Derivation and properties of a brief health status assessment instrument for use in HIV disease

Abstract

Health status measures adapted from the Medical Outcomes Study (MOS) scales have been shown to be useful and highly reliable in human immunodeficiency virus-infected populations, but acceptance of these measures has been limited in part by concerns over investigator, patient, and data burden. We sought to address these concerns by reducing the number of items in the MOS scales rather than by reducing the number of domains covered. We selected items for the shorter scales based on the static and dynamic relationships to the longer scales and to indicators of clinical and functional status in 10,399 responses from 1,934 participants, as well as in relevant subpopulations. The resulting scales have a total of 21 items, as compared to 38 in the longer instrument. Correlations between the shorter and longer scales were excellent. At 0.78 to 0.85, the reliability of the subscales was lower than that of the full scales. However, compared to an index based on the longer scales, an index based on the subscales had identical reliability, yielded nearly identical values, and was as sensitive in detecting treatment differences in a clinical trial. The resulting instrument continues to cover disability, work, utilization, and health status, but is less than half the length of our previous comprehensive questionnaires.