摘要：Objectives. To evaluate whether the quality of implementation of a water, sanitation, and hygiene program called SHEWA-B and delivered by UNICEF to 20 million people in rural Bangladesh was associated with health behaviors and sanitation infrastructure access. Methods. We surveyed 33 027 households targeted by SHEWA-B and 1110 SHEWA-B hygiene promoters in 2011 and 2012. We developed an implementation quality index and compared the probability of health behaviors and sanitation infrastructure access in counterfactual scenarios over the range of implementation quality. Results. Forty-seven percent of households (n = 14 622) had met a SHEWA-B hygiene promoter, and 47% of hygiene promoters (n = 527) could recall all key program messages. The frequency of hygiene promoter visits was not associated with improved outcomes. Higher implementation quality was not associated with better health behaviors or infrastructure access. Outcomes differed by only 1% to 3% in scenarios in which all clusters received low versus high implementation quality. Conclusions. SHEWA-B did not meet UNICEF’s ideal implementation quality in any area. Improved implementation quality would have resulted in marginal changes in health behaviors or infrastructure access. This suggests that SHEWA-B’s design was suboptimal for improving these outcomes. Scientists and development stakeholders argue that health programs proven effective in randomized efficacy trials should be translated into large-scale programs to benefit public health. 1 Substantial evidence supports the scaling up of numerous health programs. 2–4 Since the establishment of the Millennium Development Goals, the funding and motivation for scaling up has grown. 5 However, translating efficacious, small-scale programs to a large scale can present implementation challenges. 6,7 A growing body of literature documents barriers and facilitators to scaling up, yet there is little empirical evidence about how best to scale up. 7–12 In the water, sanitation, and hygiene (WASH) sector in particular, several evaluations of large-scale programs have found limited health impacts and incomplete program uptake. 13–17 As stakeholders consider scaling up other WASH programs, there is a scientific imperative to evaluate program impacts on health and to document reasons for intervention success or failure. 6 One of the largest WASH programs in history was the Sanitation Hygiene Education and Water Supply in Bangladesh (SHEWA-B) program, which was implemented by UNICEF and the Government of Bangladesh. SHEWA-B targeted approximately 20.4 million beneficiaries from 2007 to 2012. The program promoted hygiene practices and aimed to reduce diarrhea and WASH-related diseases among the poorest in rural Bangladesh. A 2009 interim assessment of SHEWA-B found little to no improvement in health behaviors (e.g., handwashing), access to hygiene or sanitation infrastructure, or prevalence of diarrhea and respiratory illness among children younger than 5 years. 18 These results could reflect a suboptimal program that needed to be better tailored to the target population or an appropriate program that needed to be better implemented. We conducted an observational study in SHEWA-B program areas to measure whether health behaviors and access to hygiene and sanitation infrastructure would have been better if SHEWA-B implementation had been of higher quality in all clusters. A positive association between implementation quality and health behaviors and access to hygiene and sanitation infrastructure would suggest that the SHEWA-B program had an appropriate design, even if it was imperfectly implemented, and a lack of association would suggest that the program needed to be better tailored to the population.
