To investigate and compare feeding practices among infants of less than 7 months of age in a rural and an urban area in Tanzania.

Cross-sectional, questionnaire-based interview of mothers and focus group discussions with extension workers and community leaders.

Eleven villages in a rural district and 10 wards in an urban district in the Morogoro region, Tanzania, west of Dar es Salaam.

Probability samples of mothers with infants of less than 7 months of age (n = 320 from each area).

Exclusive breast-feeding was rarely practised in either the rural or urban areas investigated. However, the urban mothers initiated breast-feeding earlier, discarded colostrum less frequently, breast-fed exclusively for a longer period, gave breast milk as the first feed more often and delayed the introduction of solid foods for longer than their rural counterparts. The rural mothers, on the other hand, breast-fed their previous infants slightly longer than the urban mothers.

The better performance of urban mothers could be partly due to sustained breast-feeding support in hospital settings and other campaigns which may not have reached the rural areas. In both the rural and urban areas more efforts are needed to encourage exclusive breast-feeding, to avoid premature complementation and, in the case of the urban areas, to protect extended breast-feeding.

To evaluate the long-term effects of a horticultural and nutrition education intervention in rural Tanzania.

A quasi-experimental post-test design was used.

The research was carried out in 10 villages in Singida region, Tanzania.

Mothers and their children aged 6–71 months (n = 236) from an experimental (Ilongero) and control (Ihanja) area were interviewed regarding knowledge and practices related to vitamin A nutrition. Intake of vitamin A-rich foods by the children during the 7 days prior to the interview was recorded. Stools were examined for helminths and serum samples were analysed for retinol and C-reactive protein (CRP) (n = 146) for the children aged 12–71 months.

Knowledge and practices were more favourable to vitamin A intake in the experimental area than in the control area, and an increased frequency of intake of green leaves was associated with higher serum retinol values. The experimental area had lower mean serum retinol levels (13.7 μg dl−1, n = 75) than the control area (19.3 μg dl−1, n = 71). One likely confounder was the higher helminth infestation in the experimental area (n = 75, 79%) than in the control area (n = 71, 49%) (P > 0.001). Children with helminths (n = 94) had a lower mean serum retinol level than those without (n = 52) (12.3 ± 5 vs. 24 ± 10 μg dl−1; P = 0.001).

Food-based vitamin A programmes can make sustainable improvements in knowledge and dietary practices but these may not necessarily be reflected in increases in serum retinol. Programme implementation and evaluation should take confounders into consideration as, in this case, helminth infestation.