Predictive power of a risk-assessment questionnaire across different disease states: results in an elderly managed care enrolled population

Abstract

Objective: To examine the predictive power of self-reported assessment questionnaire data to explain changes in healthcare service utilization and expenditures of a population of Medicare health maintenance organization enrollees during a 2-year period.

Study design: A prospective cohort study with a 2-year postenrollment follow-up period. Multiple robust regression analyses were conducted to examine associations among self-reported health status variables obtained from responses to the questionnaire.

Sample and methods: Participants were administered a voluntary comprehensive questionnaire at enrollment that collected self-reported information on morbidity, health status, perceptions of health, and healthcare service utilization during the preenrollment year. Questionnaire responses were combined with actual 2-year postenrollment claims data. For the complete follow-up period, 4128 patients were available.

Results: Participants with such chronic conditions as depression and diabetes were likelier than the average enrollee to have higher healthcare service utilization. Self-reported health status predictors examined in this study explained a larger percentage of the variance (as much as 20%) in such chronic conditions as cancer and depression. Despite evidence of underreporting of preenrollment healthcare service utilization, these variables were highly predictive of actual postenrollment utilization patterns.

Conclusions: Self-reported health status information collected at baseline is as predictive of postenrollment risk as are currently used traditional approaches that require archival healthcare service utilization data. In addition, this approach is sensitive to changes in healthcare service utilization across differing morbidity states in older adults.