Articles 1460 www.thelancet.com Vol 366 October 22, 2005 Introduction In most low-income countries, coverage rates for child- survival interventions are low, and millions of children die every year from diseases for which there are effective interventions. 1,2 Furthermore, there are important inequities between different social groups in nearly all low-income and middle-income countries 3,4 and even within rural populations that may appear to be uniformly poor. 5 Children belonging to the poorest families are consistently less likely to receive preventive and curative interventions than those from other families. 6,7 In low-income countries various child-survival interventions are being implemented simultaneously. 1,2 These include preventive interventions such as vaccines, insecticide treated mosquito nets, micronutrient supplementation, nutrition counselling (breastfeeding and complementary feeding), growth monitoring, and appropriate newborn care. Additionally, health systems in most countries provide many case-management interventions, including oral rehydration therapy, antibiotics, and antimalarials. We assessed the joint distribution of key preventive interventions in children younger than 5 years and investigated how many separate child-survival interventions each child is receiving and whether this number differed by the sex of the child. We also studied the role of social inequities in co-coverage, and discuss here possible implications for planning the delivery of child-survival interventions. Methods Study samples We selected low-income countries for which Demographic and Health Surveys (DHS) 8 datasets obtained since 1999 were available for secondary analysis and included the variables needed for analysis of intervention coverage rates and equity. We attempted to include countries from different regions of the world. Brazil was also included, despite the fact that the survey data were from 1996, as an example of a middle-income country with high coverage of most child-survival inter- ventions. The data refer to national probability samples of children younger than 5 years from Bangladesh (1999/2000), Benin (2001), Brazil (1996), Cambodia (2000), Eritrea (2002), Haiti (2000), Malawi (2000), Nepal (2001), and Nicaragua (2001). 9–17 DHS rely on multi-stage sampling procedures, including region of the country, states or provinces, municipalities, census tracts, and households. Weighting was used to reproduce national samples. Details are available elsewhere. 8 Outcome variables We assessed coverage of preventive interventions for children age 12–59 months, including those living in the Lancet 2005; 366: 1460–66 Universidade Federal de Pelotas, CP 464, 96001-970 Pelotas, RS, Brazil (Prof C G Victora MD); London School of Hygiene and Tropical Medicine, London, UK (B Fenn MSc, Prof B R Kirkwood MA); and 2081 Danby Road, Ithaca, NY, USA (J Bryce EdD) Correspondence to: Prof Cesar Victora cvictora@terra.com.br Co-coverage of preventive interventions and implications for child-survival strategies: evidence from national surveys Cesar G Victora, Bridget Fenn, Jennifer Bryce, Betty R Kirkwood Summary Background In most low-income countries, several child-survival interventions are being implemented. We assessed how these interventions are clustered at the level of the individual child. Methods We analysed data from Bangladesh, Benin, Brazil, Cambodia, Eritrea, Haiti, Malawi, Nepal, and Nicaragua. A co-coverage score was obtained by adding the number of interventions received by each child (including BCG, diphtheria-pertussis-tetanus, and measles vaccines), tetanus toxoid for the mother, vitamin A supplementation, antenatal care, skilled delivery, and safe water. Socioeconomic status was assessed through principal components analysis of household assets, and concentration indices were calculated. Findings The percentage of children who did not receive a single intervention ranged from 0·3% (14/5495) in Nicaragua to 18·8% (1154/6144) in Cambodia. The proportions receiving all available interventions varied from 0·8% (48/6144) in Cambodia to 13·3% (733/5495) in Nicaragua. There were substantial inequities within all countries. In the poorest wealth quintile, 31% of Cambodian children received no interventions and 17% only one intervention; in Haiti, these figures were 15% and 17%, respectively. Inequities were inversely related to coverage levels. Countries with higher coverage rates tended to show bottom inequity patterns, with the poorest lagging behind all other groups, whereas low-coverage countries showed top inequities with the rich substantially above the rest. Interpretation The inequitable clustering of interventions at the level of the child raises the possibility that the introduction of new technologies might primarily benefit children who are already covered by existing interventions. Packaging several interventions through a single delivery strategy, while making economic sense, could contribute to increased inequities unless population coverage is very high. Co-coverage analyses of child-health surveys provide a way to assess these issues.