Abstract
A number of data sets can be used to estimate the size of the nursing home population that has mental illness; however, estimates vary because of differences in methods of data collection. This study sought to compare estimates from three nationally representative data sets of the number of nursing home residents who have a mental illness, determine which data set provides the best national-level estimate, and identify the types of policy and monitoring questions that can best be answered with each.
The study compared estimates of the number of nursing home residents who had either a primary or any diagnosed mental illness from the National Nursing Home Survey (NNHS), the Minimum Data Set (MDS), and the Medicaid Analytic eXtract (MAX) files.
The NNHS produced the most valid national-level estimates of residents with a mental illness--nearly 102,000 with a primary diagnosis in 2004 (6.8% of residents), of which about 23,000 were under age 65 and 79,000 were aged 65 and older. However, data from the NNHS cannot be broken down to the state level; therefore, state- and facility-level estimates would have to be generated with the MDS or MAX data sets.
Policy makers and program managers need to be aware of the strengths and limitations of the data they use in order to make informed decisions. Users of the NNHS, MDS, and MAX data sets should be aware of the differences in recorded diagnoses among the three, especially the relatively limited diagnoses in the MAX and imprecise diagnoses in the MDS.