Clinical Research Methodology 1: Study Designs and Methodologic Sources of Error

Abstract

Clinical research can be categorized by the timing of data collection: retrospective or prospective. Clinical research also can be categorized by study design. In case-control studies, investigators compare previous exposures (including genetic and other personal factors, environmental influences, and medical treatments) among groups distinguished by later disease status (broadly defined to include the development of disease or response to treatment). In cohort studies, investigators compare subsequent incidences of disease among groups distinguished by one or more exposures. Comparative clinical trials are prospective cohort studies that compare treatments assigned to patients by the researchers. Most errors in clinical research findings arise from 5 largely distinguishable classes of methodologic problems: selection bias, confounding, measurement bias, reverse causation, and excessive chance variation.