The agency relationship and clinical decision-making in general practice

Abstract

In response to rising health care costs and unexplained variations in medical practice, the agency relationship in health care has attracted the attention of health economists. While the asymmetry of information in this relationship has been recognised as an important source of market failure, relatively little consideration has been given to identifying and measuring agency problems, or to evaluating the proposed solutions to these agency problems. The objective of this thesis is to demonstrate the importance of the agency relationship in explaining the clinical decisions of general practitioners. The thesis is in three parts. Part I considers the working details of the agency relationship in general practice, the agency problems (hidden information and hidden action) that arise under conditions of asymmetric information and the possible solutions to these agency problems. The key feature of agency in health care is the requirement for the general practitioner to act as a common agent of two principals: the patient and third-party funder. As principals the patient and funder face a variety of hidden information and hidden action problems, with a mixture of contractual and non-contractual mechanisms used to resolve these agency problems. It is argued that the worst effects of hidden action are likely to be avoided for the patient, although the potential for exploitation of the funder is much greater. Many of the sources of hidden information are also likely to remain unresolved. Part II looks for evidence of hidden information and hidden action problems in general practice. These agency problems are considered from the perspective of the patient's socio-economic status (SES) and her entitlement to subsidised health care. Patient SES might influence medical decision-making through the communication skills and attitudes of general practitioners, where these characteristics may cause them to misspecify the utility function of patients from lower socio-economic groups. Patient subsidy entitlement may also influence decision making through the financial constraints placed on patients and the income objectives of general practitioners. The empirical results indicate that patient SES is associated with the decisions to prescribe, to follow-up and to refer. Presuming the influence of SES is attributable, in part, to the (unobserved) characteristics of general practitioners, these findings are suggestive of a hidden information problem for patients of lower SES. The results also suggest that the incentives arising from the structure of public entitlements has a part to play in explaining the prescribing decisions of general practitioners. This response is shown to be due entirely to a hidden action problem. Part III considers whether the budget holding scheme, as a response to the hidden action problem for the funder, has been successful in altering the prescribing behaviour of general practitioners. The incentives built into this scheme are consistent with the usual solutions to the hidden action problem: general practitioners are given financial incentives to provide patients with cost-effective care, rather than access to all health care that is effective (as favoured by patients). The empirical results indicate that budget holding leads to a small but significant change in patient drug use. Where the general practitioner held a referral budget for pharmaceuticals, the probability of giving a prescription fell by 1% and the prescription cost, where one was given, fell by 47 cents per encounter. Although these results are suggestive of national savings for 1997 of less than $8 million, the incentives in this scheme do not appear to be strong enough to achieve desired objective.