Geospatial pattern of HIV seropositivity and its predictors among women in Ethiopia. A spatial and multiscale geographically weighted regression analysis

Abstract

Background: Although promising efforts have been made so far, HIV remains a public health concern. Women in Ethiopia are disproportionately affected by HIV, accounting for a majority of new infections and AIDS-related deaths. However, the geospatial distribution of HIV among women in Ethiopia is not well understood, making it challenging to develop geographically targeted measures. Besides, to accelerate the pathway of decreasing HIV prevalence and plan geographically specific interventions, understanding the geospatial distribution of HIV seropositivity and its predictors among women plays a significant role.

Methods: A spatial and multiscale geographically weighted regression analysis was conducted using the 2016 EDHS dataset, comprising 14,778 weighted samples of women in the reproductive age group. The EDHS sample underwent two-stage stratification and selection. The data were extracted between October 18 and 30, 2023. Non-spatial analysis was carried out using STATA version 17. Additionally, ArcGIS Pro and Sat Scan version 9.6 were used to visually map HIV seropositivity. Global Moran's I was computed to evaluate the distribution of HIV seropositivity. The Getis-Ord Gi* spatial statistic was utilized to identify significant spatial clusters of cold and hot spot areas. Geographically weighted regression analysis was subsequently performed to identify significant predictors of HIV seropositivity. Significance was established at a P-value <0.05 throughout all statistical analyses.

Results: HIV seropositivity among women in Ethiopia is distributed non-randomly (Global Moran's I = 0.16, p-value <0.001 and Z-score = 7.12). Significant hotspot clustering of HIV seropositivity was found in the Addis Ababa, Harari, Dire Dawa, and Gambela region. Poor wealth index, being divorced and widowed, having more than one sexual partner, and early first sexual experience (<15 years) were found to be predictors of geographical variation of HIV seropositivity among women.

Conclusion: HIV seropositivity among women in Ethiopia varies geographically. Thus, deploying additional resources in high hotspot regions is recommended. Programs should focus on improving the economic empowerment of women to prevent the from engaging in risky sexual behaviors. Furthermore, comprehensive sex education programs in schools and community settings regarding the consequences of early first sexual debut might play a role in reducing HIV seropositivity among women in Ethiopia.

Fig 1. Schematic representation of the sampling procedures used to study the geospatial pattern of HIV seropositivity and its predictors among women in Ethiopia, EDHS 2016.

N.B EAs = Enumeration areas.

Fig 2. Spatial autocorrelation analysis of HIV seropositivity among women in Ethiopia.

Fig 3. Incremental autocorrelation analysis of HIV seropositivity among women in Ethiopia.

Fig 4. Hot spots analysis of HIV seropositivity among women in Ethiopia.

Fig 5. Ordinary kriging interpolation to predict HIV seropositivity prevalence among women in Ethiopia.

Fig 6. Spatial clustering of high and low rate of HIV seropositivity among women in Ethiopia.

Fig 7

The spatial mapping of geographically weighted regression coefficients by poor wealth index (A) and early sexual experience (B) to predict the hotspot of HIV seropositivity among women in Ethiopia.