Case-crossover study
Why this level?
The design family and its defining features have not been specified for this hierarchy. NHMRC 2009 sourceThe intervention table names cohort, case-control and interrupted time series. It has no row for a self-controlled design, where each person is their own comparison. A rung built for between-person comparison would misdescribe what this controls for: everything about the person that does not change.
What question does this help answer?
Whether a brief exposure TRIGGERS an acute event: whether the hour after heavy exertion, a new drug dose, or an air pollution spike carries higher risk than a comparable period in the same person.
When to use it
Use it when the exposure is transient, the outcome is abrupt and rare, and the risk period is short and can be specified in advance. Because each person serves as their own control, everything stable about them is controlled by construction.
When not to use it
Do not use it for a sustained exposure, a gradual outcome, or when exposure prevalence is changing over time, since the design cannot separate a trend from an effect.
How it works in plain language
Only cases are studied. For each, exposure during a HAZARD PERIOD immediately before the event is compared with exposure during one or more CONTROL PERIODS at other times. The comparison is within the person, so age, sex, genetics, chronic illness and personality are all held fixed without measuring any of them.
What the hazard and control periods look like
Why Mallard may recommend it
Mallard may recommend it when the described exposure is brief and the outcome is acute, and when a conventional control group would be hard to define credibly.
Key assumptions and design requirements
- The exposure is TRANSIENT and the outcome has an abrupt onset.
- The hazard period length is specified in advance on clinical grounds, not chosen to maximise the estimate.
- Control periods are chosen to break any time trend, for example by sampling symmetrically around the event.
- Exposure prevalence is stable over the comparison window, or the design accounts for the drift.
- The analysis is conditional, treating each person as their own stratum.
What to watch out for
It controls only for things that are STABLE within a person over the comparison window; anything that changes over time, including season, day of week, or a trend in the exposure itself, remains confounded and must be handled by design. It requires a transient exposure and an abrupt outcome, and it is invalid when exposure prevalence is drifting.
Clinical example
Question: Does starting a new sedative trigger a fall? Design: For each patient who fell, compare sedative exposure in the 24 hours before the fall with exposure in matched 24-hour windows a week earlier and a week later in the same patient.
How to report it
No reporting guideline dedicated to case-crossover designs was found in a PubMed search on 2026-09-16. STROBE is the guideline to use. Add the choices it does not name: the length and placement of the case and control windows, and how a trend in exposure within the same person was handled. A reporting guideline is not a design manual: use it to make the study transparent, not as a substitute for careful design. Check the target journal's own requirements as well, which this page cannot know.
- The case-crossover design: a method for studying transient effects on the risk of acute events (opens in a new tab) · Am J Epidemiol · doi 10.1093/oxfordjournals.aje.a115853
- Observational research methods. Research design II: cohort, cross sectional, and case-control studies (opens in a new tab) · Emerg Med J · doi 10.1136/emj.20.1.54
- Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement: guidelines for reporting observational studies (opens in a new tab) · BMJ · doi 10.1136/bmj.39335.541782.AD