To assess the feasibility of using existing claims‐based algorithms to identify community‐dwelling Medicare beneficiaries with disability based solely on the conditions for which they are being treated, and improving on these algorithms by combining them in predictive models.
Data on 12,415 community‐dwelling fee‐for‐service Medicare beneficiaries who first responded to the Medicare Current Beneficiary Survey () in 2003–2006.
Logistic regression models in which six claims‐based disability indicators are used to predict self‐reported disability. Receiver operating characteristic () curves were used to assess the performance of the predictive models.
The predictive performance of the regression‐based models is better than that of the individual claims‐based indicators. At a predicted probability threshold chosen to maximize the sum of sensitivity and specificity, sensitivity is 0.72 for beneficiaries age 65 or older and specificity is 0.65. For those under 65, sensitivity is 0.54 and specificity is 0.67. The findings also suggest ways to improve predictive performance for specific disability populations of interest to researchers.
Predictive models that incorporate multiple claims‐based indicators provide an improved tool for researchers seeking to identify people with disabilities in claims data.