A recent paper argued for use of models in the evaluation of complex interventions where:
- correlations are not linear,
- components interact,
- feedback loops are incorporated, and
- they adapt over time.
But they leave it there – they do not say how to model the components, still less how parameters can be derived from such models for use in decision models, such as health economic models. CLAHRC WM has developed and published on the use of such models in policy and service delivery research. We show how causal chains can be mapped and how probabilities can be propagated across such causal chains.[2-4] Along with Alec Morton (University of Strathclyde) and Gavin Stewart (Newcastle University), we are leading a workshop on Bayesian causal models at the forthcoming Society of Social Medicine meeting, and will give examples of this work in forthcoming issues of the News Blog.
— Richard Lilford, CLAHRC WM Director
- Rutter H, Savona N, Glonti K, et al. The Need for a Complex Systems Model of Evidence for Public Health. Lancet. 2017.
- Watson SI & Lilford RJ. Essay 1: Integrating multiple sources of evidence: a Bayesian perspective. In: Challenges, solutions and future directions in the evaluation of service innovations in health care and public health. Southampton (UK): NIHR Journals Library, 2016.
- Lilford RJ, Girling AJ, Sheikh, et al. Protocol for evaluation of the cost-effectiveness of ePrescribing systems and candidate prototype for other related health information technologies. BMC Health Serv Res. 2014; 14: 314.
- Watson SI, Taylor CA, Chen Y-F, Lilford RJ. A Framework for the Evaluation of Service Delivery Interventions. J Health Econ. [Submitted].