Estimating Effects of Poverty on the Survival of HIV Patients on ART and Food Supplementation in Rural Haiti: A Comparative Evaluation of Socio-Economic Indicators

Abstract

Background: Because economic conditions are both a risk factor for disease and may themselves be objectives for health delivery interventions, monitoring changes in economic outcomes has become a routine priority for health and development efforts. However, the lack of formal commerce in poor agrarian communities creates challenges for measuring economic status. Data on household finances, such as income, are ideal but are time-consuming, costly, and less reliable, whereas proxy measures of wealth such as indices of durable assets are easier to measure but relatively coarse and are less sensitive to rapid changes in underlying drivers. Methods: We used data from a cohort of 528 people living with HIV/AIDS (PLHA) enrolled in a food intervention study on household demographics, agricultural production, cash income, in-kind income, household durable assets and health status, including CD4 count. We created a household economic index using principal components analysis (PCA) and compared it with three other economic indicators generated from the data (income, expenditures, poverty score). Through multivariate logistic regression analysis we evaluated the effect of the economic metric on probability of survival within the first year of study. Results: Socioeconomic status determined by PCA of durable assets, weighted by the square root of the household size, was the only consistently significant economic predictor of probability of death. It remained significant even after controlling for direct health indicators such as CD4 count. There was no significant correlation between CD4 count and the economic indicators, which may be attributable to uniform access to ART among study participants. Conclusion: Among people who have HIV and are all enrolled in ART and food programs, household socioeconomic status is an important predictor of mortality rates, even after controlling for direct health measurements such as CD4 count and other health-related covariates. The SES indicator from PCA is also a simple metric to estimate. The study underscores that poverty is a social determinant of mortality even in the context of equal access to health services, and is suggestive of the importance of poverty alleviation activities as an important supplement to clinical interventions.