Please use this identifier to cite or link to this item:

A Multi-Agent Simulation of MSM HIV Transmission Examining the Role of Concurrent Relationships

File Description SizeFormat 
2015-08-phd-puckett_r.pdfVersion for non-UH users. Copying/Printing is not permitted4.42 MBAdobe PDFView/Open
2015-08-phd-puckett_uh.pdfFor UH users only4.44 MBAdobe PDFView/Open

Item Summary

Title: A Multi-Agent Simulation of MSM HIV Transmission Examining the Role of Concurrent Relationships
Authors: Puckett, Robert
Issue Date: Aug 2015
Publisher: [Honolulu] : [University of Hawaii at Manoa], [August 2015]
Abstract: The Human Immunode ciency Virus (HIV) is a complex virus with a high rate of death, several modes of transmission, and long asymptomatic incubation period. Over time the virus weakens the host's immune system leading to Acquired Immune De ciency Syndrome (AIDS). Transmission of HIV varies over the course of infection, from a high infectivity stage during the initial, acute infection phase, to a lower level during the asymptomatic phase, before rising again as AIDS develops. Once progressed to AIDS, a body becomes susceptible to a range of opportunistic infections and cancers. In developed countries, HIV disproportionately a ects men who have sex with men (MSM), primarily due to the practice of unprotected anal receptive (UAR) sex.
Of particular interest for study, the role of concurrent relationships as a driving force of epidemics has become a hotly debated topic in recent years. Concurrent relationships occur when a person has multiple active sexual relationships with overlapping durations. Several factors may increase the impact of concurrency on transmission including higher infectivity during the primary HIV Infection (PHI) phase of HIV, greater frequency of sex, and an increased exposure to HIV from multiple partners [70, 59]. In 1998, Morris and Kretzschmar concluded from stochastic simulations that concurrent partnerships \dramatically increase the speed and pervasiveness of the epidemic spread" [101]. Numerous subsequent publications by various authors have argued for [115, 83, 125, 58] and against [104, 113, 146] the degree of impact for concurrency in HIV epidemics around the world. Thus, the role of concurrency is far from a settled subject.
Epidemic modeling helps researchers in developing hypotheses on the driving forces, testing intervention strategies, and predicting the course of the HIV epidemic. Although many models for HIV have been developed, HIV remains a di cult subject to model owing to complex transmission modes, limitations in data collection, and variability in human behavior. Instead of studying populations in aggregate, microsimulation techniques allow for the study of individuals over time. A multi-agent system (MAS) can be used to implement a microsimulation of HIV transmission among individual people existing within a complex network of sexual partners. Through simple rules governing agent interactions, complex behavior can emerge from the society of agents.
In our research, we have built a multi-agent simulation of HIV transmission dubbed \MASHIV" used to model HIV epidemics in a single-sex community of men who have sex with men (MSM). HIV progression and mortality is based upon a CD4 compartment model adapted from the Spectrum software [62] to the realm of daily timestep multi-agent systems. MASHIV supports both duration-based dating events and probabilistic dating events. We utilize normal hazard functions to impose normal distributions upon probabilistic events, allowing local event probability generation at the agent level without centralized control. Our simulator includes support for risk reduction strategies such as negotiated safety arrangements, HIV discovery-based risk reduction, assortative/disassortative partner mixing weights, and HIV-status speci c serosorting. The simulator also supports simple antiretroviral drugs (ARV) programs for pre-exposure prophylaxis (PrEP) and antiretroviral therapy (ART).
We applied MASHIV to a simpli ed representation of the MSM epidemic in Bangkok, Thailand to investigate whether concurrency may be a driving factor of the HIV epidemic of the Bangkok MSM community. Our test results suggest that, while concurrency does have a theoretical possibility of contributing signi cantly to HIV spread under certain conditions, concurrency does not currently represent a substantial contribution to the spread of HIV to MSM in Bangkok at this time. While we calculated a positive correlation between concurrency and the number of total infections, the overall change between 0% practicing concurrency and 100% concurrency yielded only a 2% change in HIV prevalence.
Description: Ph.D. University of Hawaii at Manoa 2015.
Includes bibliographical references.
Appears in Collections:Ph.D. - Computer Science

Please contact if you need this content in an alternative format.

Items in ScholarSpace are protected by copyright, with all rights reserved, unless otherwise indicated.