relationship between health care need and standardized mortality ratios in Ontario

Needs-based health care funding methods are increasingly being considered for the Canadian health care system, particularly as a basis for allocating funds from central Ministries of Health to local and regional health authorities. Standardized Mortality Ratios are one of the commonly used measures of health care need within such funding formulae. This paper applies the non-linear least-squares method suggested by Bedard et al. (2000) to identify the empirical relationship between standardized mortality ratios (SMRs) and health care need in Ontario. The data used in this analysis allow for a more detailed examination of the SMR-need relationship. Like Bedard et al., we find the SMR-need relationship is highly non-linear; unlike Bedard et al., we find the relationship to be less-than-proportional, reversing the conclusion that needs-based formula based on the linearity assumption under-compensate high-need regions and overcompensate low-need regions. We also find that estimates derived from provider-based expenditure series are especially sensitive to alternative model specifications. Finally, needs-based funding would generate substantial inter-regional reallocations of health care resources in Ontario compared to the current funding methods.