HIV testing, HIV/AIDS Knowledge, and Sexual Behavior

This study utilizes a randomized controlled trial (RCT) to explore two important issues regarding HIV/AIDS knowledge and HIV testing: 1) how to promote HIV/AIDS knowledge and demand for HIV testing and 2) what are the causal effects of HIV/AIDS knowledge and HIV testing on sexual behavior. In this paper, we try to understand these questions directly through two rounds of experiments. During the first round experiment, three randomly selected treatment groups are offered three (overlapping) treatments: HIV education only (Group 1), HIV education and home HIV testing (Group 2), and HIV education and a conditional cash transfer for a facility-based HIV testing (Group 3). During the second round experiment, all groups are offered either home HIV testing or a conditional cash transfer. We first find that the level of HIV/AIDS knowledge significantly increases, and also find suggestive evidence of knowledge spill-over. The HIV testing rate increases by 7, 64, and 57 percentage points in Group 1, 2, and 3, respectively. We find suggestive evidence that although test take-up rate is similar in home and facility-based HIV testing, home testing is substantially more efficient to detect those with HIV. In addition, we find that HIV testing take up does not dampen future demand for HIV testing. Infection expectation is also updated for both learned negative and positive result from the testing, but the effect does not persist after six months. These outcomes correspond to the finding that the probability of having multiple partners increases in the short run, but this increase disappeared within six months. (JEL: I10, C93, D80) 1 Haile: International Food Policy Research Institute (IFPRI); Kim: Department of Policy Analysis and Management, Cornell University, [email protected]; Lee: College of Nursing, Yonsei University, [email protected]. The authors would like to thank Booyuel Kim, Cristian Pop-Eleches, Eric Verhoogen, and Leigh Linden, and seminar participants in Columbia University, Cornell University, and Myoungsung Cristian Medical School in Ethiopia. We acknowledge financial support from the Korea International Cooperation Agency (KOICA), the Program for Economic Research at Columbia University, the Center for the Study of Development Strategies and the Applied Statistics Center, and Yonsei University College of Nursing. In Ethiopia, we acknowledge technical and material support from Arsi Zone and Hetosa District Heath Bureaus, Adama University Department of Public Health, Ethiopian AIDS Resource Center, and Population Services International. We thank Kyuk Suk Chang, Senait Assefa, and Kassahun Mengistie for their excellent research assistance. All errors are our own.