To assess the impact of the 1996–2005 integrated community-based micronutrient and health (MICAH) programme on linear growth retardation (stunting) in Malawian preschool children living in rural areas.

Prospective study of three large-scale cross-sectional surveys conducted in 1996, 2000 and 2004 in MICAH and Comparison populations.

Rural areas in Malawi.

Preschool children (6·0–59·9 months) from randomly selected households (474 from the 1996 baseline survey; 1264 from 2000 MICAH areas; 1500 from 2000 Comparison areas; 1959 from 2004 MICAH areas; and 1008 from 2004 Comparison areas), who responded to a household questionnaire, were weighed and measured using standard protocols.

At the baseline in 1996, the prevalence of stunting (60·2 %) was very high. By 2000, the prevalence of stunting had declined to 50·6 % and 56·0 % (χ2 = 7·8, P = 0·005) in MICAH and Comparison areas, respectively. In 2004, the prevalence of stunting did not differ significantly between MICAH and Comparison areas (43·0 % v. 45·1 %; χ2 = 1·11, P = 0·3). Severe stunting affected 34·7 % of children at baseline, which declined to 15·8 % and 17·1 % (χ2 = 0·86, P = 0·4) in MICAH and Comparison areas, respectively, by 2004. Regional variations existed, with proportionately fewer children from the Northern region being stunted compared to their Central and Southern region counterparts.

Given the length of implementation, wide-scale coverage and positive impact on child growth in Phase I (1996–2000), the MICAH programme is a potential model for combating linear growth retardation in rural areas in Malawi, although the catch-up improvement in Comparison areas during Phase II (2000–2004) cannot be adequately explained.