Improving the efficiency and relevance of health technology assessent: the role of iterative decision analytic modelling

Decision making in health care involves two sets of related decisions: those concerning appropriate service provision on the basis of existing information; and those concerned with whether to fund additional research to reduce the uncertainty relating to the decision. Information acquisition is not costless, and the allocation of funds to the enhancement of the decision makers’ information set, in a budgetconstrained health service, reduces the ‘pot’ of resources available for health service provision. Hence, a framework is necessary to unify these decisions and ensure that HTA is subject to the same evaluation of efficiency as service provision. A framework is presented which addresses these two sets of decisions through the employment of decision analytic models and Bayesian value of information analysis, early and regularly within the health technology assessment process. The model becomes the vehicle of health technology assessment, managing and directing future research effort on an iterative basis over the lifetime of the technology. This ensures consistency in decision making between service provision, research and development priorities and research methods. Fulfilling the aim of the National Health Service HTA programme, that research is “produced in the most economical way” using “cost effective research protocols”. The proposed framework is applied to the decision concerning the appropriate management of female patients with symptoms of urinary tract infection, which was the subject of a recent NHS HTA call for proposals. A probabilistic model is employed to fully characterise and assess the uncertainty surrounding the decision. The expected value of perfect information (EVPI) is then calculated for the full model, for each individual management strategy and for particular model parameters. Research effort can then be focused on those areas where the cost of uncertainty is high and where additional research is potentially cost-effective. The analysis can be used to identify the most appropriate research protocol and to concentrate research upon particular parameters where more precise estimates would be of most value.