Portal de Periódicos da CAPES

Implications of WASH Benefits trials for water and sanitation – Authors' reply

2018; Elsevier BV; Volume: 6; Issue: 6 Linguagem: Inglês

10.1016/s2214-109x(18)30229-8

2572-116X

Benjamin F. Arnold, Clair Null, Stephen P. Luby, John M. Colford,

Global Maternal and Child Health

We appreciate the thoughtful comments from Oliver Cumming and Val Curtis and from Diane Coffey and Dean Spears regarding the Kenya and Bangladesh WASH Benefits trials.1Luby SP Rahman M Arnold BF et al.Effects of water quality, sanitation, handwashing, and nutritional interventions on diarrhoea and child growth in rural Bangladesh: a cluster randomised controlled trial.Lancet Glob Health. 2018; 6: e302-e315Summary Full Text Full Text PDF PubMed Scopus (344) Google Scholar, 2Null C Stewart CP Pickering AJ et al.Effects of water quality, sanitation, handwashing, and nutritional interventions on diarrhoea and child growth in rural Kenya: a cluster-randomised controlled trial.Lancet Glob Health. 2018; 6: e316-e329Summary Full Text Full Text PDF PubMed Scopus (318) Google Scholar Since both trials took place in populations with relatively low levels of open defecation at enrolment, we agree that the global health community should be cautious about transporting effect estimates from the trials to populations with high levels of open defecation, or to populations in urban environments with vastly different conditions. These trials were done in populations that were similar to much of rural Bangladesh and Kenya, and were chosen explicitly because of the high burden of both linear growth faltering and diarrhoea.3Arnold BF Null C Luby SP et al.Cluster-randomised controlled trials of individual and combined water, sanitation, hygiene and nutritional interventions in rural Bangladesh and Kenya: the WASH Benefits study design and rationale.BMJ Open. 2013; 3: e003476Crossref PubMed Scopus (147) Google Scholar We anticipate that the results will be generalisable to many similar rural populations with persistent growth faltering. Forthcoming trial results from Zimbabwe4Sanitation Hygiene Infant Nutrition Efficacy (SHINE) Trial TeamThe Sanitation Hygiene Infant Nutrition Efficacy (SHINE) trial: rationale, design, and methods.Clin Infect Dis. 2015; 61: S685-S702Crossref PubMed Scopus (110) Google Scholar and Mozambique5Brown J Cumming O Bartram J et al.A controlled, before-and-after trial of an urban sanitation intervention to reduce enteric infections in children: research protocol for the Maputo Sanitation (MapSan) study, Mozambique.BMJ Open. 2015; 5: e008215Crossref PubMed Scopus (45) Google Scholar will complement the WASH Benefits trials by providing evidence of the effect of water, sanitation, and handwashing (WASH) interventions on growth in populations with high baseline levels of open defecation, and—in the case of Mozambique—in a high-density, urban setting. The letters propose that WASH interventions could have possibly improved child growth had we fielded the trials in populations with higher levels of open defecation or in populations with worse drinking water sources. Yet, linear growth faltering prevailed: average length-for-age z-scores (LAZ) in control groups at the final endpoint were −1·54 in Kenya and −1·79 in Bangladesh. Low average LAZ despite low levels of open defecation and access to improved water sources for the majority of people in both populations show that prenatal or postnatal exposures, or both, beyond open defecation and water source are important determinants of linear growth faltering.6Black RE Victora CG Walker SP et al.Maternal and child undernutrition and overweight in low-income and middle-income countries.Lancet. 2013; 382: 427-451Summary Full Text Full Text PDF PubMed Scopus (4258) Google Scholar In Kenya, water supply remained a challenge because few participants had piped water to their homes, but in Bangladesh tube wells within household compounds were ubiquitous, suggesting that adequate water supply alone will be insufficient to prevent growth faltering. Both letters suggest that a more comprehensive, community-level approach to improving the environment might be necessary to influence child growth. In theory, this is certainly possible, but the trials delivered compound-level interventions because formative research in rural Bangladesh and sub-Saharan Africa showed that, among children younger than 18 months, exposure to faecal contamination occurs primarily within the compound.7Kwong LH Ercumen A Pickering AJ Unicomb L Davis J Luby SP Hand- and object-mouthing of rural Bangladeshi children 3–18 months old.Int J Environ Res Public Health. 2016; 13: 563Crossref Scopus (52) Google Scholar, 8Mbuya MNN Tavengwa NV Stoltzfus RJ et al.Design of an intervention to minimize ingestion of fecal microbes by young children in rural Zimbabwe.Clin Infect Dis. 2015; 61: S703-S709Crossref PubMed Scopus (30) Google Scholar Despite delivering intensive compound-level WASH interventions, it remains possible that the trials did not reduce faecal exposure among children enrolled in the study sufficiently to influence growth through the hypothesised subclinical pathways,9Humphrey JH Child undernutrition, tropical enteropathy, toilets, and handwashing.Lancet. 2009; 374: 1032-1035Summary Full Text Full Text PDF PubMed Scopus (512) Google Scholar despite improving many other outcomes. High diarrhoea prevalence in the Kenya trial2Null C Stewart CP Pickering AJ et al.Effects of water quality, sanitation, handwashing, and nutritional interventions on diarrhoea and child growth in rural Kenya: a cluster-randomised controlled trial.Lancet Glob Health. 2018; 6: e316-e329Summary Full Text Full Text PDF PubMed Scopus (318) Google Scholar and widespread enteric pathogen infection in Bangladesh10Lin A Ercumen A Benjamin-Chung J et al.Effects of water, sanitation, handwashing, and nutritional interventions on child enteric protozoan infections in rural Bangladesh: a cluster-randomized controlled trial.Clin Infect Dis. 2018; (published online April 13.)DOI:10.1093/cid/ciy320Crossref Scopus (40) Google Scholar and Kenya (Pickering AJ, Tufts University, personal communication) reflect high levels of transmission. Environmental measurements in the Bangladesh trial documented widespread faecal contamination that was strongly associated with the presence of animals and their faeces.11Ercumen A Pickering AJ Kwong LH et al.Animal feces contribute to domestic fecal contamination: evidence from E. coli measured in water, hands, food, flies, and soil in bangladesh.Environ Sci Technol. 2017; 51: 8725-8734Crossref PubMed Scopus (116) Google Scholar Forthcoming results from both trials will summarise intervention effects on enteric pathogens and on faecal contamination throughout the children's environment, including complementary foods. Well designed and conducted randomised trials answer specific questions with high validity—a feature that is at once valuable and limiting. It will never be possible to do randomised trials in every setting, and fielding a randomised trial that delivers even more intensive environmental interventions than WASH Benefits to entire communities rather than compounds would probably be logistically and financially prohibitive. Observational analyses could potentially help fill the evidence gap. Yet, a re-analysis of the trials leads us to urge the global community to be cautious when interpreting observational analyses of the effects of sanitation on child growth, similar to those presented by Coffey and Spears. Inspired by an analysis that the SHINE investigators4Sanitation Hygiene Infant Nutrition Efficacy (SHINE) Trial TeamThe Sanitation Hygiene Infant Nutrition Efficacy (SHINE) trial: rationale, design, and methods.Clin Infect Dis. 2015; 61: S685-S702Crossref PubMed Scopus (110) Google Scholar presented at the American Society for Tropical Medicine and Hygiene 2017 conference, we re-analysed data from the WASH Benefits trials to estimate the difference in LAZ associated with improved sanitation access at enrolment among children born into the control group—creating an observational, prospective cohort nested within each trial. Among children in the control group, improved sanitation was associated with 0·15 LAZ increase in Kenya (p=0·02) and 0·22 LAZ increase in Bangladesh (p=0·02) in adjusted, double-robust analyses (table). The inconsistency between the observational analyses and null effects in the trials, estimated in the same study populations, illustrates the danger of bias from unmeasured confounding in observational studies, which has been shown in many other examples.12Pocock SJ Elbourne DR Randomized trials or observational tribulations?.N Engl J Med. 2000; 342: 1907-1909Crossref PubMed Scopus (472) Google Scholar It also calls into question whether the observed associations between sanitation conditions and linear growth in India are causal. Sanitation facilities and open defecation practices are inextricably tied to many improvements in overall wellbeing. This cautionary example highlights the value of randomised trials for measuring the effects of exposure–outcome relationships that are deeply entwined with broader socioeconomic development. Nevertheless, we feel strongly that these findings should not diminish ongoing, ambitious efforts to achieve the UN Sustainable Development Goals (SDGs): myriad health, equity, and ethical arguments motivate elimination of open defecation and ample supply of microbiologically safe water, even in the absence of a strong link to child growth.TableLAZ among children in the control groups of the WASH Benefits trials in Kenya and Bangladesh, stratified by whether the child's household had improved sanitation at enrolmentPopulation (n)Mean LAZ (SD)Difference (95% CI)p valueAdjusted*Adjusted by use of ensemble machine learning with double-robust, targeted maximum likelihood estimation following the same methods from the prespecified adjusted analyses in the trials. Prespecified, baseline covariates included: child age, child sex, household food insecurity, birth order, maternal age, maternal education, maternal height, number of children and total individuals living in the compound, distance to water, and a broad set of household characteristics and assets. The computational notebook that created the table includes additional analysis details, plus adjusted effects using generalised linear models that resulted in similar estimates (https://osf.io/qkgp8). Data used to make the table are available on the Open Science Framework website for Bangladesh (https://osf.io/wvyn4) and Kenya (https://osf.io/uept9). difference (95% CI)p valueKenya trial control group†In the Kenya trial, improved sanitation was defined as the presence of a latrine with a slab following the standard WHO/UNICEF Joint Monitoring Program definition. In the Bangladesh trial, improved sanitation was defined as a toilet with a functional water seal. These definitions mirrored those reported in the original trials.No improved latrine1737−1·58 (1·08)refrefAccess to improved latrine364−1·33 (1·08)0·25 (0·12–0·37)<0·0010·15 (0·02–0·28)0·02Bangladesh trial control group†In the Kenya trial, improved sanitation was defined as the presence of a latrine with a slab following the standard WHO/UNICEF Joint Monitoring Program definition. In the Bangladesh trial, improved sanitation was defined as a toilet with a functional water seal. These definitions mirrored those reported in the original trials.No latrine513−1·89 (0·98)refrefLatrine with no water seal391−1·86 (1·00)....Latrine has functional water seal199−1·37 (1·01)0·52 (0·34–0·70)<0·0010·22 (0·03–0·40)0·02Median age 25 months for Kenya trial and 22 months for Bangladesh trial. LAZ=length-for-age z-scores.* Adjusted by use of ensemble machine learning with double-robust, targeted maximum likelihood estimation following the same methods from the prespecified adjusted analyses in the trials. Prespecified, baseline covariates included: child age, child sex, household food insecurity, birth order, maternal age, maternal education, maternal height, number of children and total individuals living in the compound, distance to water, and a broad set of household characteristics and assets. The computational notebook that created the table includes additional analysis details, plus adjusted effects using generalised linear models that resulted in similar estimates (https://osf.io/qkgp8). Data used to make the table are available on the Open Science Framework website for Bangladesh (https://osf.io/wvyn4) and Kenya (https://osf.io/uept9).† In the Kenya trial, improved sanitation was defined as the presence of a latrine with a slab following the standard WHO/UNICEF Joint Monitoring Program definition. In the Bangladesh trial, improved sanitation was defined as a toilet with a functional water seal. These definitions mirrored those reported in the original trials. Open table in a new tab Median age 25 months for Kenya trial and 22 months for Bangladesh trial. LAZ=length-for-age z-scores. We declare no completing interests. Effects of water quality, sanitation, handwashing, and nutritional interventions on diarrhoea and child growth in rural Kenya: a cluster-randomised controlled trialBehaviour change messaging combined with technologically simple interventions such as water treatment, household sanitation upgrades from unimproved to improved latrines, and handwashing stations did not reduce childhood diarrhoea or improve growth, even when adherence was at least as high as has been achieved by other programmes. Counselling and supplementation in the nutrition group and combined water, sanitation, handwashing, and nutrition interventions led to small growth benefits, but there was no advantage to integrating water, sanitation, and handwashing with nutrition. Full-Text PDF Open Access