Secondary data analysis: techniques for comparing interventions and their limitations

Purpose of review Secondary data analysis has become increasingly common in health services research, specifically comparative effectiveness research. While a comprehensive study of the techniques and methods for secondary data analysis is a wide-ranging topic, we sought to perform a descriptive study of some key methodological issues related to secondary data analyses and to provide a basic summary of techniques to address them. Recent findings In this study, we first address common issues seen in analysis of secondary datasets, and limitations of datasets with respect to bias. We cover some strategies for handling missing or incomplete data and a basic summary of three statistical approaches that can be used to address the problem of bias. Summary While it is unrealistic for surgeon scientists to aspire to the depth of knowledge of professional statisticians or data scientists, it is important for researchers and clinicians reading to understand some of the common pitfalls and issues when using secondary data to investigate clinical questions. Ultimately, the choice of analytical technique and the particular data sets used should be dictated by the research question and hypothesis being tested. Transparency about data handling and statistical techniques are vital elements of secondary data analysis.