The stability of utility scores: test-retest reliability and the interpretation of utility scores in elective total hip arthroplasty

Abstract

Purposes: Are utility scores for hypothetical health states stable over time even when the health of the patient changes dramatically? Can investigators who use scores for hypothetical states be confident about the stability of those scores? The first purpose is to assess the stability of standard gamble utility scores for three hypothetical health states describing mild, moderate, and severe osteoarthritis (OA) (test-retest reliability). How should investigators interpret utility scores? The second purpose is to provide evidence on the marker-state approach to assist in interpreting utility scores.

Background: SG scores for three hypothetical marker states and the patient's current state were obtained at multiple times in a longitudinal study of elective total hip arthroplasty (THA). SG scores for current health increased from a mean of 0.59 pre-surgery to 0.76 post-surgery.

Methods: Test-retest reliability was assessed using the intra-class correlation coefficient (ICC). The effects of time on scores were analysed using an analysis of covariance.

Results: At the group level the marker-state scores were stable. Mean scores for mild, moderate, and severe OA were 0.69, 0.61, and 0.41. With respect to test-retest reliability, ICCs varied from 0.49 to 0.62. In general, time did not affect the scores for the three marker states.

Conclusions: Group-level standard gamble scores are stable. At the individual level scores for hypothetical health states are somewhat stable over time. The marker states assist in interpretation indicating that, on average, THA converted moderate OA to better than mild.