Strengths and limitations of competitive versus non-competitive models of integrated capitated fundholding

Integrated budget-holding (fundholding) based on risk-adjusted capitation is commonly proposed as a central element of health system reform. Two contrasting models have been developed: the competitive model where fundholders or health plans compete for enrollees; and the non-competitive model, where plan membership is determined according to an objective attribute such as place of residence. Under the competitive model, efficiency is sought through consumer choice of plan. A range of regulatory elements may also be introduced to moderate undesirable elements of competition. Under the non-competitive model, efficiency is achieved through government regulation and the fact that the fundholder has continuing responsibility for the health of a defined population, supported by micro-management tools (such as quality assurance and selective payment arrangements). In theory, the non-competitive model encourages population-based health services planning. While both models assume risk-adjusted capitated funding, the requirements of any formula are more stringent under the competitive model. Economic theory, as well as documented health system experience, can help identify the relative strengths and limitations of each model. Concerns with the competitive model relate primarily to the capacity to develop robust risk adjusters for capitation sufficient to reduce the incentives for patient risk selection. Possible reductions in the quality of care are also a concern, compounded by difficulties for consumers in discriminating between plans. Efficiency under the non-competitive model requires a strong and appropriate regulatory/policy framework and effective use of micro-management tools. Funding equity objectives can be met through either model by the adoption of income-related contributions, but under the competitive model this may be compromised by incentives for the fundholders to select low-risk patients. Evidence drawn from regional fundholding in New South Wales (NSW, Australia), the US Veterans Health Agency and the literature on managed care in the USA illustrate these concerns. The problem of risk selection in the competitive model is a major theoretical concern, confirmed by the empirical evidence. This, together with concerns regarding other aspects of performance, suggests that the non-competitive model may be preferable, at least as an interim step in reform in public or mixed systems. Future research on this issue is clearly required.