The contribution of natural experiments to the public health evidence base: four case studies in evidence synthesis
Natural experiments arise when an event occurs or a situation arises in society or the environment that divides people into exposed and unexposed groups without allowing them to influence their exposure status. Under such conditions, for example when a policy change occurs or new legislation is enacted, researchers can analyse differences between groups to identify what effects the change had. These research designs are of increasing interest in public health because they may address confounding and selection bias better than other observational study designs. Also, they may have greater external validity than controlled trials because they happen in real-world conditions, and they can constitute an efficient form of research by making use of existing large administrative datasets.
Interest in these designs in public health is growing rapidly. In order to take advantage of natural experiments and translate them into evidence to support decision-making, systematic reviews need to be able to identify, evaluate, and synthesise findings from these study types. However, systematic review methods have largely been developed to support the synthesis of randomised trials and epidemiological studies. Study designs used to analyse natural experiments, such as regression discontinuity (RD) and instrumental variable designs, remain relatively underrepresented in systematic reviews and in the related methodology. This PhD project aims to
- investigate the contribution of natural experimental designs to public health research, specifically the evaluation of public health interventions and environmental causes of disease
- explore how systematic review methods might be applied and developed to make better use of natural experiments to inform public health and policy.
The PhD project consists of four case studies, including a meta-review of endocrine-disrupting chemicals as a cause of breast cancer, a systematic review of RD studies of health outcomes, a systematic review of RD studies of minimum legal drinking age (MLDA) legislation, and the development of a critical appraisal tool for RD studies.