Ask Mallard

Self-controlled case series

A self-controlled case series uses only people who experienced the outcome, and compares the rate of the outcome during exposed periods with the rate during unexposed periods within the same person.
NHMRC EVIDENCE LEVEL · 2009Not yet classified · InterventionProvisional design classification. Study quality and certainty of findings require separate appraisal.
Why this level?The design family and its defining features have not been specified for this hierarchy. NHMRC 2009 source

Each person contributes both exposed and unexposed time, so all fixed characteristics are controlled by construction. The intervention table has no row for that, and rating it against between-person designs would describe neither its strength nor its real weakness, which is time-varying confounding.

What question does this help answer?

Whether a time-limited exposure raises the rate of an acute outcome, with every fixed characteristic of the person controlled automatically because the comparison never crosses between people.

When to use it

Use it when exposure is intermittent and well recorded, the outcome is acute and recurrent or at least well dated, and confounding by stable personal characteristics is the main worry. It is heavily used in vaccine safety, where exposure timing is precise.

When not to use it

Do not use it when the outcome affects the chance of subsequent exposure, when the outcome is death or otherwise censors observation, or when the exposure is long-term rather than episodic.

How it works in plain language

Each person's observation time is divided into risk periods following exposure and baseline periods. The analysis compares event rates between those periods within each person, so anything that does not change over the observation window cannot confound the estimate.

What risk and baseline periods look like

A self-controlled case series. Each person's observation period is divided into a risk period after exposure and baseline periods at other times, and events are compared between those periods within the same person. The observation period is fixed in advance and is not ended by an event.Self-controlled case series: event rates in risk periods against baseline periods, withineach personRisk periodRisk periodPerson APerson BObservation period, set in advance; time outside the risk period is baselineTimeFixed characteristics of the person cannot confound the comparison. An event that stops laterexposure, or ends observation, breaks the design.Exposure measured or treatment startsOutcome occurs
Conceptual diagram. It shows the structure of the idea, not data from any study.

Why Mallard may recommend it

Mallard may recommend it when the investigator describes a dated exposure such as a vaccination or a dispensed prescription, and an acute outcome recorded in the same database.

Key assumptions and design requirements

  • The occurrence of the outcome does not alter the probability of later exposure.
  • The outcome does not censor observation, which rules out death as the outcome.
  • Risk periods are defined in advance on clinical grounds.
  • Observation windows are defined the same way for everyone and account for age or seasonal effects.
  • Time-varying confounders are handled explicitly, since only fixed ones are controlled by design.

What to watch out for

Every fixed characteristic of the person is controlled by construction, and nothing time-varying is. The method assumes the outcome does not alter the chance of subsequent exposure, which fails badly when the event itself causes the treatment to be stopped or started, and it assumes the event does not censor observation, which fails for death.

Clinical example

Question: Does a vaccine raise the short-term risk of a febrile seizure? Design: Among children who had a seizure, compare the seizure rate in the seven days after vaccination with the rate in the rest of their observation period, within each child.

How to report it

No reporting guideline dedicated to the self-controlled case series was found in a PubMed search on 2026-09-16. STROBE is the guideline to use, with RECORD where the data are routinely collected. State the observation period, the definition and length of the risk window, and whether the event itself can affect later exposure, which is the assumption the whole design stands on. 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.

Code for this method
Loading…
Recent studies that used this
Searching PubMed…
Related Learn pages
Sources & further reading
  1. Self controlled case series methods: an alternative to standard epidemiological study designs (opens in a new tab) · BMJ · doi 10.1136/bmj.i4515
  2. 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
  3. 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
  4. The REporting of studies Conducted using Observational Routinely-collected health Data (RECORD) statement (opens in a new tab) · PLoS Med · doi 10.1371/journal.pmed.1001885
The first reference sets out the assumptions plainly, including the two that most often fail: that the event does not affect later exposure, and that it does not censor observation.
Last source review: 2026-09-06
AI-generated educational content
Mallard uses reputable methodological and reporting sources, but this material may contain errors or omit important context. Verify important methodological decisions with the cited sources and, when appropriate, a qualified research methodologist or statistician.