Heterogeneity in the Relationship between the Standard-gamble Utility Measure and Health-status Dimensions

Abstract

The authors assessed the relationship between the standard-gamble utility measure and the RAND-36 health-status dimensions, taking into account possible heterogeneity among patients in the weights they assign to different health-status dimensions. A questionnaire including both measures was completed by 68 patients with symptomatic peripheral arterial disease. Conventional multiple regression analysis, assuming a ho mogeneous relationship for the total population between the standard-gamble utility and the RAND-36 health-status dimensions, demonstrated that only the dimension social functioning was significant (p < 0.05), which accounted for 10% of the variation. Assuming that the population consisted of two separate classes demonstrated superior representation of the data. Latent class analysis was used to estimate the unknown parameters and class memberships. In the first class, consisting of 65% of the patients, the relationship between the standard-gamble utility and the dimension general health perception was significant. The within-R² was 12%. The second class represented 35% of the patients and showed significant coefficients for the dimensions social functioning and role limitations due to physical problems, which accounted for 80% of the variation. The overall percentage of variation explained by latent class analysis was 49%. The results suggest that patients with symptomatic peripheral arterial disease belong to a variety of classes, all with class-specific relationships between the standard-gamble utility and the RAND-36 health-status dimensions. Key words: health status measure; health utility measure; latent class analysis; heterogeneity. (Med Decis Making 1996; 16:226-233)