Food-based nutrition education and hygiene can improve the growth of stunted children.

The paper by Salehi et al. (2004) in the current issue of the British Journal of Nutrition is a welcome and rather rare example of an evaluation of a successful nutrition-education trial targeting poorly nourished children. There are few community-based effectiveness trials that have succeeded in improving growth, and almost all such trials have been reported in technical reports and proceedings of meetings that are not widely available and seldom peer-reviewed (Griffiths et al. 1996; Allen & Gillespie, 2001). The approach and results in this study of Qashqa’i tribal people in Iran provide an outline and context for addressing several issues related to nutritional interventions to improve growth. The authors chose to start with the beliefs, attitudes, subjective norms and enabling factors (BASNEF) approach of Hubley (1988), which they modified to include a ‘knowledge’ component. In their intervention, the authors relied heavily on the support of tribal elders and teachers, who served as the ‘influential people’ specified in Hubley’s model, as keys to addressing social pressures and providing ‘enabling factors’ such as tribal cooperation and literate helpers. The authors relied heavily upon literate daughters, who delivered much of the intervention. The BASNEF approach has not been widely applied elsewhere, but it does contain many of the practical components seen in other approaches (Allen & Gillespie, 2001). Key educational messages included information on: water, waste, and food sanitation; growth monitoring; appropriate complementary foods and feeding; child requirements for specific foods; food preparation; increased access to fruits and vegetables; increased uses for fruits, vegetables and pulses, especially lentils; referral to health centres. Notably absent from the intervention components was provision of food supplements directly to mothers or children. Consequently, the impacts of the intervention resulted from behavioural responses of the mothers and caregivers to the intervention messages and adoption of the health-related practices as the new group norms during the study period. Using the mean weight-for-age Z-score (WAZ) and the mean age, the mean weight of the children at baseline was approximately 11·5 kg. Accordingly, the mean dietary energy intake (exclusive of breast milk) was approximately 490 J (117 kcal)/kg per d and probably adequate for most of the children (Butte, 1996; Torun et al. 1996). The initial average protein intake (exclusive of breast milk) was approximately 34·7 g/d (with 28 % from animal sources), so the total protein intake was approximately 3 g/kg per d. Consequently, it seems unlikely that the Qashqa’i children as a group were protein-deficient at baseline (Dewey et al. 1996). Even so, because the exact age, body weight and intake distributions were not presented, there may have been some children who had less energy and/or protein than the estimated requirements. Upon inspection of Tables 6 and 7, one sees that the intervention resulted in net mean daily increases of 20 g lentil flour, 20 g beans, 10 g milk, 29 g egg, 30 g fresh vegetables, 10 g dried vegetables and 100 g seasonal fruit. In addition, after 1 year of the intervention there were net decreases in estimated mean daily intakes of wheat flour (30 g) and sugar (15 g), presumably due to food substitutions. The corresponding intervention-related net increases in macronutrients reported were 4·4 g animal protein (from 9·7 g at baseline) and 5·9 g vegetable protein (from 25·0 g at baseline). The intervention did not change the estimated mean daily intake of total dietary energy (5 583 J (1 335 kcal)). Thus, the intervention added new plant foods to the diet (lentils, beans, vegetables, fruits) and increased intakes of milk and eggs, without altering average energy intake. It seems most likely that the chief nutritional benefits of the intervention were due to multiple micronutrients. The additional foods would have enhanced intakes of a wide range of micronutrients, including Fe, Zn, Mg, retinol, folate, carotenoids, riboflavin, and vitamins C and B12. Moreover, the added vitamin C from the fruits would have increased the availability of Fe (Allen, 1998), and the combination of foods may have had other synergistic effects (Jacobs & Steffen, 2003). At baseline the mean height-for-age Z-score (HAZ) for the intervention or ‘test’ group in the study of Salehi et al. (2004) was 22·1. This corresponds to a prevalence of stunting (HAZ ,22) of approximately 54 %, which is considered very high compared with other developing countries (World Health Organization, 1995). Mean WAZ at baseline was 21·4, corresponding to a prevalence of approximately 27 % underweight children (WAZ ,22), a level considered high (World Health Organization, 1995). The starting points are important because higher prevalences of stunting and underweight are associated with larger responses to nutrition interventions (Caulfield et al. 1999). The net intervention-related gain in length or height (intervention mean minus control mean) was 0·0163 m or HAZ 0·41 for the Qashqa’i children, and that for weight was 0·74 kg or WAZ 0·45. Because the children ranged from 0–59 months old at baseline, the Z-scores (adjusted for gender and age) are more easily interpreted than the absolute metric gains. The net reductions in the prevalences of stunting and underweight British Journal of Nutrition (2004), 91, 657–659 DOI: 10.1079/BJN20041127 q The Authors 2004