The evidence surrounding the safety of statins is conflicting despite a number of systematic reviews and meta-analyses. We will use a regression discontinuity analysis (RDA) to assess the effects of statins on both intended and unintended consequences. RDA is a statistical technique that allows for causal inference when a decision rule (such as being above or below a cut-off value on a continuous measure) is used to assign treatment. It is a quasi-experimental method that (like RCTs) reduces the effect of confounding of unobserved variables. The assumption being that patients lying just either side of the cut-off value are similar, in terms of observed and unobserved characteristics. Focussing on a small window surrounding the cut-off value should result in treatment assignment being the only difference between patients. We will therefore look at patient QRISK2 score as the exposure and our outcomes will include cardiovascular disease, future cholesterol levels and future QRISK2 scores (to determine efficacy of statins) and commonly reported side effects such as muscle pain and weakness, nausea and diabetes development (to determine safety). RDA is a novel method in clinical and epidemiological studies and it will provide insights into the causal effects of statins on side effects and also their efficacy.
Kate Tilling - Chief Investigator - University of Bristol
Theresa Redaniel - Corresponding Applicant - University of Bristol
Lauren Scott - Collaborator - University of Bristol
Ruta Margelyte - Collaborator - University of Bristol