Spatial Analysis and Contextual Correlates of Health Risk Behaviors among Unmarried Women in Sub-Saharan Africa.

Chukwuedozie Ajaero, University of Nigeria Nsukka(UNN)
Oliver Razum, Bielefeld University, Germany

We estimated the contextual correlates of health risk behaviours (HRBs) in sub-Sahara Africa using demographic and health surveys from 14 countries. Data analysis involved descriptive statistics, hotspot analysis using spatial analysis and multilevel regression. The prevalence of HRBs ranged from 26% in Nigeria to 59% in Zimbabwe and Liberia. The results also revealed cold and hot spots of HRBs across sub-national levels in the region. We found significantly increased or decreased odds of HRBs in all the countries except for Sierra Leone (OR =0.89). Countries with increased likelihoods of HRBs are Liberia (OR=1.42), Cameroun (OR=1.14) and Zimbabwe (OR=1.44). The models of the contextual-level variables generally maintained the pattern in null models but more importantly had more countries with significantly increased odds of HRBs. Furthermore, the intercept of most of the models of individual-level factors and full models of HRBs showed significantly decreased likelihood of HRBs.The individual factors which either significantly increased or decreased odds of HRBs in most of the countries were age, and educational attainment. Our results underscore the importance of interventions that will consider the inequities of HRBs and the associated spatial, contextual and socioeconomic factors in each country

Keywords: Geographic Information Systems (GIS), Neighbourhood/contextual effect analysis, Health and Morbidity, Multi-level modeling

See extended abstract.