Impact Of Wash On Key Social And Health Outcomes

EVIDENCE PAPER The Impact of WASH on Key Health & Social OutcomesIntroduction AimsThis paper was commissioned by UNICEF and undertaken by theDFID-funded Sanitation and Hygiene Applied Research for Equity(SHARE) research programme consortium.This evidence paper aims to provide evidence for specific elementsof UNICEF’s forthcoming WASH Strategy, 2016-2030. In particular, itseeks to present the evidence on the importance of WASH to otheroutcomes beyond child diarrhoea.A key rationale for investing in WASH is the importance of WASHto other Sustainable Development Goal (SDG) outcomes. Theessential inputs that the WASH sector provides, in the form ofservices and hygiene promotion, have multiple impacts beyondthe WASH outcome itself, such as nutritional status, or education.Furthermore, these far-reaching effects of WASH can be felt beyondthe immediate impact, can have a cumulative effect throughout thelife course of an individual, and can often also affect the lives oftheir offspring (Ben-Shlomo & Kuh, 2002; Campbell et al., 2014).This paper describes the contribution of WASH to outcomes in othersectors and summarises the evidence for investment in these areas.The paper considers the following outcomes to which UNICEF iscommitted: diarrhoea, nutrition, complementary food hygiene,violence and female psychosocial stress, maternal and newbornhealth, MHM, school attendance, oral vaccine performance, NTDs,and disability.With this in mind, the objectives of this evidence paper arespecifically to: R eview the best available evidence with regard to strategicpriorities of UNICEF; P rovide an accessible guide to existing evidence on how WASHcan affect women and child health and well-being and otherdevelopment outcomes, with a particular focus on outcomes thatinclude but go beyond those traditionally measured by the WASHsector (see below for topics); the available evidence on the benefits of WASH Presentinterventions on health; I dentify what we do and do not know, and assess the robustnessof the available evidence relating to the impact of WASH and theeffect of WASH interventions on these outcomes.The Impact of WASH on Key Health & Social Outcomes PAGE 9

EVIDENCE PAPER The Impact of WASH on Key Health & Social OutcomesThis evidence paper does not make specific recommendations onwhat UNICEF should or should not do, but instead identifies keypoints for consideration in defining and implementing UNICEF’sStrategy for WASH 2016-2030 in the following areas:1. Assessing the scale of the problem2. Evidence of impact3. Evidence of what works4. Remaining knowledge gapsTo achieve this, each thematic chapter addressing a differentoutcome will cover:1. The problem: The extent to which this issue affects childhealth and well-being;2. Can WASH have an impact?: An assessment of the plausibleimpact of WASH; UNICEF/UNI47396/Pirozzi3. The effect of WASH interventions: A review of the evidencespecifically for the effects of WASH interventions.Liberia, 2007. A girl carries a large pail of water, outside her school in the village of Selega inthe north-central Lofa County.The Impact of WASH on Key Health & Social Outcomes PAGE 10

EVIDENCE PAPER The Impact of WASH on Key Health & Social OutcomesInterpreting the evidence on WASHWASH brings together several interventions, which are frequentlyimplemented by multiple agencies, often delivered separately butsometimes together. These interventions affect a wide range ofdirect outcomes, beyond just health outcomes. As a result, theevidence is complex and, therefore, difficult to classify.Nonetheless, expectations on the quality of evidence needed tojustify interventions have increased in recent years, and consensushas formed around rules of best practice for analysis, weighing andcombination of such evidence.In many policy-making domains, the systematic review and theRandomized Controlled Trial (RCT) have emerged as the goldstandard for quality of evidence as they are judged to reducesystematic error – or bias – to the greatest extent possible (Jüniet al., 2001). However, aside from RCTs and their meta-analysis,there are a wide range of observational study designs, includingecological, cohort, cross-sectional and case-control studies, someof which do not have a specified intervention and/or control.Increasingly in the WASH sector, various econometric methods arealso being employed to interrogate cross-sectional and longitudinaldata to address important questions (Spears, 2012).Beyond this, there are of course a wide range of qualitativeapproaches which can be employed in isolation or in combinationwith quantitative methods, and which are essential to manyareas of research, in particular those which are highly sensitive.For example, eliciting information from people about violence –possibly of a sexual nature – experienced while tending to theirurinary, defecation or MHM needs, can be a difficult process,provoking feelings of shame or inadequacy. Beyond these methodsand approaches, a very broad range of research disciplines isactively engaged in WASH research; epidemiologists, economists,microbiologists, geographers, anthropologists, statisticians,and engineers, to name but a few. As a result of this, the WASHliterature may be unwieldy, but it is rich and voluminous, reflectingthe broad challenge of delivering interventions which require bothchanges in infrastructure and in behaviour, and which influencepeople’s lives in many different ways.Assessing the quality of such a body of evidence is difficult to doobjectively. The GRADE approach uses algorithms for weighing andcombining evidence from these different levels (Guyatt et al.,2008) to reduce the partiality of human judgement in an objectivemanner. However, careful consideration is required to interpret theThe Impact of WASH on Key Health & Social Outcomes PAGE 11

EVIDENCE PAPER The Impact of WASH on Key Health & Social Outcomesoutcome of GRADE when navigating questions with high biologicalplausibility, but where very few intervention studies have beenconducted. Four specific challenges relating to the WASH literatureshould be highlighted:1. Subjective outcomes: There are ethical and logical argumentsagainst studies using the principal health outcome of interest:death from diarrhoea. However, the alternative outcomes,particularly self-reported diarrhoea morbidity, have provento be more subjective and subject to bias than was originallybelieved. For example, the 2007 systematic review on pointof-use (POU) water treatment by Clasen and colleagues drewthe conclusion from nearly 40 rigorous RCTs that drinking waterquality improvements were associated with reductions of nearly50% in diarrhoea rates. However, when the handful of blindedstudies were isolated they showed no impact on diarrhoea –suggesting that the overall impact may have resulted from aplacebo effect or courtesy bias (Schmidt & Cairncross, 2009).This weakness in the evidence base is relevant to any behaviourchange intervention for which, by its very nature, allocationcannot usually be blinded.2. Logistical challenges of randomization: There are political,ethical and practical complications associated with randomisingan intervention like water supply and sanitation, or evenallocating it by individual household. This is because of themuch appreciated non-health benefits of WASH—for example,time saved on water collection (Churchill et al., 1987)—andthe impossibility of providing water and sanitation without theknowledge of the studied population.3. Complex exposure-outcome relationship: In addition to thethree basic dimensions of WASH, there are various levels ofservice and a variety of combinations of the three. For example,practically every intervention study of sanitation is in fact astudy of water and sanitation.4. I mportance of context: A standpipe isolated in the desert is adifferent level of service from a standpipe in a village where halfthe households already have household connections. Quite apartfrom the variation in technology between different settings,there are often important differences in programme designand execution; hygiene promotion implemented effectively inone setting may have been much less effective in another. Anepidemiological study in this sector is thus meaningless unless itis seen in the context of the setting in which it was carried out.The Impact of WASH on Key Health & Social Outcomes PAGE 12

EVIDENCE PAPER The Impact of WASH on Key Health & Social OutcomesThere is clearly a tension that exists between achieving internaland external validity that should be taken into account whendesigning studies. If a study is to provide high-quality evidenceof health impact, it must be designed to exacting standards ofrigour, eliminating the potential for confounding and for bias dueto extraneous factors. However, often the more ‘rigorous’ thestudy the more it achieves internal validity, potentially at thecost of becoming less relevant to the wider context and existingprogrammes and policy issues.Box 1: Bradford Hill’s ‘viewpoints’ for assessing causality1. Consistency – in a systematic review theimpact was similar for the more rigorousstudies (Curtis & Cairncross, 2003);2. S trength of association – in a study focusedon domestic transmission of a single pathogen,handwashing prevented 85% of secondarycases (Khan, 1982);3. T emporal sequence – handwashing by mothersjust before preparing the family’s food has agreater impact than at other times (Luby etal., 2011);4. D ose response – one study found the impactof a sewer system construction project ondiarrhoea in a neighbourhood increased withthe proportion of households connected to thesewers (Barreto et al., 2007);5. Specificity – for example, water treatmentaffects diarrhoea, but not malaria;6. Coherence – (i.e. laboratory and epidemiologyresults cohere) – more faecal bacteria indrinking water is associated with morefrequent diarrhoea (Moe et al., 1991);7. Biological plausibility – given the number offaecal pathogens present in a community’swaste, it is not surprising that excretadisposal helps to prevent excreta-relateddisease (Feachem et al., 1983);8. Analogy – in particular, sanitation helps toprevent intestinal worm infections; it cantherefore be expected to prevent transmissionof other faecal pathogens, such as thosecausing diarrhoea;9. Experimental evidence – this refers tointervention studies, ideally randomizedtrials, many of which have been carried outfor household water treatment.Review methodsThis evidence paper is not a systematic review. Our methods have beenheuristic, based on existing systematic reviews where possible andexploratory reviews on a range of topics, and supplemented with morerecent studies. Wherever possible, we rely on published systematicreview-based meta-analyses to estimate the magnitude of effect for agiven WASH intervention on a given outcome.This paper takes a broad perspective, allowing for a range of exposuresand outcomes, a variety of settings in which studies have been carriedout and the application of judgement based on an assessment ofthe available evidence. In assessing causal evidence, our approachThe Impact of WASH on Key Health & Social Outcomes PAGE 13

EVIDENCE PAPER The Impact of WASH on Key Health & Social Outcomeswas generally informed by the criteria or ‘viewpoints’ famouslyproposed by Bradford Hill (1965). Box 1 broadly illustrates these,with reference to WASH. For the purpose of this evidence paper,we have used a pragmatic set of five applied viewpoints from whichto appraise the evidence base for WASH interventions. That is,whereas Bradford Hill’s viewpoints are for assessing the evidencefor causality in an association, the following viewpoints are used inthis paper to appraise the strength of support for implementationof each intervention. The first viewpoint is internal validity, whichassesses the rigour of the studies demonstrating cause and effect,including randomization, blinding, etc. The second assesses the easeof going to scale, which requires relevance to programme conditionsin the field. The third looks at the sustainability of the intervention,assuming reasonable effort is devoted to maintaining it. The fourthand fifth evaluate other substantial health benefits in addition toimpact on diarrhoea and significant non-health benefits respectively.For this evidence paper, we have been asked to show whichrelationships are supported by firm evidence and which by relativelyweak evidence. Throughout this paper, we consider the type ofevidence but also seek to grade the strength of the evidenceaccording to the following three categories: G ood evidence: several good quality studies consistently showan effect. For example, randomized trials with a low risk of bias,or observational studies showing a large effect size with a lowpotential for confounding; S uggestive evidence: studies show an effect, but statisticalsupport is weak due to insufficient study size. Or studies showsignificant effects, but there is a risk of bias and confounding; Weak evidence: no studies have been done, or where they beendone, they have shown inconclusive results.The Impact of WASH on Key Health & Social Outcomes PAGE 14

EVIDENCE PAPER The Impact of WASH on Key Health & Social OutcomesWASH and diarrhoeaThe problemDiarrhoea is defined as the passage of three or more loose orliquid stools per day (World Health Organization [WHO], 2013). Butglobally, diarrhoeal diseases are caused by infectious agents such asbacteria (e.g. E. coli, salmonella, shigella, campylobacter), viruses(e.g. rotaviruses, noroviruses and adenoviruses), and protozoa (e.g.cryptosporidium, amoeba and giardia). However, the aetiology ofdiarrhoeal diseases varies from region to region. Rotavirus is themain cause of severe and moderate diarrhoea (Lozano et al., 2013;Kotloff et al., 2014). Only a small proportion of diarrhoea casesresult from non-infectious conditions (such as intoxication or noninfectious inflammatory diseases) (WHO).Most diarrhoeal deaths are among children under the age of five(Prüss-Üstün et al., 2014), and within low-income countries, thevery poor suffer much more from diarrhoea than others (Howling &Kunst, 2010). In both low and middle-income countries, diarrhoealdisease is the second leading cause of morbidity and mortalityamong children under the age of five (Lim et al., 2012; Walker etal., 2013; Murray et al., 2015), and the leading cause of death insub-Saharan Africa (Prüss-Üstün et al., 2014). Approximately 1.5million children under the age of five died of diarrhoeal disease in2012 (Prüss-Üstün et al., 2014).Diarrhoeal disease can also affect a child’s nutritional status, withthe associated health and socio-economic consequences (discussedin the following section). One multiple country study found that25% of stunting in children under the age of two could be due tofive or more diarrhoeal episodes (Checkley et al., 2008). Longterm exposure to faecal pathogens may also partially explainenvironmental enteric dysfunction (EDD) (Humphrey, 2009).While most diarrhoeal diseases associated with poor WASH tend tobe endemic, some are epidemic in nature – notably, cholera andtyphoid fever. Cholera is an acute diarrhoeal disease that can killwithin hours if left untreated, and it is a continual public healthproblem in many parts of the world. Researchers have estimatedthat every year there are roughly 1.4 million to 4.3 million cases,and 28,000 to 142,000 deaths per year worldwide (Ali et al., 2012).The majority of reported cholera cases and deaths occur in Africa(Gaffga et al., 2007). Furthermore, the continent suffers fromexplosive outbreaks that result in high levels of both morbidity andmortality.The Impact of WASH on Key Health & Social Outcomes PAGE 15

EVIDENCE PAPER The Impact of WASH on Key Health & Social OutcomesHow does WASH influence diarrhoealdiseases? Recent updates in knowledge1. Can WASH affect diarrhoeal disease?Diarrhoeal diseases are characteristically transmitted via the faecaloral route. Poor WASH increases an individual’s exposure to faecalpathogens through multiple pathways, as demonstrated in the‘F-diagram’ below.Figure 1: The LESoil(fields)Hands(fingers)SOURCE: Cumming & Cairncross (2016); adapted from Wagner & Lanoix (1958)and Kawata (1978)It has been estimated that in 2012 a total of 842,000 diarrhoeadeaths were caused by inadequate WASH (502,000 from water,280,000 from sanitation and 297,000 from hand hygiene). Thisrepresents over half of diarrhoeal diseases, or an estimated 1.5%of the total disease burden (Prüss-Üstün et al., 2014). Given whatwe know about disease transmission routes and possible barriers tothese, the most recent estimate suggests that adequate WASH couldprevent the deaths of 361,000 children under the age of five, or 5.5%of deaths in that age group (Prüss-Üstün et al., 2014). A differentestimate, which includes WASH in addition to other interventionssuch as oral rehydration treatment and exclusive breastfeeding,suggests that 95% of diarrhoeal deaths in children under the age offive could be prevented by 2025, as a result of targeted scale-up ofsuch proven interventions (Bhutta et al., 2013).The Impact of WASH on Key Health & Social Outcomes PAGE 16

EVIDENCE PAPER The Impact of WASH on Key Health & Social OutcomesAs diarrhoeagenic pathogens spread by many different interactingpathways, the different components of WASH interventions need tobe well coordinated to be effective—although evidence is lackingon how best to combine different approaches. There is little doubt,however, that improving access to adequate amounts of water froman adequately distanced source, hygienic sanitation facilities andpromotion of handwashing with soap should be the cornerstones ofintegrated WASH campaigns (Cairncross et al., 2010). UNICEF/UNI85770/HoltzSanitation and hygiene promotion are still the two most effectiveinterventions for controlling endemic diarrhoea (Laxminarayan etal., 2006). An additional potentially critical intervention would be toimprove food hygiene, which may prevent many diarrhoea deaths,especially in hot climates where food hygiene is difficult to maintain(Curtis et al., 2011). For more information read this paper’s sectionon Complementary Food Hygiene.Niger, 2009. A child near traditional clay water containers in his family’s courtyard in thevillage of Foura Guirké, in the southern Maradi Region.With regard to cholera, although it is largely perceived to be awaterborne disease, person-to-person transmission, limited accessto sanitation, an inadequate water supply and poor hygienicpractices may contribute to the rapid

A key rationale for investing in WASH is the importance of WASH to other Sustainable Development Goal (SDG) outcomes. The essential inputs that the WASH sector provides, in the form of services and hygiene promotion, have multiple impacts beyond the WASH outcome itself, such as nutritional status, or education.