Larger HIV care units are not necessarily better! impact of units characteristics on treatment outcomes. Results from the French national representative ANRS-VESPA2 study

Background: As a result of the last economic crisis, French HIV care services tend to be collapsed together for reducing costs. To date, the impact of health care centers organization on patients'' outcomes has been rarely investigated. We aimed to investigate whether virological (VS) and immunological (IS) successes are explained not only by individual factors, but also by the hospitals┬┤ characteristics among a representative sample of PLWH in France.
Methods: ANRS Vespa2 is a cross-sectional survey conducted in 2011 on 3022 adult PLWH attending French hospitals with caseloads ranging between 72 and 3300 patients. Data collection was based on structured face-to-face interviews, patient records and hospital characteristics. Patients receiving ARV treatment since at least one year were selected (n=2612). VS was defined as having undetectable viral load at the time of the survey; IS was defined as having either CD4 count≥500cells/mm3 or a specific CD4 retrospective gain estimate since ARV initiation (100 (1-st year), >200 (years 2-4) or 350 cells/mm3 (4th year)). Potential explanatory variables for both VS and IS were identified using bivariate analysis. Two-level logistic models were estimated by adjusting according to caseload epidemiological characteristics at the hospital level).
Results: Among the study sample, 79% and 71% patients were in VS or IS, respectively. VS and IS were explained by both structural and individual factors: 6.6% and 3.1% of the total model variability was explained by differences between hospitals, respectively. Hospitals with unit having HIV caseloads < 1700 (AOR:2.0;IC95[1.2,3.3]) and units with HIV care as dominant activity (AOR:1.56;IC95[1.08,2.27]) were associated to the highest probability of VS, after adjustment on patients┬┤ characteristics, i.e. non-European citizenship, income and self-reported adherence. Patients followed-up in Paris area hospitals were less likely to be in IS (AOR:0.73;IC95[0.56,0.95]) adjusting for age and self-reported adherence.
Conclusions: Centralization of care is a challenge in a period of economic crisis. However, public health authorities should i) encourage the development of hospital units with HIV predominant activity; ii) monitor and correct the way HIV care is organized in order to optimize patients'' response to ARV treatment and to assure high standard of quality of care especially in units with large HIV caseload.

L. Sagaon Teyssier1,2,3, L. Fressard1,2,3, M. Suzan-Monti1,2,3, M. Préau1,3,4, P. Carrieri1,2,3, F. Lert5,6, R. Dray-Spira7,8, B. Spire1,2,3, ANRS-VESPA2 Study Group
1INSERM UMR 912 (SESSTIM), Marseille, France, 2Aix Marseille Université, UMR_S 912, IRD, Marseille, France, 3ORS PACA, Observatoire Régional de la Santé Provence-Alpes-Côte-d'Azur, Marseille, France, 4GREPS, Université Lyon 2, Lyon, France, 5INSERM U 1018, Centre de Recherche en Epidémiologie et Santé des populations, Villejuif, France, 6Université de Versailles Saint-Quentin en Yvelines, Villejuif, France, 7INSERM, UMR_S 1136, Pierre Louis Institute of Epidemiology and Public Health, Team Research in Social Epidemiology, Paris, France, 8Sorbonne Universités, UPMC Univ Paris 06, UMR_S 1136, Pierre Louis Institute of Epidemiology and Public Health, Team Research in Social Epidemiology, Paris, France