Reporting heterogeneity in the measurement of health and health-related quality of life

Objective: To investigate heterogeneity (systematic and observable variation) in health or health-related quality-of-life reports across population groups, for a given level of 'true health'. Design: The investigators undertook full sit-down face-to-face interviews with 1999 individuals representing Jewish Israelis aged 45 to 75 years who resided in urban Israeli communities in 1993. Three popular measures of health and health-related quality of life were used: a categorical subjective evaluation, the 36-item Short Form health survey and a visual analogue health-related quality-of-life rating scale. The empirical analysis of the various relationships was based on the use of multiple-indicator linear structural equation models with latent variables. The model was estimated by the 2-stage least squares (2SLS) method. Results: In general, the results confirm the existence of substantial measure-specific heterogeneity in reporting (with differences associated with age, gender and education level) for all three measures. Conclusions: The existence of heterogeneity in reporting renders the results of quality-of-life investigations sensitive to the sample used, and considerably limits the ability to generalise from these results and make comparisons with other populations. Economic evaluations based on self-reports are thus sensitive not only to the measure used, but also to the sample used. The application of results to populations differing in their socioeconomic and demographic structures, even if similar in the distribution of 'true health', might be misleading.