Skip to contentScholarGate
LibraryBookshelfDeskReview StudioAssistant
Sign in
On this page
IntuitionHow it worksWhen to use itStrengths & limitationsCommon pitfallsApplicationsFrequently asked🔒 Read the full methodSourcesRelated methods
Cite this pageSpotted an issue on this page? Report or suggest a fix →
Home›Epidemiology›Multicenter Case-Crossover Design — Multi-Site Self-Matched Transient Exposure Study
Process / pipelineClinical / epidemiology

Multicenter Case-Crossover Design — Multi-Site Self-Matched Transient Exposure Study

Multicenter Case-Crossover Epidemiological Study · Also known as: multi-site case-crossover study, multicenter self-matched crossover, multi-center transient exposure study, MCCO study

The multicenter case-crossover design is an observational epidemiological method that investigates whether brief, transient exposures trigger acute health events by comparing each case's exposure just before the event to their own exposure during matched control periods — with data collected from two or more independent clinical or geographic sites to increase power, external validity, and the ability to detect site-level effect modification.

ScholarGate
  1. Process / pipeline
  2. v1
  3. 2 Sources
  4. PUBLISHED
Cite this page →
Tools & resources
Download slides
Learn & explore

Read the full method

Members only

Sign in with a free account to read this section.

Sign in

Method map

The neighbourhood of related methods — select a node to explore.

Multicenter Case-Crossover Design
Case-control studyCase-crossover designCase-Time-Control DesignMulticenter cohort studyNested case-control

When to use it

Use the multicenter case-crossover design when the research question concerns whether a brief, transient exposure increases the short-term risk of an acute, discrete outcome and when data from a single center would be insufficient in power or generalizability. The design is ideal when stable between-person confounders (age, sex, comorbidities, socioeconomic status) are the main threat to validity, because self-matching eliminates them. It requires time-stamped individual-level event data and reliable reconstruction of exposure during both hazard and control windows. Do NOT use it when exposures are chronic and constant (there is no within-person variation to exploit), when the outcome is slow-onset or continuous, when time-varying confounders within the individual are likely and unmeasured (e.g., behavioral changes preceding the event), or when the number of incident cases per site is too small to provide stable site-level estimates for pooling.

Strengths & limitations

Strengths
  • Complete control for all stable individual-level confounders — genetic, behavioral, and clinical — without requiring measurement or adjustment.
  • Pooling across multiple sites substantially increases statistical power and the number of rare events available for analysis.
  • Multicenter data enable testing of effect modification by site characteristics such as climate, healthcare system, or population demographics.
  • No separate control group is needed, reducing recruitment burden and selection bias associated with identifying external controls.
  • Well-suited to environmental and pharmacoepidemiological questions where administrative or monitoring data exist across sites.
Limitations
  • Cannot control for time-varying within-person confounders — factors that change between the hazard and control windows and also affect outcome risk (e.g., recent illness, medication changes).
  • Requires that the exposure be genuinely transient; chronic or slowly changing exposures produce no within-person variation and the design provides no information.
  • Control-window selection can induce bias if seasonal trends, weekday effects, or autocorrelated exposures are not properly handled through symmetric bidirectional or time-stratified sampling.
  • Across sites, harmonizing exposure measurement methods and case definitions adds logistical complexity and potential for measurement heterogeneity.
  • Pooled estimates from a small number of sites can be sensitive to influential individual-site results; heterogeneity testing may have low power when the site count is small.

Frequently asked

Why use multiple centers rather than a single large center?

Acute events are often rare at any single site, so multicenter designs accumulate the necessary statistical power. More importantly, pooling across diverse sites improves external validity and enables tests of whether the exposure-event relationship generalizes across different populations, climates, or healthcare systems.

How should I choose control windows across sites?

The time-stratified approach — selecting control windows within the same month and day-of-week as the event — is the most widely recommended because it removes seasonal and weekday confounding without requiring assumptions about the direction of any time trend. This approach should be applied consistently across all sites, with central protocol coordination.

Do I need to adjust for confounders in the conditional logistic regression?

Stable individual characteristics are automatically controlled by the self-matching. However, time-varying within-person covariates that differ between hazard and control windows — such as concurrent medication use, acute infections, or ambient temperature when it is not the exposure of interest — should be measured and included as covariates in the model.

What if heterogeneity across sites is large?

If the I² statistic indicates substantial heterogeneity, a random-effects pooling model should be used and the sources of heterogeneity investigated. Site-level characteristics (e.g., mean pollution level, case ascertainment method, population age distribution) can be entered as moderators in a meta-regression to explain the variation.

Can the multicenter case-crossover design be used for drug safety surveillance?

Yes. It is well-suited to new-user pharmacoepidemiological questions where the transient exposure is treatment initiation and the outcome is an acute adverse event. Linking electronic health records across hospital networks provides both case identification and time-stamped prescription data needed to reconstruct hazard and control-window exposures.

Sources

  1. Maclure, M. (1991). The case-crossover design: A method for studying transient effects on the risk of acute events. American Journal of Epidemiology, 133(2), 144–153. DOI: 10.1093/oxfordjournals.aje.a115853 ↗
  2. Case-crossover study. Wikipedia. link ↗

How to cite this page

ScholarGate. (2026, June 3). Multicenter Case-Crossover Epidemiological Study. ScholarGate. https://scholargate.app/en/epidemiology/multicenter-case-crossover-design

Related methods

Case-control studyCase-crossover designCase-Time-Control DesignMulticenter cohort studyNested case-control

Which method?

Set this method beside its closest kin and read them side by side — the library lays the books on the table; the choice is yours.

  • Case-control studyEpidemiology↔ compare
  • Case-crossover designEpidemiology↔ compare
  • Case-Time-Control DesignSocial Epidemiology↔ compare
  • Multicenter cohort studyEpidemiology↔ compare
  • Nested case-controlEpidemiology↔ compare
Compare side by side →

Similar methods

Matched Case-Crossover DesignCase-crossover designMeta-analytic case-crossover designProspective Case-Crossover DesignRisk-adjusted case-crossover designBayesian Case-Crossover DesignCase-Time-Control DesignMulticenter Case-Control Study

Related reference concepts

Study Matching and StratificationCross-Sectional StudyCase-Control StudyCase-Control and Cohort Studies in Outbreak InvestigationEpidemiologic Study DesignsObservational Study Design

Spotted an issue on this page? Report or suggest a fix →

ScholarGate — Multicenter Case-Crossover Design (Multicenter Case-Crossover Epidemiological Study). Retrieved 2026-07-20 from https://scholargate.app/en/epidemiology/multicenter-case-crossover-design · Dataset: https://doi.org/10.5281/zenodo.20539026
Quick facts
Originator
Malcolm Maclure (single-center design, 1991); multicenter applications developed through 1990s–2000s environmental and pharmacoepidemiology literature
Year
1991 (core design); multicenter extensions 1990s–2000s
Type
Observational epidemiological design
DataType
Individual-level time-stamped event and exposure data across multiple clinical or geographic sites
Subfamily
Clinical / epidemiology
Related methods
Case-control studyCase-crossover designCase-Time-Control DesignMulticenter cohort studyNested case-control
ScholarGate

A content-first reference library for research methods — what each one is, how it works, and where it comes from.

Open data (CC-BY)

Explore

  • Library
  • Search the library…
  • Browse by field
  • Fields
  • Journey
  • Compare
  • Which method?

Reference

  • Subjects
  • Atlas
  • Glossary
  • Methodology
  • Philosophy

Your tools

  • Bookshelf
  • Desk
  • Chat

Company

  • About
  • Pricing
  • Contact
  • Suggest a method

Entries are compiled from published sources for reference. Verifying the accuracy and suitability of any information for your own use remains your responsibility.

© 2026 ScholarGate · A research-method reference library
  • Privacy
  • Cookies
  • Terms
  • Delete account