixed threshold cutoffs at the 33rd percentile. This resulted in surgeon volume being classified as low (seven cases or fewer), medium (8–18 cases), and high (19 cases or more). To simplify the analysis, we combined medium and high surgeon volumes to create a binary of low compared with nonlow volume, given that our primary objective was to investigate MIS disparities and that prior literature showed these disparities occurring primarily for low-volume surgeons. The modality of hysterectomy for benign indications was characterized as MIS or abdominal; MIS included laparoscopy, robotic-assisted surgery, and vaginal procedures. We used a tiered system of modality identification, whereby individuals who had an ICD-10-CM procedure or CPT/Healthcare Common Procedure Coding System modifier code for robotic-assisted surgery or a documented charge code for robotic instrumentation were defined as having robotic hysterectomy. Next, individuals who had a laparoscopic modifier code were defined as having laparoscopic hysterectomy, and individuals with a vaginal modifier code were identified as having vaginal hysterectomy. All remaining individuals were identified as having abdominal hysterectomy. See Appendix 2, available online at http://links.lww.com/AOG/E550 , for modality modifier codes. Recommendations on procedure modality can be based on individual and facility-level characteristics; therefore, we adjusted for many of these variables. Demographic characteristics included age (18–34 years, 35–44 years, 45–64 years, 65 years or older), and payer type (commercial, Medicare, Medicaid, other). Complexity indicators included uterine size greater than 250 g, presence of adhesions, body mass index (BMI, calculated as weight in kilograms divided by height in meters squared) 30 or higher, and Charlson Comorbidity Index score (grouped as 0, 1, 2 or higher). 15 We also examined the indications for surgery (endometriosis, leiomyomas, inflammation or infection, uterine bleeding, and chronic pelvic pain) and whether an individual had undergone concomitant procedures (removal of ovary or fallopian tube, repair of pelvic organ prolapse, and endometriosis ablation or excision). See Appendix 3, available online at http://links.lww.com/AOG/E550 , for indication and concomitant codes. Finally, facility characteristics included geographic region (Northeast, West, South, Midwest), location (rural or urban), hospital type (teaching or nonteaching), and number of beds (fewer than 100, 100–299, 300 or more) as a proxy for size. Urban was defined as a location in which core Census blocks have a minimum density of 1,000 people per square mile and are surrounded by Census blocks with a minimum density of 500 people per square mile. 14 Descriptive statistics were used to compare the baseline characteristics of patients referred to low-volume compared with non–low-volume surgeons. Categorical variables were summarized as frequencies and percentages, and continuous variables were presented as means with standard deviations or medians with interquartile ranges, as appropriate. Comparisons between groups were assessed using χ 2 or Fisher's exact tests for categorical variables and t tests or Wilcoxon rank-sum tests for continuous variables. The association between patient race and surgeon volume was calculated using a multivariable logistic regression analysis. Surgeon volume (low vs nonlow) was modeled as the dependent variable, with patient race (non-Hispanic Black vs non-Hispanic White) as the primary independent variable. The association between surgeon volume and surgical modality was also calculated using a multivariable logistic regression analysis. Surgical modality (abdominal vs MIS) was modeled as the dependent variable, with surgeon volume (low vs nonlow) as the primary independent variable. Both crude and adjusted models were estimated. Finally, we examined whether surgeon volume was a significant determinant in the relationship between patient race and surgical modality (see Appendix 4, available online at http://links.lww.com/AOG/E550 , for the directed acyclic graph used to assess this relationship). The mediation analyses used Lange et al's 16 structural equation modeling with path analysis. This method involves the decomposition of the total effect into an indirect (exposure through mediator onto outcome) and direct effect (exposure onto outcome). The sum of the direct and indirect effects (or product of the odds ratios [ORs]) was calculated to estimate the total effect. Percentile bootstrapping was applied to estimate CIs, with 2,000 resamples. The proportion mediated was calculated as the ratio of the indirect effect to the total effect to quantify the contribution of surgeon volume to the race-modality relationship. Additional interactions between exposure and covariates were not assessed. All statistical tests were performed using R Studio 4.4.1 and were two-sided, with a significance level set at α=.05.

Of the 367,593 patients, 71,138 (19.4%) were Black and 296,455 (80.6%) were White. A total of 38,605 patients were treated by low-volume surgeons (5,824 surgeons were classified as low-volume), and 328,988 patients were treated by non–low-volume surgeons (8,600 surgeons were classified as non–low volume). Surgeon volume classification was based on 10,810 unique surgeons contributing data across 29,996 surgeon-years, with an average of 2.8 years of data per surgeon contributed to the total cohort dataset. Among the unique surgeons, 3,755 were always classified as low volume, 4,986 were always classified as non–low volume, and 2,069 were reclassified when case volume thresholds were recalculated annually based on the previous year’s data. The distribution of patients across surgeon volume was significantly different. Low-volume surgeons were more likely to care for patients who were Black, had complications including chronic pelvic pain and uterine bleeding, had leiomyomas as the primary indication for surgery, and were treated at small- to medium-size hospitals (Table 1 ). In the study cohort, 13.9% of Black patients were treated by low-volume surgeons, compared with 9.7% of White patients. After adjusting for confounders, Black patients were 45.7% more likely to be treated by low-volume surgeons than White patients (OR 1.46, 95% CI, 1.42–1.50). See Table 2 for crude and adjusted results. Baseline Patient Characteristics by Surgeon Volume Surgeon Volume by Patient Race The proportion of low-volume surgeons using MIS approaches was lower than that of non–low-volume surgeons (75.7% and 88.0%, respectively). After adjusting for confounders, patients whose hysterectomy for benign indications was done by a low-volume surgeon had 2.2 times higher odds of having an abdominal procedure than patients seen by non–low-volume surgeons (OR 2.24, 95% CI, 2.17–2.30). See Table 3 for crude and adjusted results. Hysterectomy Modality by Surgeon Volume To assess whether surgeon volume was a mediator in the relationship between patient race and surgical modality, we calculated path A (race onto surgeon volume), path B (surgeon volume onto modality), and the product of these paths (ie, the indirect effect). Two models were tested, one with no covariate adjustment and the other with the relevant covariates detailed in the Methods section. The indirect effect analysis revealed a small but significant effect of surgeon volume as an important determinant of the race–modality relationship (0.004, P <.01), with the proportion mediated ranging from 2.7% (adjusted) to 3.1% (crude). The direct effect indicates that minoritized group identity significantly influenced the likelihood of receiving MIS, regardless of any mediators and measured confounders (0.143, P <.01). The results from the total, direct, and indirect analyses are summarized in Appendix 5 (available online at http://links.lww.com/AOG/E550 ) and Table 4 . Mediation of Surgeon Volume on Relationship Between Patient Race and Route of Hysterectomy

In this retrospective study, we tested the relationship between patient race, surgeon volume, and surgical modality to contextualize the disparity in access to minimally invasive hysterectomy for benign indications. We found that Black women had greater odds of being treated by low-volume surgeons who were more likely to perform the procedure abdominally. Our findings are consistent with those of previous studies: Boyd et al found that surgeons who performed at least 10 hysterectomy cases a year had an increased likelihood of using minimally invasive options. 13 Anne et al validated this relationship and found that Black patients were more likely to be treated by low-volume surgeons. 17 Our analysis also indicated that surgeon volume was a significant mediator of the relationship between patient race and modality but contributed to only 2.7% of the observed disparity. Although this has not been formally tested before, associations between surgeon volume and disparities to surgical access and outcomes have been cited numerous times in the literature. 18 , 19 The present findings represent a first step toward characterizing the disparity through a determinant (surgeon volume) that is actionable and (to our knowledge) previously untested. Thus, our study contributes to a deeper understanding of racial disparities in the use of MIS for gynecologic surgery. The clinical and economic benefits of surgery performed by high-volume surgeons in high-volume centers have been demonstrated across multiple disciplines, including gynecology. Hysterectomy, a relatively safe and low-mortality procedure (6 deaths/10,000 patients) has demonstrated similar benefits, though to a somewhat lesser extent. 20 For vaginal hysterectomy, the benefits of reduced morbidity and resource utilization are limited to high-volume surgeons exclusively performing the procedure of interest. 21 Similar benefits—both statistically significant but clinically modest—have been shown for patients undergoing laparoscopic hysterectomy by high-volume compared with low-volume surgeons. 22 Boyd et al observed similar benefits in women undergoing hysterectomy of any modality by high-volume compared with low-volume surgeons in New York State. 17 Their results showed that high-volume surgeons were more likely to perform MIS, conferring some of its inherent benefits in addition to improved perioperative management. Our study builds on these demonstrated benefits by showing a small but statistically significant effect of surgeon volume on surgical modality when considering Black women specifically. Although the effect is small, it still must be acknowledged that surgeon volume has a measurably significant effect on the likelihood of a minimally invasive hysterectomy in a group of women who have otherwise experienced lower rates of this safer approach. The strengths of our approach include the use of a large, geographically diverse database; robust mediation technique; and the characterization of the mediator as time varying. We used structural equation modeling, which allowed us to calculate the direct, indirect, and total effects and to test whether each one is a significant pathway simultaneously. Previous studies on disparities in surgical access and outcomes have used this methodology to test various determinants in cardiac and oncologic procedures. 23 – 25 Our methodologic approach allowed for a more realistic depiction of the complexity and interrelatedness of the disparity, including the use of robust standard errors and bootstrapping to improve reliability of estimates. Additionally, surgeon volume was calculated as total hysterectomy cases performed in the previous calendar year and categorized by tertiles, allowing for surgeons to move across volume classifications (low medium, high) over time. The use of previous year’s case volume and tertile thresholds were consistent with existing literature for hysterectomy. 9 , 22 , 26 Modeling surgeon volume as dynamic allows for accounting of latent variables that evolve over time, such as surgeon experience or training. Dynamic modeling also allows for contextual variation, such as referral network changes or institutional investments in robotic platforms, to be absorbed into the model. Moreover, analysis of surgeon volume allowed us to improve temporal accuracy by calculating physician case volume relative to a patient's surgery date, which preserved time-ordering in the causal inference. However, the study is limited in its race categorization, the collapsing of surgeon volume into a binary variable, and residual confounding. Our analysis was limited to Black and White patients because of the limited reporting of race and ethnicity categories within the claims database and their respective sample size constraints, thereby limiting generalization to other patient groups. We operationalized the three categories of surgeon volume as a binary outcome (low vs nonlow volume) to simplify the analysis, but this may have reduced statistical power and obscured a potential gradient of effect (ie, a dose-effect relationship). Variables such as patient preference, years of surgeon experience, and institutional MIS culture were untested in our analysis and may have contributed to unmeasured confounding that biases our estimates toward the null. These limitations affect the external validity of our study. However, in the absence of many surgeon- and facility-level variables, we used surgeon volume as an imperfect proxy measure for surgical skill and MIS capability. In this study, the analysis treats surgeon volume as a mediator instead of a covariate, which emphasizes the complexity of equitable health care delivery in issues of both quality and access. The analysis demonstrates the role that surgeon volume plays as both an institutional determinant of health and an aspect of the quality of patient care. Given this important role, interventions that support marginalized patients’ access to high-volume surgeons should be prioritized. These may include strengthening referral pathways to hospitals with MIS expertise, increasing MIS training geographically to reduce “MIS deserts,” 27 and enacting policies that incentivize resource allocation in hospitals disproportionately serving marginalized patients. These interventions could help reduce the inequity in MIS utilization by increasing pathways for racially diverse patients to access high-volume health care professionals.