Analysis of the compassionate allowance (CAL) program: A systematic data-driven approach to identifying potential CAL conditions

Each year over three million individuals apply for Social Security Disability Insurance (DI) and Supplemental Security Income (SSI) disability benefits. Over the years, SSA has instituted procedures to increase the speed of claims processing. Most recently, Compassionate Allowances (CAL) was implemented to identify the most obvious cases for allowances based on minimal objective medical information that could be obtained quickly. Periodically, SSA expands its list of CAL conditions. Through an interagency agreement supporting SSA's efforts, NIH developed a systematic, data-driven approach to identify potential conditions for inclusion on the CAL list similar to current CAL conditions based on survival times using SSA Disability Research Files (1997 - 2006). SSA data linked to the National Health Interview Survey and the National Death Index for the same timeframe were used for benchmarking. We developed survival curves for the existing CAL conditions that could be identified in the data sets. By viewing these survival curves as elements of a Skorokhod space, we constructed a test region whose boundary was defined by these elements. We used a Kolmogorov-Smirnov test statistic to determine whether survival curves from non-CAL condition lie within this region. Using this method, we identified 25 potential CAL conditions whose survival profiles were comparable to those of current CAL conditions. These conditions included cancers, severe forms of liver disease and HIV infection, and cerebral trauma. Although data limitations presented substantial analytic challenges, we demonstrated that SSA data can be used to augment SSA's programmatic decision making processes using innovative analytic approaches.