Measuring health status in the community: a comparison of methods.

STUDY OBJECTIVE--The measurement of health outcomes is central to the evaluation of medical treatment and intervention. It is generally acknowledged that such measurement ought to include an assessment of the impact of health care on the quality of life, as well as its quantity. The Health Measurement Questionnaire (HMQ) was developed as a means of identifying respondents in terms of Rosser's classification of illness states. This study examines the extent of convergent validity of the HMQ when used as a self report measure of health status, alongside the General Health Questionnaire (GHQ) and the Nottingham Health Profile (NHP). DESIGN--A randomised survey of residents of Wolverhampton was drawn from the electoral register. Interviews were conducted in the respondents' own homes. The three measurement instruments formed part of an extensive battery of questionnaires aimed at assessing a broad range of health issues. PARTICIPANTS--A total of 430 respondents were interviewed, of whom 407 completed the HMQ. Altogether 210 also completed the GHQ, and a further 207 completed the NHP. Failure in the interview protocol meant that 12 respondents did not complete either the GHQ or the NHP; these respondents did complete their HMQ. MEASUREMENTS AND MAIN RESULTS--At the descriptive level, Rosser distress categories derived from the HMQ seem to correlate well with the GHQ. There is a strong association between weighted Rosser disability/distress states and scores produced using the NHP. All three measures discriminated between "healthy" and "not healthy" subgroups of respondents. CONCLUSIONS--The results indicate strong evidence for convergent validity. There are significant levels of physical and psychological morbidity within the community. The results of this study reinforce the case for the continued measurement of health status within the general population. Low cost techniques such as the HMQ offer the prospect of such measurement.