ABSTRACT
This chapter explains the hands-on aspect of electronic health record (EHR) research by demonstrating how to create various study types from the research dataset to answer a specific research question. In addition to unique identifiers, attributes related to demographic characteristics and markers of socioeconomic status should be included in any analytic dataset. Sampling a case-control study from the research dataset is the most complicated study design. Creating a cohort from the research dataset is a slightly more complicated study design. To sample the cross section from the research dataset will require subsetting the patients based on the years 2001 through 2010 that were born at the hospital and keeping only the core variables needed for the analysis. Cross-sectional studies are frequently employed to measure prevalence and are used more for descriptive epidemiology than for analytic, hypothesis testing epidemiology. Designing a cross-sectional sample is straightforward compared to the other study designs making it an efficient study to conduct.
