Dialysis Ratings Shift When Race Leaves Mortality Formula
A Brown University analysis in JAMA Internal Medicine found that removing race and ethnicity from Medicare's dialysis mortality formula moved 27 percent of centers serving the most Black patients into a better rating category, with no change in care.
Edited and fact-checked by Stephen Sprague
Medicare's published dialysis ratings will show the formula change before any center changes its care, and the study tells patients how to read that jump.
A dialysis center can post a better mortality rating next year with the same staff, the same machines and the same patients. A Brown University analysis published in JAMA Internal Medicine on Wednesday, November 18, shows the arithmetic behind that outcome, because the researchers simulated what happens to Medicare's standardized mortality ratings when race and ethnicity come out of the formula.
The Centers for Medicare and Medicaid Services stopped collecting race and ethnicity in July 2025 across several federal forms, including the intake form completed when a patient starts treatment for kidney failure. Those fields had fed the risk adjustment Medicare applies before it judges a center on patient deaths, so the removal reaches every rating the program publishes from that data.
Working from 6,895 dialysis centers nationwide and records from 2017 to 2020, the team rated each center twice, once with the variables and once without them. Twenty-seven percent of centers serving the highest proportions of Black patients moved into a better-performing category in the simulation, nearly 22 percent of centers serving the most socially disadvantaged patients improved, and about 15 percent of centers serving fewer disadvantaged patients moved down.
"A patient might see these new mortality ratings and make their referral decision based on that, not knowing that the more favorable rating is due to a modeling change," said lead author Kalli Koukounas Green, a research scientist in Brown's Department of Health Services, Policy and Practice. Referring doctors read the same ratings, and facilities use them to decide where quality-improvement money goes, so a modeling shift travels quickly into real decisions.
Senior author Dr. Amal Trivedi put the wider point plainly for federal policymakers, saying that when the government changes health data collection, it changes the country's ability to understand public health trends and measure clinical performance. In the simulation, the ratings gap between centers serving disadvantaged patients and the rest nearly disappeared once the two variables left the calculation.
Every number in the study comes from a simulation on older records, because Medicare's new practices have not yet reached the published ratings and the size of the real-world shift is still unknown. The researchers say the next step is to measure the published ratings themselves once the change flows through, and dialysis patients choosing a center in the meantime are reading scores built on a formula that is already being retired.
The conversation.
Loading comments…
A little more perspective.
Reading is free. Verified members can comment and reply for $2 a month.