Exploring racial disparities in treatment patterns and outcomes for patients with multiple myeloma using real world data.

This retrospective observational study evaluated racial disparities among Black and White patients with multiple myeloma (MM). We included patients from a longitudinal de-identified EHR-derived database who had ≥2 visits recorded on or after 1/1/2011, documented treatment, and race listed as White or Black. Black patients (n = 1172) were more likely female (54.8%/42.9%) and younger (<65 years, 40.8%/30.8%) than White patients (n = 4637). Unadjusted median real-world overall survival (rwOS) indexed to first-line of therapy (LOT) was 64.6 months (95% CI: 57.8-74.0) for Blacks and 54.5 months (95% CI: 50.9-56.2) for Whites. Adjusted rwOS estimates (for sex, age at index date, and practice type) to either first- (aHR = 0.94; 95% CI: 0.84-1.06) or second-LOT (aHR = 0.90; 95% CI: 0.77-1.05) were similar. Unadjusted derived response rate (dRR) during first-LOT was 84.8% (95% CI: 80.7-88.1) for Blacks and 86.9% (95% CI: 85.0-88.5) for Whites (odds ratio [OR] = 0.78 [95% CI: 0.57-1.10]); in second-LOT, 67.2% (95% CI: 58.4-75.0) for Blacks and 72.4% (95% CI: 68.1-76.3) for Whites (OR = 0.72 [95% CI: 0.46-1.13]). High representation of Black patients enabled this robust analysis, albeit with limitations inherent to the observational data source, the retrospective design, and the analytic use of newly derived endpoints requiring further validation.

Ensemble machine learning for the prediction of patient-level outcomes following thyroidectomy.

BACKGROUND: Accurate prediction of thyroidectomy complications is necessary to inform treatment decisions. Ensemble machine learning provides one approach to improve prediction. METHODS: We applied the Super Learner (SL) algorithm to the 2016-2018 thyroidectomy-specific NSQIP database to predict complications following thyroidectomy. Cross-validation was used to assess model discrimination and precision. RESULTS: For the 17,987 patients undergoing thyroidectomy, rates of recurrent laryngeal nerve injury, post-operative hypocalcemia prior to discharge or within 30 days, and neck hematoma were 6.1%, 6.4%, 9.0%, and 1.8%, respectively. SL improved prediction of thyroidectomy-specific outcomes when compared with benchmark logistic regression approaches. For postoperative hypocalcemia prior to discharge, SL improved the cross-validated AUROC to 0.72 (95%CI 0.70-0.74) compared to 0.70 (95%CI 0.68-0.72; p < 0.001) when using a manually curated logistic regression algorithm. CONCLUSION: Ensemble machine learning modestly improves prediction for thyroidectomy-specific outcomes. SL holds promise to provide more accurate patient-level risk prediction to inform treatment decisions.