The aim of this article is to describe the development of a dynamic French cohort of HIV-infected patients, the methodological issues and decisions made, and the characteristics of the patients currently enrolled. Data are collected during medical encounters. Data quality is ensured by automated checks during data capture, by regular controls, by annual assessments, and by ad hoc processes before any scientific analysis is performed. In September 2007, 10,458 patients representing 59,383 patient-years of follow-up were followed in our centres, including 446 with a first HIV diagnosis in the past year. Among these recently diagnosed patients, 25.6% presented with late diagnosis. Our cohort included 3017 women (28.8%). The women were less likely to be receiving highly active antiretroviral therapy (HAART) than men, and when treated were less likely to be receiving nonnucleoside reverse transcriptase inhibitor (NNRTI)-based regimens. Our network includes medical centres in overseas territories (1105 patients living overseas). In this particular population, women represented 38.5% of the patients, and the probable route of infection was heterosexual in 75.7% of the patients. Despite epidemiological and social disparities, more patients had nondetectable viral loads when receiving HAART in overseas departments than in metropolitan France. The Nadis Cohort represents a collaboration of major French HIV treatment centres. In September 2007, the cohort database contained up-to-date information on more than 10,000 patients, of whom a significant proportion were women. As a consequence of the choices made when building the cohort and the efforts made to ensure the quality of the database, scientific studies are regularly performed using this cohort.
