Water, Sanitation, Hygiene, and Soil-Transmitted Helminth Infection: A Systematic Review and Meta-Analysis

In a systematic review and meta-analysis, Eric Strunz and colleagues examine whether improvements in water, sanitation, and hygiene (WASH) practices are associated with reduced risk of infections with soil-transmitted helminths. Please see later in the article for the Editors' Summary


Introduction
More than a billion people are infected with soil-transmitted helminths (STHs) and many more live in high risk areas [1]. The global burden of STH infection is estimated at between 5 and 39 million disability-adjusted life years, largely attributable to anemia, stunting, and reduced cognitive development [2][3][4]. Humans are infected after ingesting eggs (A. lumbricoides and T. trichiura) or through penetration of the skin by infective larvae in the soil (hookworm [A. duodenale and N. americanus] and S. stercoralis) [1]. Current control strategies have focused on preventive chemotherapy through mass drug administration (MDA), in which at-risk populations are treated once or twice per year with benzimidazoles, primarily albendazole (usually given as a single oral dose of 400 mg) or mebendazole (500 mg) [5]. While preventive chemotherapy can greatly reduce morbidity from helminth infection, reinfection typically occurs rapidly after treatment [6].
Long-term STH control and eventual elimination require improvements to water, sanitation, and hygiene (WASH) access and practices [7]. The history of STH in the United States of America, South Korea, and Japan-where WASH improvements acted in concert with deworming to eliminate STH as a public health problem-supports the need for an integrated control paradigm [8][9][10]. WASH interventions are diverse, potentially including improvements in water access (e.g., water quality, water quantity, and distance to water), sanitation access (e.g., access to improved latrines, latrine maintenance, and fecal sludge management), and hygiene practices (e.g., handwashing before eating and/or after defecation, water treatment, soap use, wearing shoes, and water storage practices) [11][12][13][14][15][16][17][18][19][20]. Interventions often include multiple components, e.g., building ventilated-improved pit latrines while also providing hygiene education. Work in the WASH sector is often motivated by the view that access to clean water and adequate sanitation is a human right, but health outcomes are also broadly considered, with diarrheal disease burden representing a common measure of impact [21][22][23].
The successful integration of WASH into a disease control program has already been demonstrated for trachoma, whichlike STH-is also considered a neglected tropical disease (NTD). The World Health Organization (WHO) endorses the ''SAFE'' strategy for trachoma control: surgery to correct advanced stages of trachoma, antibiotics to treat active infection, facial cleanliness to reduce disease transmission, and environmental change (including increased access to water and improved sanitation) [24]. The SAFE strategy explicitly calls for the implementation of improved access to, and use of, water, sanitation, and hygiene through improvements in delivery and/or specific interventions.
Such a fully integrated strategy-including guidelines and targets-does not yet exist for STH control, in part because evidence examining the relationship between WASH and STH is limited. A seminal review by Esrey and colleagues found few investigations that evaluated the association between WASH and STH infection [25]. A recent systematic review and meta-analysis by Ziegelbauer and colleagues found that individuals who have access to and use of sanitation facilities were at lower odds of infection with STH compared to individuals without sanitation [26]. Additional empirical evidence that links WASH improvements to reductions in STH infection is scarce, and an improved evidence-base may lead to better coordination between the NTD and WASH sectors [27,28].
To fill this gap, we conducted a systematic review and set of meta-analyses to examine evidence of association between STH infection and WASH. We expanded the study's focus to include up-to-date meta-analyses for water and hygiene components, in addition to sanitation. We only used adjusted effect estimates in meta-analyses to help account for potential confounding and followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines for systematic reviews. Our use of the Grades of Recommendation, Assessment, Development and Evaluation (GRADE) approach for quality assessment also provides a comprehensive accounting of the limitations of available evidence. We hypothesized that improvements in WASH would be associated with reductions in odds of STH infection. Thus, the purpose of this study was to quantitatively summarize the relationship between WASH access or practices on STH infection, while also synthesizing available data that did not qualify for meta-analysis.

Search Strategy, Inclusion Criteria, and Data Extraction
Our review adheres to the PRISMA and Meta-analysis of Observational Studies in Epidemiology (MOOSE) reporting guidelines (see Texts S1 and S2) [29][30][31]. The methods protocol is available in Text S3. A study investigator (ECS) and two research assistants (Rachel Stelmach [RS] and Claire Still [CS]) systematically searched PubMed, Embase, Web of Science, and LILACS for relevant articles from inception to October 28, 2013. We also indexed relevant studies from the bibliography of reviews by Ziegelbauer and colleagues [26] and Asaolu and Ofoezie [32]. Abstracts without published articles were considered eligible for inclusion. Additionally, we requested available unpublished research from the US Centers for Disease Control and Prevention, The Carter Center, The Task Force for Global Health, the WHO regional offices, and the authors' personal collections.
The native search engines within PubMed, Embase, Web of Knowledge, and LILACS were used to search each respective database using Boolean operators. The search included two clusters of terms: one for STH (i.e., helminth, soil-transmitted helminth, geohelminth, ascaris, lumbricoides, trichuris, trichiura, hookworm, ancylostoma, duodenale, necator, americanus, strongyloid*, stercoralis) and one for WASH (i.e., sanitation, sanitary engineering, water supply, waste management, environment*, excre*, faec*, fecal, feces, hand washing, handwashing, hygiene, latrine*, toilet*, water, soap). Results had to contain at least one term from both clusters. ''Extensive search'' was enabled when searching with Embase. Because Embase only allowed for exporting up to 5,000 records, results were stratified by date in order to screen and export all results in smaller segments. All search records were exported to bibliographic files and imported into Endnote X5 (Thomson Reuters), which was used to manage and screen search results. Titles, and when available, abstracts were scanned by an investigator (ECS) and also independently by research assistants (RS and CS) to determine possible relevance. Final selection was based on the full text of all potentially applicable articles. Ambiguous articles were examined by a senior reviewer (MCF).
Publications in all languages were considered. Studies in English, Spanish, Portuguese, and French were screened by investigators directly. Chinese-language articles were reviewed by a study collaborator (Shuyuan Huang [SH]) who assessed eligibility and extracted relevant data for the research team. Relevant data from all eligible studies was abstracted by a reviewer (ECS) and independently by assistants (RS and CS). Extracted data included study design, setting, year, population characteristics, WASH components measured, diagnostic approach, STH species, and relevant effect measures. Odds ratios (ORs) served as the primary effect measure in the reviewed literature. We collected both crude and adjusted estimates if available. Excel 2007 (Microsoft) was used to input and manage data using a long format to accommodate multiple effect estimates per study.
An article was eligible for inclusion if it presented a measure of effect between WASH and STH (e.g., an OR). For studies that pooled multiple intestinal parasites (e.g., Giardia intestinalis and STH) into one outcome measure, we contacted authors to request disaggregated data. We did not exclude studies based on methodology or population characteristics. Studies that evaluated multiple WASH components were included, as long as the components could be assessed separately from deworming medications and other non-WASH interventions.
There are few standard definitions for WASH access and practices, and it is difficult to measure WASH behaviors objectively [33]. We were unable to consistently connect water and sanitation variables reported in retrieved studies to the WHO and UNICEF Joint Monitoring Program's water and sanitation ladder definitions [34,35]. For this review, ''treated water'' is defined as the use of any chemical or physical treatment of water to change its potability, whether conducted at the source or at the point of use. Two specific forms of treatment included boiling and filtering water at home. ''Piped water'' describes access to, or use of, water collected from a piped infrastructure, regardless of where the water is accessed (public/private) or how well maintained the infrastructure may be. ''Sanitation access'' was our primary sanitation exposure, defined as access to, or use of, any latrine. We did not exclude studies that lacked information about latrine quality, so access to sanitation could refer to anything from simple pit latrines to flush toilets. For hygiene, ''washing after defecation'' refers to the availability of handwashing resources (e.g., a wash basin) near sanitation facilities or reported handwashing behavior after defecation. ''Soap use or availability'' could refer to washing with water alone or no washing as the comparison group. Further, these definitions do not incorporate any criteria for compliance or consistency, since such details were rare in retrieved literature.

Statistical Methods
We conducted meta-analyses for groups of effect estimates that related similar WASH access or practices (e.g., latrine availability and/or use became ''sanitation access'') to a common outcome. Potential outcomes included infection with a specific STH (i.e., A. lumbricoides, T. trichiura, hookworm, and S. stercoralis) or any STH generally. Note that ''any STH'' reflected infection with an individual species or co-infection with multiple species when authors reported aggregated STH infection results. Meta-analyses were performed for groups of independent effect estimates that numbered three or greater and shared a similar exposure and infection outcome. A study that measured several WASH components could contribute to multiple meta-analyses, but could only supply one effect estimate for any single meta-analysis.
We employed random-effects models to account for the expected heterogeneity between studies [36]. Only adjusted estimates were utilized to limit the impact of confounding on pooled effect measures [37]. When necessary, we inverted estimates to reflect the effect of WASH, rather than the absence of WASH. This inversion was necessary in order to ensure enough study estimates were available for meta-analysis, but could have resulted in additional heterogeneity. For example, the inverse of ''no sanitation access'' may be similar to, but distinct from, ''sanitation access'' when assessed by questionnaire due to bias associated with socially desirable responses. Further, the presence of WASH access or practices may not necessarily be the same as the inverse effect of their absence, especially if important confounders or effect modifiers remain unexplored. Estimates of effect not included in meta-analyses were summarized in the text. The meta-analysis package MAIS for Stata version 12 (StataCorp) was used to perform the random-effects meta-analyses with the DerSimonian and Laird method [38]. The natural log of reported ORs was the dependent variable. CIs use the 95% level unless otherwise noted.

Bias Assessment and Evidence Quality
We used the GRADE framework to assess potential sources of bias within studies and determine overall strength of evidence for each meta-analysis [39]. The GRADE approach is used to contextualize or justify intervention recommendations with four levels of evidence quality, ranging from very low to high. These levels correspond to how likely it would be for further research to alter conclusions drawn from the current evidence. ''High quality'' Table 1. Criteria for study bias assessment.

Criteria Description
Infection diagnostics Is a diagnostic assay clearly mentioned? Is there any form of quality control in the diagnostic process (e.g., a senior technician doing spot-checks)?
Exposure assessment Was exposure assessment (e.g., access to clean water, washing hands) ascertained via a self-reported survey response (unreliable) or observed directly by investigators (more reliable)? Is there any attempt to gauge proper use of water, hygiene, or some form of ''quality control'' for the exposures?

Confounding assessment
Are only crude estimates computed? Has matching and/or multiple logistic regression been undertaken to control for important potential confounders?
Response rate Is the response rate (or loss-to-follow-up) similar for infected versus non-infected individuals?

Selective reporting
Is there evidence of selective reporting within an article (e.g., outlining certain variables of interest in the methods but not providing any data on them in the results)? doi:10.1371/journal.pmed.1001620.t001 suggests that it is very unlikely for conclusions about effect estimates to change, whereas ''very low quality'' suggests that any estimate of effect is highly uncertain [40]. We formed our key bias categories from the literature, GRADE recommendations [41], and two instruments highlighted by the Cochrane Collaboration [42]: the Downs and Black tool [43] and the Newcastle-Ottawa scale [44]. We focused on five potential sources of bias in our assessment of individual studies: (i) diagnostic approach for assessing STH infection; (ii) exposure assessment; (iii) confounding assessment; (iv) response rate; and (v) selective reporting. Each study received one of three rankings for each source of bias: low risk, unclear risk, or high risk. Detailed criteria for these categories are available in Table 1. Bias was assessed independently by ECS and one of the two research assistants (RS and CS), compared, and reviewed by a senior assessor (DGA or MCF) if necessary.
We assessed the overall quality of evidence for each metaanalysis after considering seven key characteristics. Each metaanalysis could receive a quality grade of very low, low, moderate, or high [45]. Meta-analyses of observational studies were classified as ''low'' by default, but could be downgraded (because of imprecision, indirectness, inconsistency, publication bias, and potential confounding) or upgraded (because of magnitude of effect, dose-response relationship, and potential confounding) on the basis of the overall strength of the evidence. Inconsistency (i.e., heterogeneity) was assessed with Moran's I 2 and Cochran's Q-test [46]. I 2 provides an estimate of the proportion of variability in a meta-analysis that is explained by differences between the included studies instead of sampling error [47]. If a study exhibited an I 2 value over 50%, there was potential cause for concern, and the Q-test was also checked for a p-value less than 0.10. Values for I 2 over 70% or Q-test p-values lower than 0.05 resulted in the automatic downgrading of a body of evidence.
Publication bias was assessed through a visual inspection of funnel plots, though Egger's test also informed our interpretation [48]. Evidence quality was downgraded due to ''imprecision'' if the pooled effect estimate's 95% CI overlapped with the null (i.e., statistical significance at the 0.05 level). Although we provide CIs for pooled point estimates, imprecision remains a valuable criterion since not all consumers of reviews understand the importance of CIs and statistical uncertainty.
Evidence quality was upgraded owing to large magnitude of effect if the meta-analysis yielded a pooled OR less than 0.33 or greater than 3.0 [41]. Traditionally, risk ratios (RRs) are considered to show a large magnitude if they are less than 0.5 or greater than 2.0. However, ORs overstate the effect size compared to RRs, especially when initial risk (i.e., the prevalence of the outcome of interest) is high [49]. Because STH infection is relatively common, a more conservative threshold was needed for ORs in order to qualify as a large magnitude of effect.
Evidence quality could also be upgraded or downgraded on the basis of any unaccounted sources of potential confounding that would likely have a predictable direction on the effect estimate. For example, hygiene behaviors are typically over-reported in surveys, which could reduce the measured strength of effect for hygiene practices since the exposure group includes those who did not practice hygiene [50][51][52].
Due to the breadth of the review, indirectness was not a common concern, but would be more important for future reviews that focus on specific populations, settings, or interventions. Doseresponse relationships were assessed by examining studies where Table 2. Criteria for meta-analysis GRADE assessment.

Criteria Description Imprecision
Caused the evidence quality to be downgraded if the pooled effect estimate's 95% CI overlapped with the null (i.e., one for odds ratios). In this context, imprecision is synonymous with a pooled estimate being statistically non-significant at the 0.05 level. Imprecision is used to downgrade evidence quality because some consumers of reviews (e.g., policymakers and practitioners) often do not fully understand statistical uncertainty.

Indirectness
Did not cause any evidence quality to be downgraded. Our review had a broad scope that aimed to collect a wide array of evidence exploring different populations and contexts. Traditionally, indirectness refers to issues that may limit the generalizability of evidence's reported results to the review's specified research question. This could be caused by differences in study population, study design, co-interventions, etc.

Inconsistency
(i.e., heterogeneity) Assessed with Moran's I 2 and Cochran's Q-test [46]. If a study exhibited an I 2 value over 50%, there was potential cause for concern, and the Q-test was also checked for a p-value less than 0.10. Values for I 2 over 70% or Q-test p-values lower than 0.05 resulted in the downgrading of a body of evidence.

Publication bias
Assessed through a visual inspection of funnel plots, though Egger's test also informed our interpretation [48]. Detecting publication bias is difficult when dealing with dichotomous outcomes, especially when there is significant between-study heterogeneity. In such circumstances, the popular Egger's test is usually inappropriate, with the potential to result in many false positives. For this reason, qualitative funnel plot analysis served as our primary assessment tool, though we also computed Egger's statistics to inform our judgment. Tests described by Rü cker et al. [135] and Peters et al. [136] were also considered, but not performed.
A large magnitude of effect (also called ''effect size'') Could upgrade overall evidence quality if pooled odds ratios were less than 0.33 or greater than 3.0 [41]. The standard criteria for risk ratios and hazard ratios is that effect estimates be less than 0.5 or greater than 2.0. However, since odds ratios will show a greater magnitude than risk ratios, especially when an outcome is common, a more conservative cut-off value is needed. No firm rules have been established in the literature, so we increased the relevant effect size magnitude for odds ratios by 50%.

Evidence of a dose-response relationship
Can upgrade evidence quality. Dose-response relationships were assessed by examining studies where exposures were discretized into ranked categories, e.g., analyzing ''always washes hands'' versus both ''sometimes'' and ''never.'' A doseresponse relationship was considered possible if the point estimates improved between the ordinal categories, especially if relevant confidence intervals did not overlap.

Potential confounding
Can upgrade a body of evidence if there are plausible factors that may be artificially weakening the observed pooled measurement. In the case of hygiene, individuals are known to overreport handwashing behaviors, which would systematically lower any apparent benefits. Potential downgrades are also possible, however, especially if established confounding variables are not taken into account by an analysis.  Table 2.

Retrieved Studies
The search yielded a total of 47,589 articles from PubMed (n = 21,718), Embase (n = 18,188), Web of Knowledge (n = 7,502), and LILACS (n = 181), with 42,882 unique records. Our PRISMA flow diagram is available in Figure 1. After reviewing titles and abstracts, we examined 397 articles more intensively: 264 were excluded for lacking a relevant effect measure, 30 were excluded for aggregating non-STH infections in the outcome, and 11 were excluded for being review or editorial articles (see Tables 3-5 for included studies and S1 for excluded ones). We contacted 11 authors to obtain additional data [53][54][55][56][57][58][59][60], but only three authors responded [61][62][63]. A total of 94 studies ultimately met our inclusion criteria, yielding over 450 estimates of effect. Retrieved data included findings from one unpublished investigation [64] and one publication with information about two related studies [65].
Most included studies were published in English (n = 86), though articles in Portuguese (n = 4), Chinese (n = 2), and Spanish (n = 2) were also included. Studies researched populations in Asia (n = 42),     Africa (n = 29), and the Americas (n = 23). Studies investigated access and practices relating to water (n = 56), sanitation (n = 79), and hygiene (n = 53) ( Figure 2); the most commonly explored were access to sanitation (n = 63), access to water (n = 45), handwashing (n = 30), and wearing shoes (n = 27). Studies reported investigating infection with A. lumbricoides (n = 69), T. trichiura (n = 60), hookworm (n = 63), S. stercoralis (n = 12), and any STH collectively (n = 52). Tables 6 and 7 illustrate the number of articles in which both specific WASH components and helminth infections were investigated. Of 94 studies, 89 were observational: 75 used a cross-sectional epidemiologic design, 13 were prospective, and the remaining was a case-control study. Most studies investigated multiple potential risk factors for STH infection. Exposure status for WASH access and practices was typically determined through self-report, although 15 studies also used some form of observation to validate self-reported information. All included studies reported the diagnostic method used to assess helminth infection, with the Kato-Katz technique most frequently mentioned (n = 63). To assess the independent effect of WASH components on STH infection, authors typically used multiple regression analysis (n = 68), though adjusted effect estimates were often not reported for WASH covariates if they were not statistically significant. Not all multivariable models were reported with a full list of included covariates either. Slightly more than one-third of the studies (n = 33) reported at least one non-significant adjusted effect estimate. Study bias assessment is presented in Table S2. Metaanalysis results are available in Table 8 and grades summarized in Table 9.

Water
Water-related access and practices were generally associated with lower odds of STH infection. We conducted meta-analyses to examine the association of piped water access and use of treated water on STH infection. Using treated water (filtered or boiled) was associated with lower likelihood of having any STH infection (k = 3, OR 0.46, 95% CI 0.36-0.60). The quality of evidence for the analysis was low, as all three studies were observational ( Figure 3). Use of piped water was not associated with STH infection in general (k = 5, OR 0.93, 95% CI 0. 28-3.11). The quality of evidence for the pooled estimate was very low due to high heterogeneity (I 2 = 98.6%, 95% CI 98%-99%, Q p-value, 0.01) among the studies ( Figure 4). The heterogeneity could have stemmed from multiple factors, as the five studies shared few methodological characteristics. Use of piped water was associated with reduced likelihood of A. lumbricoides infection (k = 4, OR 0.40, 95% CI 0.39-0.41) and T. trichiura infection (k = 3, OR 0.57, 95% CI 0.45-0.72). Evidence quality for these two meta-analyses was low, based on four studies and three studies respectively ( Figures 5  and 6). We did not find a sufficient number of studies to conduct a similar meta-analysis for hookworm infection, although Nasr and colleagues found a significantly lower adjusted odds of infection (OR 0.59, 95% CI 0.34-0.91) for Malaysian children with access to piped water [66]. Other researchers found no statistically significant associations between piped water access and hookworm infection [67,68].
Other water-related exposures for STH infection were reported in the literature, but not with sufficient frequency for metaanalyses. In one study examining storage of water, Quintero and colleagues found a significantly higher adjusted odds of T. trichiura infection for Venezuelan children and adults collecting water in ''inappropriate'' receptacles (OR 1.12, 95% CI 1.09-1.15) [69]. Limited evidence also was retrieved on the influence of water source location; Belyhun and colleagues [70] found a beneficial
We found limited evidence that use of shared or private sanitation facilities influenced odds of STH infection. Worrell and colleagues [74] found in Kenya that participants using toilets located outside of their household premises had significantly increased odds of infection with any STH. In contrast, another study found that sharing latrines with neighboring households, compared with private latrine use, was associated with significantly lower odds of hookworm infection [71]. Few details were provided to contextualize this finding.

Hygiene
Three randomized controlled trials, two carried out in China and one in the Peruvian Amazon, found strong benefits for interventions that focused on promoting hygiene in schools [75][76][77]. Xu and colleagues [75] assessed a randomized intervention that promoted handwashing with soap, both before eating and after defecation among 657 school children in three schools. All infected children were treated at baseline. At the 1-year follow-up, A. lumbricoides prevalence for children in the experimental group had declined by 35.7% (pre-intervention prevalence, 68.3%; postintervention cumulative infection rate, 43.9%) compared with an increase in the control group of 78% (pre-intervention, 41.4%; post-intervention, 73.7%); this was a statistically significant difference (p,0.01). The study's primary limitation was that schools were the unit of randomization, with two primary schools becoming controls and the third receiving the intervention. With so few clusters, it is highly possible that confounding factors were not comparable between the control and experimental groups.
More recently, Bieri and colleagues [76] reported on a singleblind, unmatched, cluster-randomized intervention trial involving 1,718 children (aged 9-10) in 38 schools over the course of one school year. Schools were randomly assigned to a health-education package, which included an entertainment-education cartoon video, or to a control package, which only displayed a healtheducation poster. All participants were treated with albendazole at baseline. At follow-up at the end of the school year, knowledge about STH was significantly higher in the intervention group, and almost twice as many intervention children (63.3% versus 33.4%, p,0.01) reported washing their hands after defecating. The incidence of STH infection (predominantly T. trichiura and A. lumbricoides) was also significantly improved in the experimental Table 4. Cont.

Author [cite ID],
Year -Country  schools: 50% lower in the intervention group than in the control group (4.1% versus 8.4%, p,0.01). Gyorkos and colleagues [77] conducted an open-label, clusterrandomized controlled trial using a hygiene education intervention in Peruvian primary schools. Within paired groups, 18 schools (1,089 fifth grade student participants) were randomly allocated to receive albendazole and the hygiene intervention or albendazole alone. The health intervention included a helminth-oriented class for students, a health curriculum workshop for teachers, and educational print materials. Four months after the intervention, the experimental group showed a significant reduction in A. lumbricoides intensity compared to deworming alone (adjusted incidence rate ratio [IRR] 0.42, 95% CI: 0.21-0.85). T. trichiura and hookworm intensity did not show statistically significant improvements in the experimental group, nor did prevalence of any single STH species. Children in the intervention group showed significant improvements in STH knowledge and water treatment behaviors compared to the control, but not in most other hygiene practices (e.g., handwashing). The authors also noted that the prevalence of hookworm was low (about 5% compared to 30% for A. lumbricoides and 50% for T. trichiura) and that albendazole was less efficacious against T. trichiura than it was against A. lumbricoides.
Our meta-analyses of hygiene-related observational evidence provided estimates that are consistent with findings from these randomized controlled trials. Soap use or availability was significantly associated with lower odds of STH infection at the 5% level (k = 3, OR 0.53, 95% CI 0.29-0.98). The quality of the evidence was low, though the possibility of respondents' overreporting hygiene behaviors could have underestimated the strength of the association (Figure 11). Handwashing, both before eating (k = 3, OR 0.38, 95% CI 0.26-0.55) and after defecating (k = 3, OR 0.45, 95% CI 0.35-0.58), was associated with lower odds of A. lumbricoides infection (Figures 12 and 13). Both analyses were of low quality due to the observational evidence available. Handwashing after defecation also was associated with reduced odds of any STH infection (k = 5, OR 0.47, 95% CI 0.24-0.90). This meta-analysis had very low evidence quality due to high heterogeneity among estimates from the five pooled studies (I 2 = 88%, 95% CI 74%-94%, Q p-value,0.01, Figure 14). All studies used Kato-Katz for diagnosis, but varied considerably in  Table 6. Number of studies (n = 94) that investigated STH species and WASH domains.  most other study characteristics, including population age, baseline prevalence, and geographic setting. Balen and colleagues reported limited evidence of a dose-response effect for handwashing; respondents who more frequently washed their hands with soap after defecation had lower odds of infection with any STH, but confidence intervals of the handwashing groups overlapped [78]. Washing vegetables was found to be associated with lower odds of STH infection in two studies. Steinmann and colleagues [79] found washing vegetables to be negatively associated with A. lumbricoides infection in school children (OR 0.69, 95% CI 0.50-0.95), while Hohmann and colleagues [80] found washing was associated with lower odds of T. trichiura (OR 0.50, 95% CI 0.31-0.79) and any STH infection (OR 0.71, 95% CI 0.51-0.99).
Our meta-analysis found evidence of a strong association between wearing shoes and lower odds of hookworm infection (k = 5, OR 0.29, 95% CI 0.18-0.47). The quality of the evidence was moderate, upgraded due to the magnitude of effect ( Figure 15). Wearing shoes was also associated with lower odds of infection with any STH (k = 3, OR 0.30, 95% CI 0.11-0.83). The evidence quality for that analysis was low, downgraded by heterogeneity (I 2 = 74%, 95% CI 12-92%, Q p-value = 0.02) ( Figure 16) but upgraded by a strong effect magnitude. Heterogeneity could have been introduced by many different factors, as the studies shared few characteristics. Three studies found mostly non-significant associations between geophagy (i.e., consumption of soil) and STH infection [81][82][83]. In adjusted models, households with dirt floors in India and Venezuela were found to have higher odds of T. trichiura and A. lumbricoides infection than were houses with other more elaborate flooring material [69,84]. Young children living with dirt floors in Colombia also showed higher odds of infection with any STH compared to those with tile or cement floors [85].

Integrated Interventions
In a cluster-randomized controlled trial, Freeman and colleagues examined a comprehensive WASH intervention in Kenyan schools that included hygiene promotion, water treatment and storage, and installation of sanitation infrastructure [27]. The intervention reduced reinfection prevalence (OR 0.56, 95% CI 0.31-1.00) and egg count (IRR 0.34, 95% CI 0.15-0.75) of A. lumbricoides, but not of T. trichiura or hookworm. Effects of the intervention differed by sex, with girls in the intervention group showing a significantly reduced A. lumbricoides infection intensity compared to the control group; boys in the intervention group did not show any significant difference from controls. Shoe-wearing and geophagy also emerged as effect modifiers for hookworm and T. trichiura infection intensity, respectively.
Dumba and colleagues found no statistically significant benefit of a participatory hygiene and sanitation transformation (PHAST) intervention when compared with a control group that only received deworming [86]. PHAST uses training sessions to encourage communities to identify problems in their own environment, decide what aspects need to be improved, and then implement changes. Parents or guardians of participating children in 19 villages received three PHAST education sessions. Participants in both control and experimental villages received albendazole and showed significant reductions in helminth prevalence compared with baseline, but the prevalence in the experimental group did not decline more than that among the control children. This study grouped Hymenolepis nana and Enterobius vermicularis with STH in analysis, but only a handful of participants were infected by H. nana or E. vermicularis, whereas STH prevalence was very high (.80%). Table 7. Number of studies that investigated STH species and WASH access and practices. Cells with high numbers but no meta-analysis (no footnote) indicate that effect measures were not reported (selective reporting), reported measures were not statistically adjusted, or that the WASH access and practice was too diverse to be effectively grouped in a meta-analysis (e.g., handwashing can be measured before eating or after defecating).
a Water Types and Latrine Types refer to studies that measured multiple sanitation comparisons, not just ''latrine versus no latrine.'' For example, a study could examine water collected from rivers, wells, or piped connections. b Gray cells indicate that a meta-analysis was conducted for that WASH variable and STH outcome. doi:10.1371/journal.pmed.1001620.t007

Strongyloides stercoralis
We found 12 studies that investigated the relationship between WASH and S. stercoralis infection, but only located relevant effect estimates in five. Among school children in Cambodia, Khieu and colleagues found crude associations between infection and handwashing, shoe-wearing, and sanitation access [87]. Hall and colleagues found mixed results for a range of sanitation-related exposures, with some evidence that open defecation and use of community latrines were associated with higher odds of S. stercoralis infection in children [72]. In a multivariable model using data from a rural Peruvian community, Yori and colleagues found that wearing shoes never or occasionally (versus more frequently) was associated with higher odds of infection (OR 1.89, 95% CI 1.10-3.27) [88]. Knopp and colleagues did not find a significant association between S. stercoralis infection and home latrine ownership or handwashing after defecation [89].

Discussion
We conducted a systematic review and meta-analysis of the relationship between WASH access and practices and STH infection. Our analysis revealed that WASH access and practices are generally, but not universally, associated with lower odds of STH infection. Particularly strong associations emerged between wearing shoes and hookworm infection (OR 0.29, 95% CI 0.18-0.47), piped water use and A. lumbricoides infection (OR 0.40, 95% CI 0.39-0.41), and treated water use and infection by any STH (OR 0.46, 95% CI 0.36-0.60). Pooled estimates for all metaanalyses, except for two (i.e., piped water use for any STH and sanitation access for hookworm), indicated at least a 33% lower odds of STH infection associated with specific WASH behaviors or access (Table 8). All but two of the meta-analyses were statistically significant at the 5% level.
On the basis of the evidence available, this review primarily draws upon observational studies. Observational research typically has greater risks to internal validity than randomized controlled trials, but such research is also key to providing a broad evidence base. When conducted well, randomized controlled trials provide the strongest evidence of a causal relationship between an exposure (e.g., an intervention) and an outcome. In the WASH context, however, conducting RCTs can be ethically and financially challenging. Traditional randomized designs can be costly and require that a subset of the target population be allocated to the control group, receiving only a limited intervention. Observational studies can be conducted more quickly and affordably in a wide array of contexts, allowing for WASH access and practices to be investigated in different social-ecological systems. This diversity is critical, since the effectiveness of specific WASH interventions can vary widely across settings, and interventions will most likely provide the greatest impact after being tailored to local conditions. Looking forward, a stepped wedge design represents a powerful compromise between ethics, operational feasibility, and internal validity. With a stepped wedge approach, the rollout of an intervention is randomized so that all participants eventually receive the study benefits, but at different times. Because many WASH interventions require staggered implementation owing to limited financial and human resources, randomizing the order in which communities are visited is often feasible. Combined with longitudinal data analysis, this design allows for robust assessments that can integrate with many interventions without radically altering implementing organizations' plans.
This review highlights important gaps in the WASH and STH body of literature. For example, only a few of the studies that met our inclusion criteria investigated the impact of sharing latrines (n = 6) or latrine maintenance (n = 3) on STH infection. The effect of treating water (n = 7) and geophagy (n = 10) were also infrequently explored. S. stercoralis was by far the least commonly investigated STH infection, reflecting another important knowledge gap. A total of 35 studies contributed data to the 14 meta-analyses. A lack of standardized WASH definitions across studies limited our ability to pool results via additional meta-analyses. More consistent use of the Joint Monitoring Program's water and sanitation ladder definitions would aid future review efforts. Additional metaanalyses could have been conducted if all reviewed studies had provided relevant adjusted estimates of association. For example, many studies investigated the relationship between ''toilet sharing'' on any STH infection and ''water access'' on hookworm infection, but a dearth of reported adjusted estimates stymied meta-analyses of these relationships (Table 7).
Few studies analyzed the relationship between fecal egg count, a proxy for intensity of infection, and WASH [27,81,90], even though intensity of infection represents a more relevant predictor for morbidity than prevalence alone [91]. A lack of measures on this relationship represents a considerable gap in the literature, though many studies did report broadly on intensity of infection. Zero-inflated modeling strategies have recently shown promise in analyzing fecal egg count datasets, which often contain excess zero counts due to some individuals not harboring infections [92][93][94]. Contemporary analysis of existing data represents a potentially cost-effective mechanism for yielding additional insights into this topic.
Our findings build upon past reviews by Asaolu and Ofoezie [32] and Ziegelbauer and colleagues [26], which both concluded that WASH represents a valuable strategy for STH control. Although Asaolu and Ofoezie did not conduct a meta-analysis, their comprehensive review found broad evidence of reductions in STH prevalence and intensity resulting from multiple types of WASH interventions. Asaolu and Ofoezie concluded that improvements in sanitation systems and hygiene practices were important tools to not only sustain preventive chemotherapy benefits, but also help protect the uninfected. Results from our meta-analyses support their conclusion using systematically aggregated quantitative data. Ziegelbauer and colleagues focused more specifically on latrine access and use, conducting a rigorous meta-analysis using primarily crude odds ratios. The results from our meta-analyses, which drew upon adjusted odds ratios, are consistent with their findings and lend additional support to the value of sanitation improvements for STH control. Our metaanalyses also broadened focus to include water and hygiene components, allowing for a quantitative summary of currently available evidence across the three core WASH domains.
Our analysis of the relationship between access to a piped water source and STH infection yielded significantly protective associations for A. lumbricoides and T. trichiura, but not for any STH infection generally. The meta-analysis of any STH yielded strong heterogeneity statistics, reflecting a spread in observed effects. While the inclusion of hookworm infections in the ''any STH'' analysis may seem like a possible source of the variability, we found no clear evidence to support this explanation. The only study that analyzed hookworm infection and piped water use with an adjusted model found a significantly protective association, so other sources of heterogeneity should be considered.
The presence of heterogeneity can be systematically investigated by statistics like Moran's I 2 and Cochran's Q, but these global tests do not themselves uncover specific causes of heterogeneity. Diversity among studies can originate from a plethora of sources: population, setting, diagnostic approach, study design, analytic method, definitions, and so on. Without additional subgroup analysis or meta-regression, which both require a large body of studies, it is difficult to investigate the myriad potential causes of heterogeneity. Without clarification, the presence of heterogeneity indicates that pooled results are averaging multiple related, but   distinct effects. For example, access to piped water could have different levels of benefit depending on distance to the source [95,96], water quality [70,97], or other unknown factorsespecially when studies use different diagnostic assays and are conducted in a variety of community settings.
Concerning sanitation, our meta-analyses of access to sanitation yielded considerably lower odds of infection with A. lumbricoides, T. trichiura, or any STH for those with latrine access. We did not find evidence of a statistically significant association between sanitation and hookworm, though the pooled estimate suggested reduced  odds of infection. Our sanitation findings were comparable to those found by Ziegelbauer and colleagues, who asserted that improved sanitation access should be prioritized alongside preventive chemotherapy to achieve a sustainable reduction in helminthiasis burden. They found an overall pooled odds ratio of 0.51 (95% CI 0.44-0.61) for the effect of sanitation availability and use, while we found an odds ratio of 0.66 (95% CI 0.57-0.76). Species-specific results were similar as well, with the exception of hookworm. Differences in the magnitude of our findings may be attributed to the use of adjusted measures in our analysis, since Ziegelbauer and colleagues used unadjusted estimates. In addition, we did not include separate estimates for sanitation use and access. Taken together, these two reviews support the hypothesis that improved access to, and use of, sanitation prevents STH infection. Additional research could help explore the complementarity of sanitation promotion with MDA.
For hygiene, three randomized controlled trials provided strong evidence linking hygiene practices-especially handwashing with soap-to reductions in STH infection [75][76][77]. However, not all hygiene interventions may be effective in reducing STH infection [86]. Our meta-analyses of the effect of handwashing before eating and after defecation for A. lumbricoides infection, along with handwashing after defecation and soap use for any STH infection, also yielded significant results that suggest protective effects. Accurately assessing handwashing is challenging; self-reported and observed measures are often highly biased [33]. Many studies rely on self-report, but individuals have consistently been shown to over-report handwashing behaviors [98]. Heterogeneity was exhibited in the analysis of handwashing after defecation, suggesting that the benefits of handwashing may vary considerably depending on circumstances and definitions. Beyond handwashing, our analysis also showed that wearing shoes was associated with significantly lower odds of infection with hookworm and any STH.
These results may be of interest to several audiences. Researchers can take note of the gaps in the literature identified by this review and focus investigation on key outstanding questions (e.g., the impact of WASH on S. stercoralis infections). Policymakers should understand that, despite gaps in data, these findings provide a broad evidence base in support of WASH for STH control-especially from randomized trials for hygiene interventions. WASH practitioners will recognize that these findings provide further support for their efforts and, we hope, will consider partnering with STH researchers to evaluate future interventions.

Strengths and Limitations
Our review included only adjusted effect estimates in metaanalyses, which lends greater strength to our pooled results [37]. Many different variables were controlled across studies, which may contribute to heterogeneity. However, this variation in adjusted models may also serve as a small buffer against the inherent heterogeneity across observational studies. Different covariates will vary in importance for different populations and circumstances, so a broad review like ours may benefit from pooling estimates from    . Meta-analysis of the association between sanitation access and hookworm infection [65,124,[126][127][128]. Note: Chongsuvivatwong et al [65]. reported on two separate studies in their 1996 article. doi:10.1371/journal.pmed.1001620.g010 Figure 11. Meta-analysis of the association between soap use and infection with any STH [70,73,129]. doi:10.1371/journal.pmed.1001620.g011 models that were adapted by researchers to best fit their data and contexts. There are many factors that could confound the relationship between WASH access or practices and STH prevalence, including socioeconomic status, age, and gender. Consideration of only crude associations would likely overstate the magnitude of effect for WASH exposures or even misinterpret the true direction of effect [99]. Limiting our focus to adjusted measures of effect reduces the number of eligible studies, which may impact the generalizability of our results. This strategy also amplifies the impact of selective reporting, since many authors reported only statistically significant adjusted estimates.
Evidence quality was typically ''low''-the default GRADE for observational research-meaning that our confidence in pooled effect estimates is limited, and that the true effect may be markedly Figure 12. Meta-analysis of the association between handwashing before eating and infection with A. lumbricoides [80,125,130]. doi:10.1371/journal.pmed.1001620.g012 Figure 13. Meta-analysis of the association between handwashing after defecation and infection with A. lumbricoides [66,80,116]. doi:10.1371/journal.pmed.1001620.g013 Figure 14. Meta-analysis of the association between handwashing after defecation and infection with any STH [73,80,112,116,131]. doi:10.1371/journal.pmed.1001620.g014 Figure 15. Meta-analysis of the association between wearing shoes and hookworm infection [65,118,132,133]. Note: Chongsuvivatwong et al. [65] reported on two separate studies in their 1996 article. doi:10.1371/journal.pmed.1001620.g015 different from the results reported here [40]. A much stronger case can be made for the benefit of hygiene because of the evidence provided by recent randomized controlled trials, but results from our meta-analyses suggest that the protective effect of hygiene practices on STH infection may be variable depending on context.
Publication bias also represents a concern. Five meta-analyses (piped water for any STH and A. lumbricoides, wearing shoes for hookworm and any STH, sanitation access for hookworm) showed evidence of publication bias in funnel plot assessments. However, two of those plots (piped water for A. lumbricoides and sanitation access for hookworm) showed that larger studies yielded more protective associations, suggesting that the results from those analyses may be underestimating the true relationship strength. This was unexpected-and possibly caused by the natural heterogeneity across observational studies-since larger studies are traditionally expected to show smaller magnitudes of effect. Heterogeneity creates great difficulty in assessing publication bias accurately with statistical tests, so it is impossible to know how pronounced publication bias may be throughout our metaanalyses [100].

Conclusion
A vibrant discussion continues in the literature about the role of MDA in measurably mitigating morbidity from STH infection at the population level [101][102][103][104][105][106]. MDA alone is unlikely to permanently interrupt STH transmission. Our review provides evidence that WASH is a valuable component for STH control strategies, but guidelines and targets for the integration of these approaches are needed. Increased attention towards WASH for STH also has great potential to catalyze synergies with integrated NTD control programs, while jointly elevating awareness of WASH and NTDs [5,28,107]. Additional high-quality research into the potential of integrated WASH interventions is merited, specifically on the complementarity of WASH and MDA. Recent and ongoing research continues to build an evidence-base that can guide policymaking and programmatic decisions [27,28,108]. Increased collaboration between the health and WASH sectors represents a key enterprise for the future of NTD control and elimination [109,110].     Editors' Summary Background. Worldwide, more than a billion people are infected with soil-transmitted helminths (STHs), parasitic worms that live in the human intestine (gut). These intestinal worms, including roundworm, hookworm, and whipworm, mainly occur in tropical and subtropical regions and are most common in developing countries, where personal hygiene is poor, there is insufficient access to clean water, and sanitation (disposal of human feces and urine) is inadequate or absent. STHs colonize the human intestine and their eggs are shed in feces and enter the soil. Humans ingest the eggs, either by touching contaminated ground or eating unwashed fruit and vegetables grown in such soil. Hookworm may enter the body by burrowing through the skin, most commonly when barefooted individuals walk on infected soil. Repeated infection with STHs leads to a heavy parasite infestation of the gut, causing chronic diarrhea, intestinal bleeding, and abdominal pain. In addition the parasites compete with their human host for nutrients, leading to malnutrition, anemia, and, in heavily infected children, stunting of physical growth and slowing of mental development.

Supporting Information
Why Was This Study Done? While STH infections can be treated in the short-term with deworming medication, rapid re-infection is common, therefore a more comprehensive program of improved water, sanitation, and hygiene (WASH) is needed. WASH strategies include improvements in water access (e.g., water quality, water quantity, and distance to water), sanitation access (e.g., access to improved latrines, latrine maintenance, and fecal sludge management), and hygiene practices (e.g., handwashing before eating and/or after defecation, water treatment, soap use, wearing shoes, and water storage practices). WASH strategies have been shown to be effective for reducing rates of diarrhea and other neglected tropical diseases, such as trachoma; however, there is limited evidence linking specific WASH access or practices to STH infection rates. In this systematic review and meta-analysis, the researchers investigate whether WASH access or practices lower the risk of STH infections. A systematic review uses predefined criteria to identify all the research on a given topic; a meta-analysis is a statistical method that combines the results of several studies.
What Did the Researchers Do and Find? The researchers identified 94 studies that included measurements of the relationship between WASH access and practices with one or more types of STHs. Meta-analyses of the data from 35 of these studies indicated that overall people with access to WASH strategies or practices were about half as likely to be infected with any STH. Specifically, a lower odds of infection with any STH was observed for those people who use treated water (odd ratio [OR] of 0.46), have access to sanitation (OR of 0.66), wear shoes (OR of 0.30), and use soap or have soap availability (OR of 0.53) compared to those without access to these practices or strategies. In addition, infection with roundworm was less than half as likely in those who practiced handwashing both before eating and after defecating than those who did not practice handwashing (OR of 0.38 and 0.45, respectively).

What Do These Findings Mean?
The studies included in this systematic review and meta-analysis have several shortcomings. For example, most were cross-sectional surveys-studies that examined the effect of WASH strategies on STH infections in a population at a single time point. Given this study design, people with access to WASH strategies may have shared other characteristics that were actually responsible for the observed reductions in the risk of STH infections. Consequently, the overall quality of the included studies was low and there was some evidence for publication bias (studies showing a positive association are more likely to be published than those that do not). Nevertheless, these findings confirm that WASH access and practices provide an effective control measure for STH. Controlling STHs in developing countries would have a huge positive impact on the physical and mental health of the population, especially children, therefore there should be more emphasis on expanding access to WASH as part of development guidelines and targets, in addition to short-term preventative chemotherapy currently used.
Additional Information. Please access these websites via the online version of this summary at http://dx.doi.org/10. 1371/journal.pmed.1001620.

N The US Centers for Disease Control and Prevention also
provides detailed information on roundworm, whipworm, and hookworm infections N The World Health Organization provides information on soil-transmitted helminths, including a description of the current control strategy N Children Without Worms (CWW) partners with Johnson & Johnson, GlaxoSmithKline, the World Health Organization, national ministries of health and education, non-governmental organizations, and others to promote treatment and prevention of soil-transmitted helminthiasis. CWW advocates a four-pronged, comprehensive control strategy-Water, Sanitation, Hygiene Education, and Deworming (WASHED) to break the cycle of reinfection N The Global Network for Neglected Tropical Diseases, an advocacy initiative dedicated to raising the awareness, political will, and funding necessary to control and eliminate the most common neglected tropical diseases, provides information on infections with roundworm (ascariasis), whipworm (trichuriasis), and hookworm N WASH for the Neglected Tropical Diseases is a repository of information on WASH and the neglected tropical diseases (NTDs) such as soil-transmitted helminthiasis, and features a resource titled ''WASH and the NTDs: A Manual for WASH Implementers.'' N Two international programs promoting water sanitation are the World Health Organization Water Sanitation and Health program and the World Health Organization/United Nations Childrens Fund Joint Monitoring Programme for Water Supply and Sanitation