Objective: The US Public Health Service Panel on Cost-Effectiveness recommends using the general public to assess utilities for health states for cost-effectiveness analyses. We assessed whether people's utilities for their own health affect their utilities for hypothetical health states for scleroderma (SSc), a rheumatic disease.

Methods: Subjects, recruited through flyers and advertisements in local newspapers, completed 2 health status measures, the SF-12 version 2 and the Health Assessment Questionnaire-Disability Index (HAQ-DI); rated their own health on a 0-100 rating scale (RS); and completed computer-assisted time trade-off (TTO, range: 0.0-1.0) and standard gamble (SG, range: 0.0-1.0) utility assessments for their own health. We then presented brief descriptions of several SSc health states derived from the Quality of Well-being scale and asked subjects to imagine living in each of those states for the rest of their lives. Each subject assessed 3 of the following 5 SSc health states: mild SSc, moderate SSc, severe SSc, moderate SSc with lung disease, and severe SSc with lung disease. To assess the impact of subjects' utility scores for their own health on utilities for SSc, we performed multivariable regression analyses for each of the 3 valuation methods (RS, TTO, and SG) on each of the 5 SSc health states, controlling for demographics, SF-12 Physical Component Summary (PCS) and Mental Component Summary (MCS) scores, HAQ-DI scores, and self-utility scores.

Results: A total of 218 adults completed the study; 135 (62%) were female, 143 (66%) were Caucasian, and 62 (28%) were African-Americans. The median scores for the PCS-12 (52.0), MCS-12 (49.0), and HAQ-DI (0.12) were close to population norms. The median (25th, 75th percentile) values for subjects' own current health state were RS: 85 (75, 90); TTO: 0.998 (0.93, 1.00); and SG: 0.992 (0.88, -1.00). Multivariable analyses revealed statistically significant positive associations between self-utility scores and valuations for 3 of the 5 SSc health states; the R2 in the RS models ranged from 0.11 to 0.14; in the TTO models, they ranged from 0.07 to 0.22; and in the SG models, from 0.07 to 0.37. Other significant independent variables included demographics, health status scores, or both.

Conclusion: Peoples' valuation of their current health impacts their valuations of hypothetical health states. Thus, utilities assigned by the general public may be biased.