Measurement of health literacy to advance global health research: a study based on Demographic and Health Surveys in 14 sub-Saharan countries

Abstract

Background: Education and literacy are key determinants of health, but these do not ensure health literacy. A definition of health literacy from the US Institute of Medicine (IOM) is widely accepted, but there is no consensus on the traits by which health literacy can be measured. Some questions in Demographic and Health Surveys (DHS) represent domains of health literacy as defined by the IOM (ie, the capacity to interpret, obtain, and understand health information; and the ability to make appropriate health decisions), and could be a valuable untapped resource for the measurement of health literacy. Building on pilot work in Zambia, we aimed to derive a measure of health literacy from information recorded in the DHS that aligned with the IOM definition. Methods: We accessed data from surveys conducted between 2006 and 2015 in 14 sub-Saharan African countries: Cameroon, Democratic Republic of the Congo, Ethiopia, Ghana, Guinea, Côte d'Ivoire, Lesotho, Rwanda, Niger, Namibia, Sierra Leone, Swaziland, Togo, and Zambia. Surveys were the same in all countries, but translated to local languages as appropriate. We used factor analysis to extract one measure of health literacy and assessed its reliability and validity. We reclassified the continuous measure into tertiles and collapsed observations in the lower two groups to derive a dichotomous indicator. We then used χ2 tests to assess whether the prevalence of (high) health literacy varied by demographics and country, accounting for sample weights. Findings: We included 259 684 survey respondents aged 15–49 years. We identified eight survey questions that corresponded to elements of the four domains of IOM-defined health literacy. The questions addressed school and reading comprehension (capacity to interpret); reading magazines or listening to the radio or TV (capacity to obtain); learning family planning information (capacity to understand); and knowledge of a place to get an HIV test or acquire condoms (make appropriate health decisions). The derived continuous measure of health literacy demonstrated internal consistency (Cronbach's α = 0·72), and good content validity. Using the dichotomous indicator, overall prevalence of health literacy was 35·2% (95%CI 34·9–36·6%). Health literacy varied by sex (women 34·1% vs men 39·2%, p<0·001) and education level (primary education or less 8·9%, some secondary education 69·4%, and completed secondary education or higher 84·4%, p<0·001). Health literacy also varied greatly between countries, from 8·5% in Niger to 63·9% in Namibia (p<0·001) Interpretation: Our study derives a robust indicator of IOM-defined health literacy in a large population in sub-Saharan Africa. Researchers could use this indicator with DHS datasets to measure health literacy in other countries, and ultimately examine how health literacy relates to health behaviour and outcomes, including for HIV/AIDS and domestic violence. Funding: None.