Automated utility assessment of global health

The objective of this study was to characterize the performance of an automated utility assessment instrument for measuring preferences for overall health. The study population consisted of 83 subjects recruited from the cafeteria of a large tertiary care hospital. We assessed utilities for current health relative to perfect health and death using the rating scale, time tradeoff and standard gamble metrics. To validate the instrument, we compared utilities with the General Health subscale of the SF-36 Health Survey instrument, satisfaction with current health, and degree of bother due to current health. We evaluated interview failure rate based on irrational orderings of two practice assessments (monocular and binocular blindness) or inability to complete the interview. As expected, utility for overall health was statistically significantly associated with the General Health subscale score and measures of satisfaction with current health and degree of bother. There is substantial variation in utilities among patients with similarly severe overall health, and substantial overlap in utilities among subjects with different levels of overall health. The failure rate in the study was acceptable (9.6%). Automated assessment of utility for overall health provides a feasible means for estimating individual preferences.