The proposed Medicare prospective payment system for dialysis includes 18 case mix adjusters, 17 of which only increase payment.  CMS calculated these adjusters using several data sources, including Medicare paid claims, and reduced the base payment 22% to offset the associated incremental costs.  OBJECTIVES: Determine the ability of a large dialysis organization to detect CMAs using all available data sources, and calculate the financial impact of differences in CMAs ascertained by unit data, versus those reported by CMS. METHODS: Four dialysis units were randomly selected in each of 20 geographic regions.  CMAs were ascertained via chart reviews, electronic medical records, hospital discharge summaries, paper charts, healthcare professional notes, and discussions with on-site healthcare professionals.   RESULTS: The final sample included 100 clinics, with an average of 73.4 patients per unit.  CMS provided estimated CMAs for 89 of these units.  In 75 of 89 clinics, the case mix adjuster detected in the study was lower than that ascertained by CMS (mean difference=0.09).  In 14 units, the CMA was higher than reported by CMS (mean difference=0.04).  The average CMA for the 89 units was 1.21 versus 1.28 reported by CMS for these same units, with uneven geographic distribution on the difference.  The inability to replicate CMS CMAs would result in a 7% decrease in payment, with a differential from CMS projected payments of >$350 million over a 4 year period. CONCLUSIONS: Without access to Medicare claims data, we were unable to replicate CMS CMAs, representing the potential for significant underpayment for dialysis units under the proposed prospective payment system.