Do resource utilization and clinical measures still vary across dialysis chains after controlling for the local practices of facilities and physicians?

Background: Because of adverse survival effects, anemia management and financial incentives to increase doses of erythropoiesis-stimulating agents (ESAs) have been controversial. Prior studies showed more aggressive anemia management in dialysis facilities owned by for-profit chains, but have been criticized for not accounting for practices of individual physicians and facilities.

Objective: To improve understanding of how dialysis practices and resource utilization are influenced by physicians, facilities, and chains.

Design: Mixed models with chain fixed effects and facility and physician random effects.

Setting: Medicare hemodialysis patients in 2004.

Participants: A total of 234,158 patients, 3995 facilities, 4838 physicians, and 7 chain classifications were included.

Measurements: Spending per session for dialysis-related services billed separately from the dialysis treatment and for ESAs. Achievement of hematocrit (HCT) and urea reduction ratio (URR) targets.

Results: Of the 4 largest for-profit chains, 3 had higher resource use than independents, with differences up to $17.92 higher ESA/session. Utilization was positively associated with achieving target HCT. Despite incurring lower costs, patients treated by a large nonprofit chain were as likely as patients of independents to achieve the HCT target. The largest chains were more likely than independents to achieve the URR target. Substantial variation occurred across physicians and facilities, and adjustment for chain only modestly decreased this variation.

Limitation: Chains' methods of influencing practices were not directly observed.

Conclusions: Chains appear to have the ability to implement protocols that shift practices, but not the ability to substantially reduce local variation. Assertions that chain effects found by earlier studies were spurious are not supported.