The Organ Procurement and Transplantation Network (OPTN) collects intermittent survey data on immunosuppressive medication use that are studied frequently as research measures. Pharmacy billing claims may provide an accurate measure of immunosuppression use over time. Herein is characterized the agreement of Medicare pharmacy claims for immunosuppressive medications with OPTN reports. Data were drawn from the United States Renal Data System. Participants received a kidney transplant in 2000 to 2001 and had an OPTN record and a Medicare pharmacy claim for an immunosuppressive drug at transplant discharge and 6 mo and 1 yr after transplantation. The concordance (kappa) of the OPTN and claims (+/-30 d of survey) for indicated medication use was compared, and sensitivity, specificity, and predictive values for claims were computed, assuming OPTN as a "gold standard." Clinical trial participation and regimen changes were examined as explanations for discordance. A total of 4357 eligible subjects were identified. Concordance over observation ranged from excellent for calcineurin inhibitors (kappa > 0.86) to generally very good for adjunctive agents (kappa = 0.49 to 0.75) to poor for corticosteroids (kappa <0.15). Claims demonstrated high positive predictive values (> or =97%) but low negative predictive values (< or =13%) for OPTN-reported corticosteroid use. Regimen changes (28 to 75%) but not clinical trial participation (< or =21%) were identified frequently among cases with discordant indications of nonsteroid medication use. Close agreement of Medicare billing claims and the OPTN for indicated use of nonsteroid immunosuppressive medications supports both as useful measures of drug exposure. Low detection rates of OPTN-indicated corticosteroid use within claims require further examination.