Cohort studies were developed primarily for etiological research purposes. In the field of health monitoring and surveillance, cohorts are neither the most effective tool nor the most cost-effective. Surveillance must rely primarily on other tools. Repeated population-based surveys involving health examinations and biological sampling should be developed. These large cross-sectional surveys should not be confused with cohorts, as they do not involve long-term follow-up. These two approaches are not competing but complementary. Cohorts have their utility and place in surveillance, and the French Institute for Public Health Surveillance (InVS) has been using cohorts for several years. They do, however, have limitations that stem primarily from their initial lack of representativeness or the erosion of that representativeness over time. They do not allow for the monitoring of long-term trends when the cohort is “closed,” that is, when the same subjects are followed throughout the study, with no new inclusions. For the InVS, from an operational standpoint, the following recommendations can be made: - to be able to take the initiative to establish cohorts when necessary to fulfill its missions; - seek collaborations, including contractual ones, with research organizations that manage cohorts so as to be able to utilize the data collected by their teams for the purposes of health intelligence and surveillance (including the ELFE cohort, which the InVS recommended be established); - become more involved in policy-making, coordination, and funding bodies to advocate for the needs of health monitoring and surveillance prior to the implementation of research policies; - actively contribute to efforts to share tools. The InVS cannot be the sole leader in the discussion regarding the value of cohorts for public health in France. A discussion should be organized within a broader framework involving other organizations, particularly the Inserm Institute of Public Health. (R.A.)