O13.2 Can the UNAIDS Modes of Transmission Model Be Improved? A Comparison of the Original and Revised Model Projections Using Data from Nigeria. (13th July 2013)

• Record Type: Journal Article
• Authors: Prudden, H J; Watts, C H; Vickerman, P; Bobrova, N; Heise, L; Ogungbemi, M K; Momah, A; Blanchard, J F; Foss, A M
• Is Part Of: Sexually transmitted infections. Volume 89(2013)Supplement 1
• Page Start: A51
• Page End: A51
• Publication Date: 2013-07-13
• Subjects: HIV -- modelling -- modes of transmission; Sexually transmitted diseases -- Periodicals; HIV infections -- Periodicals; 616.951005
• DOI: 10.1136/sextrans-2013-051184.0156
• Languages: English
• ISSNs: 1368-4973

Abstract

Abstract : Mathematical modelling has increased our understanding of the HIV epidemic and played a key role in decision making. The UNAIDS Modes of Transmission (MoT) model has been used by 29 countries to analyse their HIV epidemics, with the results helping to guide and focus interventions. However, the simplistic compartmentalisation of the population within the model has raised concerns over its reliability. We compared the MoT model projections for Cross River, Nigeria, with a revised MoT model that incorporates additional heterogeneity (including subgroups for sero-discordant partnerships and individuals engaging in transactional sex) and updated parameters. We categorised population subgroups into 'high-risk', representing core groups, their bridging partners and sero-discordant partnerships; 'medium-risk', who are regular partners of 'high-risk' individuals; and 'general population', who are not linked to 'high-risk' or 'medium-risk' groups. Sensitivity analyses were undertaken and model projections assessed which population categories generated the highest incidence of HIV. The original MoT model projections suggest 73% of HIV infections occur in general population subgroups and 21% amongst 'high-risk' groups. Following revisions to the MoT, model projections estimate 76–94% (95%CrI) of new HIV infections are expected to occur in high-risk groups, who make up just 23% of the population, compared to 6–24% amongst general population subgroups. Changes in the distribution of infections result from the introduction of sero-discordant partnerships and 'transactional sex' groups (a relatively large subset of the population, often ignored in modelling analysis) in the model, which are classified as 'high-risk'. The UNAIDS MoT remains an accessible and potentially useful model that can help inform intervention priorities. However, our findings strongly suggest that the current model may produce misleading findings, especially in more concentrated HIV epidemic settings. Results from this study indicate the need for UNAIDS to conduct a formal review of the MoT, and for further revisions to be made.