Geographic Distribution of HIV-Stigma Among Women Of Child-Bearing Age In Rural Kenya

Abstract

Background: HIV-stigma is considered to be a major driver of the HIV/AIDS pandemic, yet there is a limited understanding of its epidemiology, especially at the structural/community level. Here we describe geographic patterns of two types of HIV-stigma in a population of women of child-bearing age: internalized-stigma (associated with shame) and externalized stigma (associated with blame), and explore whether individuals with similar attitudes towards people living with HIV are more likely to reside in the same geographic area. Methods: A cross-sectional sample of 405 women who gave birth within a one year period between January - December, 2010 was surveyed from the constituency of Gem, Kenya, one of three regions in the Western Kenya Health and Demographic Surveillance Area (HDSA), a 13 x 20 km region with a population of 220,000. Two forms of HIV-related stigma, self-reported HIV status, and other demographic variables were measured using a standardized, validated questionnaire. Latitude/Longitude coordinates of participants' residences were obtained with GPS devices. Residential locations of participants were compared with respect to whether or not individuals reported each form of stigma at different spatial scales using the K-function, a second order spatial data analysis used to measure spatial clustering of binary outcomes. Generalized additive models (GAMs) were used to assess whether spatial clustering of each stigma indicator occurred beyond that explained by the spatial patterns of individual-level characteristics such as age, income, education, and other socio-economic variables. Results: Among 373 women surveyed with complete GPS data, the median age was 25 years (IQR, 22-30 years), 12% self-reported positive HIV status, 45.5% reported at least one of three indicators of harboring internalized HIV-stigma (an indicator of shame) and 89.4% reported at least one of four indicators of harboring externalized stigma (an indicator of blame). There was strong evidence for a geographic trend in rates of externalized stigma among the respondents, with those who reported no form of externalized HIV-stigma being more likely to reside in the Southwestern portion of Gem compared to the Northeastern portion of the region, controlling for individual-level factors (p = 0.02). In contrast to blame, we did not observe spatial clustering for internalized stigma (shame) beyond that of complete spatial randomness (p = 0.36). Conclusions: The spatial trend observed for rates of externalized stigma compared to the random spatial distribution of internalized stigma may point to differences in the underlying social processes leading to each form of stigma. Externalized stigma may be driven more by dominant cultural beliefs disseminated within communities (i.e., churches, health facilities, or other leaders), whereas internalized stigma may be the result of individual-level characteristics outside the domain of community influence. Geographic studies of stigma can indicate higher risk areas and provide a first step in generating hypotheses as to potential community-level etiologies of stigma. Further data and hypothesis-testing is needed on community-level attributes that might promote lower rates of externalized stigma or `high tolerance' areas. These data may inform community-level interventions to decrease HIV-related stigma.