Retrospective Study
A retrospective study is an observational study that starts from the outcome and looks backward to identify exposures, typically as a case-control design comparing past exposures of cases (with the disease) to controls (without).
Why look backward instead of forward?
When the outcome is rare or time is short, starting from cases is far more efficient than waiting for outcomes to develop in a forward cohort.
Use when:
- The outcome is rare; a Prospective Study would need an enormous cohort
- Time and budget are limited
- A natural experiment has already happened (e.g. a contamination event)
Strengths
- Fast and cheap: outcomes already exist
- Efficient for rare diseases: start with cases, no huge unaffected cohort needed
Weaknesses
- Recall bias: sick people remember past exposures differently
- Selection Bias: controls may not be truly comparable
- Temporal ambiguity: harder to nail down whether exposure preceded outcome
- Records may be incomplete
Doll & Hill (1950)
Case-control study comparing lung-cancer patients to healthy controls; one of the first studies linking smoking to lung cancer.