Methods

Innovative methods are used to analyse complex health policy issues and are central to the economic evaluation of interventions and policies. Quasi-experimental econometric methods, such as robust difference-in-differences, event studies, regression discontinuity designs and instrumental variable analysis, are used to evaluate the efficacy and impact of health and social care policies. These approaches are crucial in establishing causal relationships and assessing the effects of policy interventions.

Researchers working in high-income settings and in global health also apply spatial econometrics and estimate peer effects to explore geographic and interpersonal impacts on health outcomes. Techniques like the changes-in-changes method assess heterogeneous treatment effects, and dose-response functions analyse continuous treatments. Using big GIS data enables the approximation of difficult-to-measure aspects such as healthcare access and utilisation.

CHE contributed to innovations across a broad set of economic evaluation methods. For example, new advances in microsimulation for policy analysis have enabled evaluation of early childhood interventions from a long-term/lifecycle perspective, quantifying impacts on inequalities and for different kinds of individuals in terms of cross-sectoral outcomes. Recent work expanding the outcomes and values considered in economic evaluation have allowed for the incorporation of inequality concerns (distributional cost-effectiveness analysis), inclusion of wider outcomes and costs whilst consistently accounting for opportunity costs (extended impact inventory approach) and a general approach for evaluating public policy (social decision-making analysis). New frameworks have been developed to compare different policies for the pricing of pharmaceuticals and the appropriate share of value to incentivise research and development. CHE has pioneered work on assessing the marginal productivity of healthcare and other types of public expenditure to inform the practical implementation of cost-effectiveness analysis such as the appropriate level of cost-effectiveness 'thresholds'.

Within health technology assessment, recent methods developments have included approaches to structured elicitation of expert judgement in the absence of formal evidence, novel approaches for synthesising evidence including for the evaluation of multi-indication drugs and statistical techniques to assess effectiveness and costs using evidence from real-world healthcare settings.

Machine learning techniques, including Causal Forests and Double-debiased Machine Learning, are increasingly used to refine the estimation of individualised treatment effects. This assists policy-makers in understanding which groups benefit the most from health interventions. Policy learning methods also play a critical role in algorithmically determining the most effective treatment allocations under budgetary and other constraints.