What is a retrospective study?

In a retrospective cohort study, exposure and outcome data were both gathered before the study began (Setia). Hess defines a retrospective study as one that uses existing data recorded for purposes other than research. For cohort studies, the STROBE explanatory paper reports a usage in which the distinction rests on when data were collected relative to when the study idea arose; STARD 2015 likewise asks diagnostic accuracy studies to say whether data collection was planned before or after the tests.

The word is applied to more than one design. A retrospective cohort selects participants by exposure and reads the outcome from records; a case-control study selects them by outcome and looks back at exposure. Setia holds that the sampling method decides which of the two a study is and, citing Porta’s Dictionary of Epidemiology, notes that case-control participants are often recruited prospectively. STROBE finds the words ill defined (one usage reserves “retrospective” for case-control studies) and does not use them in its own checklist.

According to Setia, a retrospective cohort is faster and cheaper than a prospective one, and Talari and Goyal value retrospective studies for rare diseases. Talari and Goyal tie the weaknesses to charts that were not designed for research, so gaps in the information are unavoidable, and they add that selection and recall bias affect results; Setia notes that exposure may have been measured in different ways across people. STARD’s explanation, written for diagnostic accuracy studies, says retrospective studies may overlook eligible patients but can sometimes mirror everyday practice better.