Biomedical research has been struck with the problem of study findings that are not reproducible. With the advent of large databases and powerful statistical software, it has become easier to find associations and form conclusions from data without forming an a-priori hypothesis. This approach may yield associations without clinical relevance, false positive findings, or biased results due to "fishing" for the desired results. To improve reproducibility, transparency, and validity among clinical trials, the National Institute of Health recently updated its grant application requirements, which mandates registration of clinical trials and submission of the original statistical analysis plan (SAP) along with the research protocol. Many leading journals also require the SAP as part of the submission package. The goal of this article and the companion article detailing the SAP of an actual research study is to provide a practical guide on writing an effective SAP. We describe the what, why, when, where, and who of a SAP, and highlight the key contents of the SAP.
Keywords: SAP; reproducibility; research methodology; statistical analysis; transparency; validity.
© 2019 John Wiley & Sons Ltd.