Enhancing predictive accuracy: Assessing the laparoscopic hysterectomy readmission score in a diverse endometrial...

The laparoscopic hysterectomy readmission score (LHRS) was developed to identify patients at low risk for readmission following laparoscopic surgery. Its performance in patients with endometrial cancer (EC), a population increasingly undergoing laparoscopic hysterectomy, remains unassessed, particularly across racial and ethnic groups. This study evaluates LHRS discrimination and calibration in predicting readmission among patients with EC overall and by race and ethnicity.

Using the National Surgical Quality Improvement Program database (2014-2020), we identified patients who underwent laparoscopic hysterectomy. The LHRS was calculated as previously described by Jennings et al. Model performance was assessed using the area under the receiver operating curve (AUC), true positive rate (TPR) and false positive rate (FPR). Logistic regression estimated 30-day readmission odds in the overall cohort and by race and ethnicity. The LHRS was analyzed as a continuous (0-8) and dichotomous (<3 vs. ≥3) variable.

Among 27,981 patients with EC (Asian = 1367; American Indian/Alaska Native = 115; Black = 2120; Latine = 1676; Native Hawaiian/Pacific Islander = 305; White = 22,398), readmission rates were low (3.1 %). The LHRS demonstrated poor discrimination, with an AUC only marginally better than chance overall (0.52, 95 % CI = 0.51, 0.53) across all racial and ethnic groups. TPRs and FPRs were uniformly low. Patients with LHRS ≥3 has higher readmission odds (OR = 3.10, 95 % CI = 2.31, 4.15) with similar associations in race and ethnicity-stratified models.

Although a higher LHRS was associated with increased readmission odds, the score failed to predict readmission better than chance in this EC cohort, underscoring the need for improved risk stratification tools for readmission in this population.