Participatory women's groups: ready for prime time?

www.thelancet.com Vol 381 May 18, 2013 1693 Conventional thinking behind global eff orts to prevent the deaths of mothers and children younger than 5 years starts with the identifi cation of an intervention with a clear biological pathway for reducing morbidity and mortality, such as a drug, vaccine, micronutrient, or health-related behaviour. The effi cacy of this intervention is then tested in randomised trials, followed by meta-analyses of the trials’ fi ndings. If results are encouraging, the intervention is scaled up, most frequently by inte gration with existing delivery platforms such as facility or community-based health services, or mass campaigns. The innovative approach taken by Audrey Prost and colleagues from the Institute of Child Health, London, UK, and from four low-income and middle-income countries, in The Lancet, brings a breath of fresh air to our fi eld of knowledge. Based on the ideas of conscientisation— or creating critical consciousness—pioneered by the Brazilian educator Paulo Freire, they promoted the generic empowerment of women in poor communities, rather than recommending specifi c interventions or behaviours. People who promote somewhat ethereal ideas such as empowerment are seldom the same people as those who are keen on doing randomised controlled trials. The greatest accomplishment of this group of investigators is to combine both. Prost and colleagues present a meta-analysis of seven randomised controlled trials of women’s groups carried out in Malawi, India, Bangladesh, and Nepal, and describe reductions in rates of maternal mortality (odds ratio 0·66, 95% 0·33–0·99) and neonatal mortality (0·77, 0·65–0·89) associated with exposure to women’s groups. Although the results of the meta-analysis are in general convincing, they also leave many unanswered questions. The sizeable estimates of reduction in maternal and newborn deaths (albeit with large 95% CI) are in contrast with the paucity of evidence of any improvement in coverage for most interventions that we would expect to mediate such an eff ect. In Prost and colleagues’ study, several of the original investigators did not measure potential changes in interventions that might have led to an eff ect. In some studies, and in the meta-analysis, the proportion of deaths prevented was often larger than that of pregnant women who participated—even if only once—in group meetings, suggesting some type of herd immunity that is hard to understand. Also, at least six of the seven studies were undertaken by the same group of researchers—and, although their methods seem to be thorough, it is always reassuring to have independent confi rmation by other investigators. Interventions with a strong behaviour-modifi cation component often work well at small scale when closely Participatory women’s groups: ready for prime time? *Karen A Grépin, Jeni Klugman Robert F Wagner Graduate School of Public Service, New York University, New York, NY 10012, USA (KAG); and The World Bank, Washington, DC, USA (JK) [email protected]