LAUSR

The validity of the standardized transfusion ratio, a quality measure for dialysis facilities, may have been affected by the transition from International Classification of Diseases, Ninth Revision (ICD-9) to International Classification of Diseases, Tenth Revision (ICD-10) procedure coding in October 2015. We analyzed Medicare Part A claims for inpatient care among dialysis patients in 2014-2017 and investigated billing patterns for blood transfusion during the last year of ICD-9 coding and the first and second years of ICD-10 coding. We identified 2205 hospitals with a steady volume of dialysis patient admissions. In nearly one-third (31.7%) of hospitals, the apparent incidence of blood transfusion during hospitalization fell >50% between the last year of ICD-9 coding and the first year of ICD-10 coding. Between the first and second years of ICD-10 coding, the apparent incidence of blood transfusion during hospitalization fell >20% in 24.5% of hospitals and rose >25% in 14.8% of hospitals. Furthermore, hospital-specific changes in the apparent incidence of blood transfusion among dialysis patients and all Medicare beneficiaries were highly correlated. These findings suggest that the standardized transfusion ratio reflects differential misclassification of transfusions among hospitals. Alternative measures to judge the quality of anemia management, such as attainment of hemoglobin within a target range, may be more appropriate.