Practice-Level Severity Case Mix and Treatment Patterns for Premenopausal Noncancerous Hysterectomy.

Abstract

Background: Hysterectomy for noncancerous conditions is a patient-preference-sensitive procedure. Therefore, gynecological practices may provide hysterectomy at varying levels of symptom severity. We assess whether practice-level severity case mix associates with segregation of patients by race and ethnicity or insurance status. Methods: In this case series, we analyzed electronic health records of 1,590 noncancerous hysterectomy patients across 20 clinical practices within a large health care system in the U.S. South (2014-2017). By abstracting 12-month presurgical medical notes, we developed severity scores for bleeding, pain, and bulk symptoms. The practice-level severity case mix measure distinguished six practices where ≥18% of patients had below median scores for bleeding, pain, and bulk. Log-binomial models estimated prevalence ratios (PRs) for severity case mix by race and ethnicity and insurance, adjusting for age, body mass index, gynecological conditions, previous abdominal surgeries, and prior uterine sparing treatments. Results: Patients at practices with lower severity case mix differed in surgical indications, had fewer uterine-sparing treatments before undergoing hysterectomy, and were largely (96%) privately insured. Compared to White patients, Hispanic patients underwent hysterectomy less frequently at lower severity practices (PR: 0.52 [0.33-0.82]) while Black patients showed no difference based on the point estimate (PR: 1.00 [0.87-1.14]). Publicly-insured and uninsured patients were less likely than privately-insured patients to receive hysterectomy at lower severity practices (PR: 0.13 [0.05-0.36] and PR: 0.28 [0.12-0.68], respectively). Conclusions: Publicly insured and uninsured patients receiving hysterectomy-including nearly all Hispanic patients-were concentrated in practices with a higher symptom severity case mix.