Variation in the apparent importance of health-related problems with the instrument used to measure patient welfare.

AIMS: This paper investigates the distributional implications for eight population groups of using six different instruments to measure wellbeing and to prioritise access to health services. Specifically, it examines the importance of different physical and psycho-social problems for the scores obtained using each instrument and whether scores differ because of differences in the concept measured by the instrument or because of the instrument's construction.
METHODS: Patients with seven chronic conditions and a sample of the 'healthy' public were administered six instruments: two utility instruments; two self-rating scales; a subjective wellbeing instrument and the ICECAP measure of capability. Scores were regressed upon the subscales of the SF-36 and the AQoL-8D. Each instrument's 'problem mix' was measured by the numerical importance of the subscales for the instrument's score and compared with the problem mix of patients constructed from all of the instruments.
RESULTS: The apparent importance of different problems varied significantly with the instrument used to assess welfare but not with the chronic conditions. The correspondence between an instrument's problem mix and the patients' problem mix was highly variable.
CONCLUSION: Different instruments give prominence to different physical and psycho-social problems and consequently favour different groups of patients. Budgetary decisions which appear to be based on efficiency criteria such as the cost per quality-adjusted life year (QALY) conceal distributive effects attributable to the instrument used in the analysis. The effects are additional to the ethical questions considered in making an equity-efficiency trade-off as they arise from the measurement of efficiency.