|
Abstract:
|
Assessment of patients' functional status across the continuum of care requires a common instrument. Different health care settings rely on different instruments. For example, the Functional Independence Measure (FIM) is used to evaluate the functional status of patients who stay in inpatient rehabilitation facilities (IRFs). After discharge from IRFs, the Minimum Data Set (MDS) is collected for patients that are transferred to skilled nursing facilities (SNFs), while the Outcome and Assessment Information Set (OASIS) is collected for patients receiving home health care. To compare patients that are discharged from IRFs to either SNFs or home health, a single measure of functionality is required. Assuming that all the patients have observed FIM measurements and treating the unmeasured MDS or OASIS items as missing, we propose a variant of the predictive mean matching method, which relies on Item Response Theory to impute the unobserved functionality items. Using simulations, we compared the proposed approach to existing methods, and showed that it performs well for estimating the overall functionality status, while preserving the correlation structures among functionality items.
|
