XVIII International AIDS Conference

Abstract

Back to the PAG
Back
Sign In

What are mathematical models of HIV/AIDS progression modeling? A review of the literature

Y.N. Caro Vega, A. Colchero Aragones, A. Valencia Mendoza, S. Bautista-Arredondo

Instituto Nacional De Salud Pública, Health Economics Division, Cuernavaca, Mexico

Background: Since the beginning of the epidemic, mathematical models were developed to represent HIV/AIDS natural progression. These studies have used a wide range of approaches and to address a wide range of policy questions. Previous reviews have analyzed specific topics, such as drug resistance or the impact of a vaccine. Our aim is to conduct a systematic review of HIV/AIDS progression models to identify their mathematical structure, the main topics and research questions posed and the main assumptions used to represent HIV progression.
Methods: A systematic review of HIV progression models was conducted, covering papers from 1991 to 2008. We found 3,455 articles, but only 84 met the inclusion criteria. We classified the models by their mathematical structure: state transition models, differential equations (DE), and others. We also identified the models by their research question, temporal horizon and the stage of the disease (natural history versus treatment), characteristics of the population at baseline, sensitivity analysis and validation methods used to support the models and their results.
Results: Among the selected papers, 56 used transition models, 20 applied DE and 8 had other structures. Most studies based their models in epidemiological and clinical data of observed cohorts. Regarding the type of question, most papers deal with prevalence and incidence of HIV/AIDS in specific places, evaluation of prevention programs, therapy optimization, initiation of ARV and cost-effectiveness of specific drugs. Transition models are used in every research question. DE models are frequently used to describe inter-cellular relations. Sensitivity analysis is used in 90% of the papers and 78% mention the use of a validation method.
Conclusions: HIV models can help decision makers to select effective interventions. Understanding the structure, research questions, general characteristics and main assumptions of HIV/AIDS progression models facilitates the comparison between studies and assessment of their limitations.


Back - Back to the Programme-at-a-Glance


Contact Us | Site map © 2010 International AIDS Society