OBJECTIVES: The aim of the study was to determine whether there is a relationship between social deprivation and time of HIV diagnosis in France. METHODS: Prospectively collected data from a multicentre database were used in the study. Patients with a first HIV diagnosis between 1 January 2014 and 31 December 2015 were selected from the database. Deprivation was measured using the European Deprivation Index (EDI), which is an ecological index constructed from the address of residence and based on the smallest geographical census unit, in which individuals are classified so as to be comparable with national quintiles. Time of diagnosis was classified as being at an early, intermediate, late, or advanced stage of disease. Age, gender, distance from home to HIV centre, most probable route of infection, and hepatitis B or C coinfection were considered in the analysis. Because of a strong interaction between gender and most probable route of infection, we constructed a 'population' variable: men who have sex with men (MSM), heterosexual men and women. RESULTS: Of 1421 newly diagnosed patients, 44% were diagnosed either late or at an advanced stage of disease, and 46.3% were in the highest deprivation quintile. Using multivariate logistic regression, 'population' [odds ratio (OR) 0.62 (95% confidence interval (CI) 0.48-0.78) for MSM compared with women] and age [OR 1.39 (95% CI 1.07-1.80), 1.72 (1.32-2.23) and 1.86 (1.40-2.47) for the second, third and fourth quartiles, respectively, compared with the first quartile] were found to be related to late diagnosis. EDI level was not related to late HIV diagnosis. CONCLUSIONS: 'Population' seems to be more relevant than EDI to define evidence-based interventions to limit late diagnosis.