Last Modified

• March 20, 2019

Creator

• Schaffer, Elisabeth
  • Affiliation: Gillings School of Global Public Health, Department of Health Policy and Management

Abstract

• Background: Men in sub-Saharan Africa are less likely than women to test for HIV. The gap in male testing coverage leads to increased morbidity and mortality for HIV-positive men compared to HIV-positive women and to missed opportunities to prevent transmission. One approach that has demonstrated potential to increase male testing is community-based testing. Objective: The objective of this dissertation was to provide evidence-based guidance to optimize the delivery of community-based HIV testing to promote uptake by men in sub-Saharan Africa. Methods: I administered a discrete choice experiment (DCE) to a random sample of 203 adult male residents of rural Uganda. The DCE elicited stated preferences for attributes of community-based testing that can be modified to promote testing. I used a random parameters logit model to estimate preferences and simulate uptake under alternative service delivery models. I estimated additional random utility models and conducted covariate analyses to investigate preference heterogeneity. I incorporated predictions of testing uptake into a decision analytic model to evaluate the cost-effectiveness of alternative service delivery models to increase HIV testing and diagnosis. Key Findings: Participants stated strong preferences for access to antiretroviral therapy (ART) at the time of testing. The predicted uptake of testing under alternative service delivery models increased 26-44 percentage points when immediate access to ART for HIV-positive persons was provided. Incentives of US $0.85 also influenced participants’ choices and increased the predicted testing uptake 6-12 percentage points. It was unclear whether preference heterogeneity could be attributed to participants’ characteristics. The most cost-effective service delivery models to increase HIV testing were not necessarily most cost-effective at increasing diagnosis. Changes of only a few percentage points in the probability that men who access a given service delivery model would test positive greatly increased the likelihood that the service delivery model was cost-effective at increasing diagnosis. Conclusion: The stated preference methods used in this dissertation reveal opportunities to improve community-based HIV testing to encourage uptake by men. Further research is warranted to corroborate the external validity of stated preferences to predict revealed preferences and to determine how community-based service delivery models can effectively reach undiagnosed HIV-positive men.

Date of publication

• August 2018

Keyword

• Mathematical modeling
• Public health
• HIV testing
• Sub-Saharan Africa
• Discrete choice modeling
• Predicted uptake

DOI

• https://doi.org/10.17615/f7jq-3n61

Resource type

• Dissertation

Rights statement

• In Copyright

Advisor

• Stearns, Sally
• Gonzalez, Juan Marcos
• Wheeler, Stephanie
• Thirumurthy, Harsha
• Chamie, Gabriel

Degree

• Doctor of Philosophy

Degree granting institution

• University of North Carolina at Chapel Hill Graduate School

Graduation year

• 2018

Language

• English

Parents:

This work has no parents.

Items

Thumbnail	Title	Date Uploaded	Visibility	Actions
	Schaffer_unc_0153D_18083.pdf	2019-04-10	Public	Press to Select an action Download
	PREMIS_Events_Metadata_0_7691edb8-1a51-4546-abb0-45c79d847881.txt	2019-04-10	Public	Press to Select an action Download
	original_metadata_file_7691edb8-1a51-4546-abb0-45c79d847881.xml	2019-04-10	Public	Press to Select an action Download