State- and end-stage renal disease-network determinants of peritoneal dialysis utilization in the United States: A multi-level Beta-regression analysis.

BackgroundChronic kidney disease and end-stage renal disease (ESRD) significantly burden the U.S. healthcare system. Despite its benefits, such as cost savings and increased autonomy, peritoneal dialysis (PD) is underutilized. This study examines how state and ESRD network-level variations impact PD utilization across the United States, addressing a gap in previous investigations that have not fully disentangled the effects of measurable patient-level factors from harder-to-capture influences.MethodsWe analyzed publicly available data from the United States Renal Data System, Centers for Medicare and Medicaid Services, and the U.S. Census Bureau covering 50 states and two territories (2015-2020). Multi-level Beta regression models assessed how state-level predictors are associated with PD utilization, accounting for state and network-level variations. Ranked random effects were compared to 2020 PD rates to identify overperforming or underperforming regions.ResultsBetween 2015 and 2020, PD utilization rose from 9.6% to 12.8%. The proportions of incident dialysis patients who were male, ≥ 65 years, and White, respectively, and dialysis facilities per 100,000 people did not significantly affect PD utilization, whereas population density was significantly negatively associated. The final model revealed that unmeasured variations in PD utilization were significantly explained by both state (intraclass correlation coefficient (ICC) = 0.47) and network (ICC = 0.53) factors.ConclusionThe chosen demographic, provider, and geographical factors explain only 18.5% of PD utilization. About half of the remaining variation resides at the state level and half at the ESRD network level, verifying the importance of unmeasured factors at both levels. We provide adjusted PD utilization rankings-identifying overperforming and underperforming states and networks-where future research can identify disparate effective and ineffective regional policies with the aim of optimizing PD uptake.