Participants were recruited in September 2002, using a two-stage cluster sampling methodology. A structured questionnaire was administered to mothers in their home settings. Information on health, household socio-economic status, child feeding practices and anthropometric measurement was gathered. Conditional logistic regression analysis was applied taking into account the hierarchical relationships between potential determinants of poor anthropometric status.
The mean Z-score for weight-for-height was −0.2 (95% confidence interval (CI) −0.1, −0.7), for height-for-age was −1.1 (95% CI −1.2, −0.9) and for weight-for-age was −0.7 (95% CI −0.8, −0.6). Wasting was independently associated only with a history of fever in the 2 weeks prior to the survey (odds ratio (OR) = 4.4, 95% CI 1.5, 13), while underweight was associated with a history of fever (OR = 2.4, 95% CI 1.3, 4.4) and cough (OR = 3.0, 95% CI 1.3, 6.8). Stunting was positively associated with a wider range of factors, including: history of a fever episode (OR = 1.7, 95% CI 1.0, 2.9), lack of a latrine in the household (OR = 2.7, 95% CI 1.5, 4.9), failure to de-worm children 12 months or older (OR = 1.7, 95% CI 1.1, 2.8), and being born to a non-formally educated mother compared with mothers educated above primary school (OR = 2.1, 95% CI 1.1, 4.0).
In analyses guided by the hierarchical interrelationships of potential determinants of malnutrition, wasting and underweight turned out to be independently predicted by morbidity (proximal) factors. Stunting, however, was predicted by socio-economic (distal), environmental and health-care (intermediate) factors in addition to morbidity. Strategies aimed at improving the growth of infants and young children in rural communities should address morbidity due to common childhood illness coupled with environmental and socio-economically oriented measures.