When Subjects Are Not Randomly Assigned: The Cohort Approach

Margaret R. Karagas

Critical Appraisal Review covers various topics in epidemiology and related fields that are useful to the practicing physician in understanding and using published studies on skin disease topics. Readers may suggest future topics for this series by writing to the Editorial Office.

In this issue, Margaret Karagas discusses the second major type of epidemiologic investigation, the cohort study. This type of study allows investigators to avoid certain potential pitfalls of case-control studies that were described in the previous issue, but also has its own limitations.

Martin A. Weinstock


Background: Cohort studies are often used to answer clinical and etiologic questions.

Objective: This article is to provide an overview of the basic concepts of cohort studies that are pertinent to the evaluation of the published literature.

Methods: Fundamental principles, methods of assessing exposure and outcome status, choice of a comparison group, and statistical methods used in cohort studies are reviewed. Studies focusing on skin neoplasms are used as examples.

Results: Methods used in cohort studies are similar to those used in clinical trials. One of the distinguishing features is that the exposure (e.g., treatment) is observed rather than assigned by the study investigators. Therefore, an effort must be made to ensure comparability between exposed and unexposed groups.

Conclusion: Cohort studies have provided valuable clinical and etiologic information, but the strengths and limitations of this approach need to be recognized when evaluating the literature.


Antécédents: L’étude de cohortes est souvent utilisée pour répondre à des questions d’ordre clinique ou étiologique.

Objectif: Passer en revue les concepts fondamentaux de l’étude de cohortes qui sont utiles pour l’évaluation de la littérature publiée.

Méthodes: Revue des principes fondamentaux, des méthodes présidant à l’évaluation de l’exposition et des résultats, au choix d’un groupe de comparaison et des méthodes statistiques utilisées dans l’étude de cohortes. Exemplification à l’aide d’études axées sur les néoplasies cutanées.

Résultats: Les méthodes de l’étude de cohortes ressemblent à celles des essais cliniques. Elles s’en distinguent toutefois par le fait que le chercheur observe l’exposition à un agent, c’est-à-dire le traitement, plutôt que d’appliquer ce dernier. Il faut donc impérativement assurer la comparabilité entre le groupe exposé et le groupe non exposé.

Conclusions: L’étude de cohortes génère une importante information clinique et étiologique, mais il importe de ne pas négliger les forces et les limites de cette méthode quand il s’agit d’évaluer la littérature.

Received 1/30/97. Accepted for publication 5/21/97.

Department of Community and Family Medicine and the Norris Cotton Cancer Center, Dartmouth Medical School, Hanover, New Hampshire

Supported in part by grants CA57494 and CA23108 of the National Cancer Institute and ACS SIG-17 of the American Cancer Society.

Reprint requests: Dr. M. Karagas, Department of Community and Family Medicine, Dartmouth Medical School, 7927 Rubin 462M-3, 1 Medical Center Drive, Lebanon, NH 03756-001, US

Full text available in the print edition / Pour le texte intégral veuillez consulter la version imprimée.

JCMS 2(2) Contents