Problems of using modelling in the economic evaluation of health care

Abstract

Increasingly models are being used as an integral part of health care evaluation or health technology assessment. They are methods of representing the real world, usually by reducing complexity to simple elements in order to better understand the way a system works or to predict its effect. Models are usually expressed in mathematical form and may be deterministic or stochastic. The data may come from a variety of sources and are subject to varying degrees of uncertainty. Any health research involves the use of some sort of modelling in order to abstract quantitative or qualitative data. Even the collection of information in case notes uses a process of abstraction which simplifies and categorises a complex reality. Important policy decisions for example, about reimbursement are influenced by the results of modelling which is used to derive clinical, resource use and cost of resources data. However, despite their increasing popularity, little critical ahention has been given to the use of modelling in economic evaluation compared to the detailed discussions over discrete methodological issues such as the correct way to deal with future health benefits or production losses due to illness. Economic evaluation should play a fundamental role in health care decision making, therefore it is vital that the methods used are carefully scrutinised and refined. Because of the growing importance of health economics in policy decisions there is also the risk that the way modelling is carried out will reflect commercial or government interests. ‘New marketing techniques need