Sources Yield Differing Data on Comorbidity in Dialysis Patients

For patients starting dialysis, the comorbidity reported on the Medical Evidence Report (MER) often differs from that identified from Medicare claims, reports a study in the American Journal of Kidney Diseases.

The study included 45,357 Medicare-eligible patients starting maintenance dialysis during the second half of 2007, 2008, or 2009. The prevalence of 12 comorbid conditions was assessed from claims during the 6-month period before the index date, the MER, and claims during the 3-month period after the index date. Agreement between these three sources of data was assessed.

The prevalence of comorbidity based on claims during the 6 months before patients started dialysis generally exceeded that based on the MER. Agreement was low to moderate, with κ statistics ranging from 0.07 for drug dependence to 0.69 for diabetes. The conditions with the largest absolute variation were atherosclerotic heart disease, congestive heart failure, chronic obstructive pulmonary disease, other cardiac disease, and peripheral vascular disease. The degree of discordance varied significantly by age, race, sex, and ESRD.

The analysis of 23,930 patient-years of follow-up included 8930 deaths. In predicting risk of death, claims from the 3 months after dialysis initiation outperformed the designations from the MER, with C statistics of 0.674 versus 0.616, respectively. Based on the difference between the MER and claims data, the standardized mortality ratios differed by more than 10 percent at 26.5 percent of dialysis facilities and by more than 20 percent at 12.8 percent of facilities.

The MER is a major source of comorbidity data for risk adjustment of quality metrics for dialysis facilities. The new study shows substantial variations between comorbidity assessed by the MER compared to Medicare claims data before and after initiation of dialysis.

“These patterns may engender bias in risk-adjusted quality metrics,” the researchers write. They suggest that claims made during in the first 3 months after patients start dialysis might be a better source of data on comorbidity [Krishnan M, et al.: Comorbidity ascertainment from the ESRD Medical Evidence Report and Medicare claims around dialysis initiation: a comparison using US Renal Data System. Am J Kidney Dis 2015; 66:802–812].