Introduction - The objective of this study is to compare hospital admissions for long-term illness due to HIV (ALD7) with newly diagnosed HIV-positive cases among individuals aged 15 and older during the 2003–2009 period, in order to assess the value of medical-administrative data for HIV surveillance in France. Methods - ALD data were obtained from the three main health insurance schemes (CnamTS, MSA, and RSI), covering 96% of insured individuals. We analyzed ALD7 admissions with a diagnosis of HIV infection. The number of new HIV diagnoses was estimated based on mandatory HIV reporting, taking into account reporting delays and underreporting. ALD7 data and new HIV diagnoses were compared at the national, regional, and departmental levels. Results - The annual numbers of ALD admissions and new HIV diagnoses are similar, with a total of 48,000 and 49,000, respectively, over the 2003–2009 period. However, the estimated number of new HIV diagnoses is trending downward, while the number of ALD admissions remains stable. Analysis by sex and age group reveals some disparities in trends or numbers, which are more pronounced among women and those aged 15–34. Trends by region and by department in the Île-de-France region and overseas territories are partly consistent. In 2009, ALD admissions outnumbered new HIV diagnoses in all regions. The ratio of ALD7 admissions to HIV-positive diagnoses decreases significantly as the proportion of people born abroad increases. Discussion - Over the 2003–2009 period, the numbers of ALD7 admissions and new HIV diagnoses are relatively consistent despite certain disparities, likely due to differences in the scope of these data: incomplete coverage by the ALD of all individuals residing in France, including some foreign-born individuals, and ALD admissions not necessarily corresponding to new diagnoses. ALD7 data, however, serve as a good proxy for new HIV diagnoses at the national level, whereas regional data should be interpreted with caution. (R.A.)