Measuring local determinants of acute malnutrition in Chad: a case-control study

BackgroundIn 2010, Guéra District, Chad recorded a prevalence of global acute malnutrition of 16·1% and severe acute malnutrition (SAM) of 4·9%, both signs of a critical nutritional situation. To improve malnutrition, a problem-tree analysis is normally used to propose causal factors. We used a case-control study and qualitative methods to determine causes of SAM and their effect sizes with an aim of providing evidence for programming.MethodsA matched case-control study was conducted within a coverage assessment (the semi-quantitative evaluation of access and coverage [SQUEAC]) of a therapeutic feeding programme. Qualitative information on the determinants of SAM was translated into hypotheses and standardised indicators to be evaluated in the case-control study. Villages were selected using the centric systematic area sampling method. Within villages, all current or recovering SAM cases among children 6–59 months were found through snowball sampling with seed informants who had knowledge of children's health. Two to three controls meeting the matching criteria (within ±3 months and same neighbourhood) were selected. Interviews with carers were conducted with a structured questionnaire. A sample of 40 cases and two controls per case was sufficient to detect a fourfold increase in odds. Conditional logistic regression with backwards stepwise elimination of non-significant variables was used to evaluate potential causal factors.Findings62 cases and 157 controls were enrolled. The odds of reporting fever and diarrhoea in the previous 2 weeks was eight (odds ratio 7·55, 95% CI 2·64–21·62) and eleven times (10·72, 4·27–26·88) higher among SAM cases.InterpretationDiarrhoea and fever represent multiple causes that contribute, with other factors, to the development of SAM and warrant a specific focus for programmes. Preventative recommendations include hygiene promotion and the routine use of bednets for under-5s. Identification of the timing of the onset of disease and malnutrition is difficult and we may only show association and not cause. The ability of the matched study to detect differences among factors that do not vary at the community level (eg, use of an unprotected water source) needs to be considered.FundingHumanitarian Aid and Civil Protection department of the European Commission (ECHO) and the USAID Office of Foreign Disaster Assistance (OFDA).