Quantifying Regional and Health Care Variations to Identify Ways to Improve Hemodialysis Service Quality and Survival Outcomes

The authors examined variations in hemodialysis care and quantified the effect of these variations on all-cause mortality. Insurance claims data from April 1, 2017 to March 30, 2018 were reviewed. In total, 2895 hospital patients were identified, among whom 398 died from various causes. Controlling effects of the facility and secondary medical care areas, all-cause mortality was associated with older age, heart failure, malignancy, cerebral stroke, severe comorbidity, and the first and ninth centile of physician density. Multilevel analyses indicated a significant variation at facility level (σ22 0.27, 95% confidence interval: 0.09-0.49). Inclusion of all covariates in the final model significantly reduced facility-level variance. Physician density emerged as an important factor affecting survival outcome; thus, a review of workforce and resource allocation policies is needed. Better clinical management and standardized work processes are necessary to attenuate differences in hospital practice patterns.