Estimating health needs: the impact of a checklist of conditions and quality of life measurement on health information derived from community surveys.

BACKGROUND: Prevalence estimates of chronic disease vary according to the technique used. Questionnaire surveys may be susceptible to inaccuracies, which may be overcome by addition of a checklist of conditions. This paper presents SF-36 scores and NHS consultation rates for people reporting individual chronic diseases or disabilities in two questionnaire surveys, one of which employed a checklist and one of which did not. We aimed to document differences in estimates of disease prevalence, and to determine whether or not subjective impact on quality of life is the same in people recruited by a checklist as in those who volunteer that they have a chronic disease or disability without the prompt of a checklist. We use these data to estimate the contribution that different chronic diseases and disabilities make to the burden of disease in the community.
METHODS: Data were collected in two postal questionnaire surveys conducted in 1991 and 1997 with response rates of 72 per cent and 64 per cent. Both questionnaires included a question on long-standing illness, disability or infirmity, together with the SF-36 health status measure. Respondents to the 1991 survey were asked to specify their illness in a free text response, whereas the 1997 survey offered a checklist of conditions. Prevalence rates of each condition were calculated, together with an 'escalation factor' representing the increase in reporting of specific diseases between the surveys. SF-36 domain and component summary scores were calculated overall and for the groups reporting individual chronic diseases or disabilities. Disease-specific NHS consultation rates were calculated for both surveys.
RESULTS: The overall reported rate of chronic disease and disability increased from 28 per cent in 1991 to 42 per cent in 1997. Reported levels of mental health problems and of conditions with a perceived psychosomatic element increased substantially, whereas rates of well-defined conditions were similar. The pattern of SF-36 scores for those reporting chronic disease or disability was similar in the two surveys in spite of very different prevalence rates, and respondents reporting chronic disease had similar levels of health service use. This suggests that they were reporting conditions with similar levels of impact on quality of life. Heart disease, arthritis and mental health problems had the greatest impact on quality of life, and asthma and hypertension the least.
CONCLUSIONS: Evidence from SF-36 scores and NHS consultation rates suggests that addition of a checklist of conditions to this community health survey encouraged reporting of illnesses by the genuinely ill and not merely by those who are less severely affected by their disease. This method appears to give a more accurate reflection of health needs than information derived from routine data sources. The combination of prevalence data combined with subjective assessment of quality of life allows an alternative perspective of health needs. This approach highlights the relative importance of musculo-skeletal problems, particularly back pain, and mental health problems to the burden of disease, and the relative lack of importance of conditions such as asthma. It presents a contrast to studies based on other methods of health needs assessment.