The Growing Problem of Obesity in Sub-Saharan Africa: A Multilevel Study Using Data for 250,000 Women from 380 Provinces in 36 Countries

Janine Huisman, Radboud University Nijmegen
Jeroen Smits, Radboud Universiteit Nijmegen

Obesity, once considered a problem of developed countries, is now increasingly becoming a problem in less developed countries as well. Given the substantial health risks and economic costs associated with obesity, it is of great importance to gain understanding of the determinants of this problem. In this paper we look at these factors for women in sub-Saharan Africa, one of the least developed parts of the world, where overweight and obesity are growing quickly. For this paper we use data from the Global Data Lab (www.globaldatalab.org) for 250,000 women from 380 sub-national regions (provinces) in 36 sub-Saharan African countries. Our dependent variable is the Body Mass Index (BMI) of non-pregnant women aged 18-49. Our main independent variables are at the individual level (age, education, urban area, wealth) and sub-national level (level of development of the province, percentage of households with internet in the province and percentage of women with paid employment in the province). To determine how the effect of the various individual factors differs according to the circumstances in which the women live, we also include interactions between our individual and context characteristics.

Keywords: Health and Morbidity, Multi-level modeling , Neighbourhood/contextual effect analysis, Big data

See extended abstract.