Research waste and inefficiency in cardiovascular prediction rule development: exploration of causes and implications

Background: Research waste and inefficiency occur when research studies do not contribute to the decisions of patients, clinicians, and policymakers. In this thesis, I examine research waste and inefficiency throughout the stages of cardiovascular clinical prediction rule (CPRs) development and explore their causes and implications. Methods: This thesis includes six research projects. A mixed-methods study, which consists of a review of derivation studies, thematic content analysis, and author survey, assesses justifications for deriving a new cardiovascular CPR. Using survival analysis, I evaluate the probability for a CPR for cardiovascular disease risk to have an independent external validation study after its derivation. I assess whether external validation studies update the Framingham Wilson coronary heart disease (CHD) risk rule when it performs poorly, and external validation studies contribute to understanding the performance in meta- analyses. A systematic review evaluates the design, methods, and reporting of studies assessing the impact of cardiovascular CPRs. I conduct an online survey to understand which cardiovascular CPRs are recognised and used by GPs in the UK. Lastly, an exploratory qualitative interview study with thematic analysis explores how perceptions of and experiences with cardiovascular CPR research influence primary care providers’ (PCPs) decisions about using cardiovascular CPRs. Results: I found that the authors of derivation studies do not cite existing evidence when justifying a new cardiovascular CPR even though they believe it is important to do so. Most of cardiovascular CPRs do not have a timely conducted independent external validation study. External validation studies of Framingham Wilson CHD risk rule fail to generate the evidence of generalizability to their full potential because they do not update the CPR when it performs poorly, or do not contribute to understanding the CPR’s overall performance. The vast majority of impact studies had flaws in study design, methods or reporting. GPs’ use of cardiovascular CPRs in the UK has evolved and is focused on a few CPRs, such as the QRISK scores. PCPs’ decisions about using cardiovascular CPRs seem to be influenced by their familiarity, trust, and confidence in using research evidence. Conclusions: To reduce research waste and inefficiency in cardiovascular CPR development, it is necessary to shift the paradigm from deriving new CPRs to externally validating, assessing the impact of, and promoting the adoption of high-quality existing cardiovascular CPRs. Also, systematic efforts are needed to reduce the flaws in design, methods, and reporting of CPR studies, and strengthen evidence-base to support the use of cardiovascular CPRs.