Background and objectives: Accurate assessment of the use of immunosuppressive medications is vital for observational analyses that are widely used in transplantation research. This study assessed the accuracy of three potential sources of maintenance immunosuppression data.
Design, setting, participants, & measurements: This study investigated the agreement of immunosuppression information in directly linked electronic medical records for Medicare beneficiaries who received a kidney transplant at one center in 1998 through 2001, Organ Procurement and Transplantation Network (OPTN) survey data, and Medicare pharmacy claims. Pair-wise, interdata concordance (kappa) and percentage agreement statistics were used to compare immunosuppression regimens reported at discharge, and at 6 mo and 1 yr after transplantation in each data source.
Results: Among 181 eligible participants, agreement between data sources for nonsteroid immunosuppression increased with time after transplantation. By 1-yr, concordance was excellent for calcineurin inhibitors and mycophenolate mofetil (kappa = 0.79 to 1.00), and very good for azathioprine (kappa = 0.73 to 0.85). Similarly, percentage agreement at 1 yr was 94.9 to 100% for calcineurin inhibitors, 91.1 to 95.7% for mycophenolate mofetil, and 87.5 to 92.8% for azathioprine. Widening the comparison time window resolved 33.6% of cases with discordant indications of calcineurin inhibitor and/or antimetabolite use in claims compared with other data sources.
Conclusions: This analysis supports the accuracy of the three sources of data for description of nonsteroid immunosuppression after kidney transplantation. Given the current strategic focus on reducing collection of data, use of alternative measures of immunosuppression exposure is appropriate and will assume greater importance.