Theses Doctoral

The Effects of Antiretroviral Therapy Scale-Up on Tuberculosis and Non-Communicable Diseases Health Service Utilization and Mortality Risk among the General Population in Rural South Africa, 2009-2014

Saito, Suzue

The overall purpose of this dissertation was to examine evidence of spillover effects of HIV care and treatment service scale up in sub-Saharan Africa in the past decade. Particularly the focus was to quantify any effect HIV treatment initiation by a person living with HIV (PLHIV) may confer health benefits to the HIV negative population by increasing utilization of non-HIV services or reduce mortality risk.
This dissertation had three primary aims. The first aim was to conduct a systematic review of the effect of increasing ART uptake in high HIV prevalence communities on use of non-HIV health services, including maternal, child, in/out-patient, non-HIV laboratory, and TB diagnosis and treatment services. Overall positive effects were found on the majority of health service indicators examined for non-HIV laboratory service utilization and Tuberculosis diagnosis and treatment services. We found negative associations on the majority of indicators examined for child health services. The existing evidence did not point to clear tendencies for maternal health services and outpatient and inpatient services. Restricting the sample to studies with stronger study designs for causal inference, the positive effect on non-HIV laboratory services and the negative impact on child health services held but evidence was mixed for TB diagnosis and treatment services, maternal health services and outpatient and inpatient services.
The second aim of this dissertation was to conduct regression discontinuity quasi-experiments to determine whether exposure to health benefits from ART utilization by a person living with HIV (PLHIV) in a household affects uptake of TB, hypertension (HTN) and diabetes mellitus (DM) treatment by other household members with these conditions. The study was conducted in the comprehensive population cohort followed by the Africa Health Research Institute (AHRI) in Kwazulu-Natal (KZN), South Africa. We linked PLHIV engaged in HIV care to their cohabitating household members aged ≥15 years using a unique identifier for homesteads. Household ART utilization significantly increased treatment for diabetes (RR 1.90: 95% CI 1.07-3.40) but not for TB (RR 1.12: 95% CI 0.71-2.03) or hypertension (RR 1.31: 95% CI 0.97-1.77).
The third aim of this dissertation was to use the same regression discontinuity design and KZN cohort data as in aim 2 to determine whether exposure to health benefits from ART utilization by PLHIV in a household reduces all-cause mortality of other household members. Overall, household ART utilization did not decrease all-cause mortality (Hazard Ratio (HR) 0.95: 95% CI 0.65-1.4), however, restricting the analysis to a narrow CD4+ cell count range around the regression discontinuity threshold showed reduced all-cause mortality by 67% (HR 0.43: 95% CI 0.22-0.85) among household members of PLHIV on ART; the reduced risk was driven largely by the significant reduction noted among female household members (HR 0.21: 95% 0.08, 0.56).

Geographic Areas


  • thumnail for Saito_columbia_0054D_14761.pdf Saito_columbia_0054D_14761.pdf application/pdf 926 KB Download File

More About This Work

Academic Units
Thesis Advisors
Howard, Andrea A.
Ph.D., Columbia University
Published Here
June 24, 2018