Deconstructing the complexity of measuring food security in South Africa : a systematic review and meta-analysis (2000-2024)

Abstract

BACKGROUND : Measuring the non-observable nature of food security has remained complex mainly because of the construct’s complexity and its continuously evolving nature. The main challenges in measuring food security involve determining what is to be measured and how it is measured. In South Africa, various approaches and indicators have led to divergent food security measurement outcomes, leading to inaccurate assessment, monitoring, and targeting of context-specific food security interventions. This study analyses food access, availability, and stability measurement metrics and proposes a clear food security measurement approach for South Africa. Comprehensive reviews of food security indices with a national scope and subsequent meta-analysis to determine these indicators’ effect size, publication bias, and heterogeneity have not been adequately explored. METHODS : A systematic review and meta-analysis using the PRISMA guidelines were used to select the analysed articles. A search strategy was used to retrieve literature from the Web of Science and Scopus Databases, yielding a total of 1155 articles. Rayyan 1.6 software was used for screening articles and duplicate removal, whilst the Newcastle-Ottawa scale was used to qualitatively assess the articles. Perplexity and Quill Bot Artificial Intelligence (AI) tools were used to enhance literature search and paraphrasing, respectively, to improve the validity and reliability of the review. The inclusion and exclusion criteria resulted in a final sample of 82 articles being eligible for analysis. RESULTS : Most studies used Household Food Insecurity Access Score (HFIAS), (n = 45), Household Dietary Diversity Score (HDDS) (n = 24), Coping Strategy Index (CSI), (n = 13), and the Household Hunger Score (HHS) (n = 4). Few studies used a composite of indicators, while most studies used HFIAS alone. The indicators used provide very different estimates of the prevalence of food insecurity in South Africa. Limpopo and KwaZulu-Natal provinces had the most studies distributed across rural communities, while Cape Town City and Gauteng Province had the highest number of urban studies. Meta-analysis was done on HFIAS (n = 16) and HDDS (n = 14) indicators using a Forest plot and Funnel plot, and results showed limited heterogeneity and publication bias across the studies. CONCLUSIONS : More food security studies need to use longitudinal designs, composite indicators across different seasons, and along the urban, peri-urban, and rural settlement gradient, and panel data from national surveys. The routine national surveys need to adopt the full modules of indicators to allow for household and individual food security analysis in South Africa. We recommend the Agency Module, the Women Empowerment in Agriculture Index (WEAI), and the Women Empowerment in Livestock Index (WELI) to measure the sustainability and agency dimensions of food security.