Author Credentials



When planning or conducting research in the hospital setting, often termed Real-World Environment (RWE), therapeutic assumptions and outcomes are often different than in the Randomized Clinical Trial (RCT) where medications, devices and therapies are tested and developed. This is because RWE research has a lack of experimental control, additional confounding due to patient complications and comorbid conditions, lack of pure patient selection and compliance with therapy in the patients being treated and many other factors as well. However, when RWE experiments are conducted, sample size determination using data from the RCT is common because that is the only data that is available when the RWE research is being developed. Using RCT data to derive sample size calculations within the RWE hospital or outpatient setting, on real patients with vastly different conditions has the potential to give inaccurate results. Using newly developed Adaptive Research Designs[chow], which allow for the individual study’s own data for sample size determination is a viable and highly accurate method to prevent under or over sampling in the RWE research context. This paper outlines the proper methodology to use to conduct a “Data-Peek for Power” which is a within RWE, “Adaptive” methodology to calculate sample size without risking reductions in p-values, termed ‘alpha-spend’. Using a Data-Peek for Power is a method that allows for no alpha spend, free from multiple comparison, assessment of statistical power or sample size calculation. When needed it can easily be implemented and described in a research protocol or a proposal that is submitted to the IRB for review listing all relevant variables to be used with the data analysis methods a-priori.

Creative Commons License

Creative Commons Attribution-NonCommercial 4.0 International License
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License