Examples
Worked examples
- Is an instance
A research team pulls five years of electronic health records for patients who received a particular chemotherapy regimen and compares their documented outcomes to those of patients who received a different regimen over the same period -- both exposure and outcome were already recorded before the study began, making this a retrospective cohort study.
- Is an instance
Investigators identify patients newly diagnosed with a rare cancer (the outcome) and, working backward through medical and occupational records, compare their prior exposure history to that of matched controls without the disease -- a classic retrospective case-control design used precisely because prospectively following enough people to observe a rare outcome would be impractical.
- Is an instance
A hospital quality-improvement team reviews the prior year's surgical charts to calculate a complication rate after a procedural change -- the complications had already occurred before the chart review began, so the analysis is retrospective even though no formal hypothesis-testing study was planned in advance.
Counter-examples
Looks similar, but isn't
- Not an instance
A clinical trial that enrolls patients and randomizes them to a new drug or placebo, then follows them forward for 12 months to measure outcomes, is prospective, not retrospective -- the outcome has not occurred at the time subjects are enrolled and data collection begins after enrollment, even though the final analysis happens after the study concludes.
- Not an instance
A prospective cohort study that enrolls currently-exposed and currently-unexposed groups and follows both forward in time to see who develops the outcome is not retrospective, even though it may take years to complete and the final report is written well after enrollment -- the defining feature is that data collection was designed and began before the outcome occurred, not when the paper was written.
Editorial commentary
A retrospective study looks backward: subjects are identified and data are analyzed only after the outcome of interest has already occurred, drawing on records that were generated earlier and usually for a different purpose — clinical documentation, administrative or claims data, laboratory results, registries, or patient recall. This timing relationship, not the specific statistical design, is what makes a study retrospective. A retrospective study can be structured as a cohort study (comparing outcomes across groups defined by past exposure), a case-control study (comparing past exposure across groups defined by current outcome status), or a simple case series or chart review.
The term is most useful as a contrast with prospective designs, where subjects are enrolled and followed forward in time before the outcome has occurred. See Prospective vs. Retrospective Study for a full side-by-side comparison of design, bias exposure, and evidence strength between the two.
Why the distinction matters
Because a retrospective study reconstructs exposure and outcome from data that were not collected with the study’s hypothesis in mind, it is more exposed to specific forms of bias than a prospective design: recall bias (participants with the outcome remember and report past exposures differently than those without it), record/measurement bias (documentation quality and consistency vary across the records being reused), and selection bias arising from which records happened to be available or complete. These limitations don’t make retrospective evidence invalid — STROBE (STrengthening the Reporting of OBservational studies in Epidemiology) recognizes retrospective cohort, case-control, and cross-sectional studies as core observational designs — but they do generally place retrospective evidence lower on the hierarchy of causal-inference strength than a well-conducted prospective or randomized design, because temporal sequence and measurement consistency must be inferred after the fact rather than controlled for in advance.
Common retrospective study designs
- Retrospective cohort study: groups are defined by exposure status that already existed in the historical record, and their already-recorded outcomes are compared going forward from that historical exposure point.
- Case-control study: inherently retrospective in its basic logic — subjects are selected on outcome (case vs. control) and their exposure histories are examined backward. This makes case-control designs efficient for studying rare outcomes or diseases with long latency.
- Retrospective chart review / case series: a review of existing clinical records, often used for quality improvement, hypothesis generation, or characterizing a condition, without a comparison group.
Why it matters for research administration
The prospective/retrospective distinction has direct administrative consequences beyond study design. It affects how a protocol is classified for IRB review — a retrospective review of existing records may qualify for expedited review or, depending on identifiability and consent history, an exemption determination, in ways a prospective study enrolling new subjects typically does not. It also affects HIPAA authorization pathways: retrospective research using protected health information often proceeds under an IRB or Privacy Board waiver of authorization rather than individual patient consent, which in turn triggers its own accounting-of-disclosures obligations under the Privacy Rule. See HIPAA and Retrospective Research for the compliance-specific requirements, and Waiver of Informed Consent (45 CFR 46.116(f)) for the Common Rule consent-waiver test that most retrospective, minimal-risk record-based research relies on.
Related terms
Machine-readable encodings
Use in your systems
<role vocab="credit"
vocab-identifier="https://casrai.org/dictionary/"
vocab-term="Retrospective Study"
vocab-term-identifier="https://casrai.org/dictionary/term/retrospective-study" />{
"@context": "https://schema.org",
"@type": "DefinedTerm",
"@id": "https://casrai.org/dictionary/term/retrospective-study",
"name": "Retrospective Study",
"identifier": "https://casrai.org/dictionary/term/retrospective-study",
"description": "A retrospective study is an observational research design in which investigators identify subjects, exposures, and outcomes that have already occurred and reconstruct the exposure-outcome relationship from data recorded before the study began -- most often clinical records, administrative/claims data, registries, or prior test results -- rather than measuring events as they unfold going forward. What makes a study retrospective is not its statistical method but the timing relationship between the start of the study and the events under analysis: if the outcome had already happened before subjects were identified and data collection began, the study is retrospective, regardless of whether it is framed as a cohort, case-control, or case-series design. Because the data were typically generated for clinical care, billing, or another purpose unrelated to the study's hypothesis, retrospective studies trade weaker control over data quality and measurement consistency for speed, lower cost, and the ability to study rare outcomes or long latency periods that a forward-looking design could not practically capture.",
"inDefinedTermSet": "https://casrai.org/dictionary/domain/clinical-research#set",
"url": "https://casrai.org/dictionary/term/retrospective-study",
"sameAs": [],
"license": "https://creativecommons.org/licenses/by/4.0/",
"publisher": {
"@id": "https://casrai.org/#organization"
},
"dateModified": "2026-07-23T09:27:21",
"inLanguage": "en"
}
