The impact of NHS expenditure on health outcomes in England:Alternative approaches to identification in all-cause and disease specific models of mortality

Several recent studies have estimated the responsiveness of mortality to English National Health Service spending. Although broadly similar, the studies differ in how they identify the outcome equation. One approach uses conventional socio-economic variables as instruments for endogenous health care expenditure, whereas the other exploits exogenous elements in the resource allocation formula for local budgets. The former approach has usually been applied to specific disease areas (e.g., for cancer and circulatory disease), whereas the other has only been applied to all-cause mortality. In this letter, we compare the two approaches by using them to estimate the direct all-cause elasticity as well as disease-specific elasticities. We also calculate the implied all-cause elasticity associated with the disease-specific results. We find that the "funding rule" approach to identification can be successfully replicated and applied to disease area models. This is important because disease area models reduce the danger of aggregation bias present in all-cause analysis, and they offer the opportunity to link estimated mortality effects to more complete measures of health outcome that reflect what is currently known about the survival and morbidity disease burden in different programmes.