Can hospital use by a measure of need for health care?

Abstract

Hospital utilisation data are flawed as an indicator of need for health care, since hospital use is as much a product of the supply of services and professional decisions as any notion of population need. Current utilisation thus provides no indication of the appropriateness of different rates of hospital use among socioeconomic groups or of the distribution of unmet needs. Instead, the amount of hospital care people receive is heavily influenced by the organisation and provision of health services in the area where they live. Hospital admission rates will therefore bear a close relation to the incidence of morbidity only for the fairly small number of conditions that can be diagnosed with reasonable certainty, and for which a consensus exists within the medical profession on the need to treat the illness on an inpatient basis. Broad utilisation based measures of population needs, although derived from increasingly sophisticated modelling procedures, will thus always reflect current patterns of service provision. As a result their use in distributing resources between geographical areas will lead to a situation reminiscent of the pre-RAWP era of the Crossman Formula (1971-75), which based target revenue allocations on three elements: the Region's population (given an arbitrary double weight), its beds, and its hospital cases. One way of improving on the use of mortality data as the need proxy in RAWP after the age and sex structure of the population has been allowed for is to address directly questions of the distribution of morbidity, rather than relying on ever more indirect proxies for health resource need. An issue requiring investigation is the contention that the relation between levels of morbidity and mortality is not a linear 1:1 relation as currently assumed by RAWP. Of particular importance would be to establish the prevalence of functional restrictions and of chronic conditions which place considerable demands on the Health services, although only rarely leading to death. The recent health and Lifestyle Survey with its national sample has shown that individuals' general assessments of their own health correlate well with evidence for the incidence of a wide range of symptoms and the results of more formal physical tests. This suggests that one approach would be to conduct sample surveys representative of different socioeconomic groups in various Regions of the country. This could be complemented by the inclusion of self reported morbidity questions in the 1991 national census. Another way of improving morbidity data is to mount epidemiological studies of the population prevalence of selected common conditions requiring hospital care for which mortality is unlikely to be an adequate proxy of prevalence. Thirdly, routine systems of data collection might be extended, including the national morbidity survey in general practice and local disease registers. The search for an alternative to SMRs as a measure of need is often portrayed as a technical problem, the aim being to identify a more adequate measure of morbidity at a population level. However, a fundamental issue which frequently underlies criticisms of SMRs is whether 'need' for hospital resources should be defined solely in terms of morbidity, or whether the definition of need should also include the wider effects of deprivation and geographical variations in the level of primary care provision and of alternative inpatient facilities, since these both directly influence demands on the hospital service, especially at District level. The current approach of employing utilisation data avoids addressing these difficult questions. It merely equates need with that which is measured and, to a greater or lesser extent, reflects current rates of use. An essential prerequisite for the development of measures of need for resource allocation is thus to address directly this central issue of the nature of need and the extent to which different dimensions of need should be compensated for, at least in the short term, through the provision of additional resources in the hospital sector.