P5-S6.28 HIV prevention based on the static modes of transmission synthesis for two Indian districts: insights from dynamical modelling. (10th July 2011)

• Record Type: Journal Article
• Authors: Mishra, S; Vickerman, P; Pickles, M; Ramesh, B M; Washington, R; Issac, S; Moses, S; Blanchard, J; Boily, M C
• Is Part Of: Sexually transmitted infections. Volume 87(2011)Supplement 1
• Page Start: A337
• Page End: A337
• Publication Date: 2011-07-10
• Subjects: Sexually transmitted diseases -- Periodicals; HIV infections -- Periodicals; 616.951005
• DOI: 10.1136/sextrans-2011-050108.584
• Languages: English
• ISSNs: 1368-4973

Abstract

Abstract : Background: The Modes of Transmission (MOT) synthesis uses a static HIV transmission model to predict distribution of incident infections along subgroups over 1 year, and directs HIV prevention along this distribution. Because the MOT does not consider where sustained transmission is most likely to occur, and does not use parameter combinations fitted to observed epidemic characteristics, its relevance for planning interventions may be limited. Methods: We fitted a dynamical HIV/STI heterosexual transmission model to districts Mysore and Belgaum, India. The MOT and dynamical models estimated the proportion of new HIV infections over 1 year due to transmission between female sex workers/clients, their non-commercial partnerships, and low-risk partnerships. We compared predictions from the dynamical model to MOT results using prior and posterior (fitted) parameters. Intervention impact was illustrated using the dynamical model. Results: Using prior inputs, the MOT predicted that commercial sex accounted for 66.2–70.6% of incident infections among males, whereas 71.7–74.2% of incident infections among females were due to bridging infections from clients. There was less variability in MOT results when fitted inputs were used. The majority of the remaining new infections in males and females were due to transmission within low-risk partnerships. In contrast, the dynamical model predicted a higher contribution of commercial sex among males (90.7–91.2%), a higher contribution of bridging infections among females (70.5–86.9%), and that <1.5% of infections were due to low-risk partnerships. Dynamical modelling predicted that any intervention that reduces transmission by 20% applied among commercial sex partnerships could decrease overall HIV incidence by 12% in the first year and by 21% in 5 years see Abstract P5-S6.28 table 1 . Applying this intervention among non-commercial partnerships of clients reduces overall incidence by 9% in years 1 through 5 because clients continue to become infected from their commercial partnerships. Conclusion: Prior inputs for the MOT will not reflect observed HIV prevalence, and as a result, will produce greater variability in MOT predictions. Allocating resources along a 1-year distribution of incident infections can prioritise prevention to the wrong subgroups because they do not account for the dynamic effects of interventions. Improved methods of epidemic appraisals are urgently needed to guide prevention programming.