The Effect of India’s Total Sanitation Campaign on Defecation Behaviors and Child Health in Rural Madhya Pradesh: A Cluster Randomized Controlled Trial 102486 Sumeet R. Patil1,2*, Benjamin F. Arnold2, Alicia L. Salvatore3, Bertha Briceno4, Sandipan Ganguly5, John M. Colford Jr.2, Paul J. Gertler2,6 1 Network for Engineering and Economics Research and Management (NEERMAN), Mumbai, Maharashtra, India, 2 School of Public Health, University of California, Berkeley, California, United States of America, 3 Stanford School of Medicine, Stanford University, Stanford, California, United States of America, 4 Water and Sanitation Program, the World Bank, Washington (D.C.), United States of America, 5 National Institute for Cholera and Enteric Diseases, Kolkata, West Bengal, India, 6 Haas School of Business, University of California, Berkeley, California, United States of America Abstract Background: Poor sanitation is thought to be a major cause of enteric infections among young children. However, there are no previously published randomized trials to measure the health impacts of large-scale sanitation programs. India’s Total Sanitation Campaign (TSC) is one such program that seeks to end the practice of open defecation by changing social norms and behaviors, and providing technical support and financial subsidies. The objective of this study was to measure the effect of the TSC implemented with capacity building support from the World Bank’s Water and Sanitation Program in Madhya Pradesh on availability of individual household latrines (IHLs), defecation behaviors, and child health (diarrhea, highly credible gastrointestinal illness [HCGI], parasitic infections, anemia, growth). Methods and Findings: We conducted a cluster-randomized, controlled trial in 80 rural villages. Field staff collected baseline measures of sanitation conditions, behaviors, and child health (May–July 2009), and revisited households 21 months later (February–April 2011) after the program was delivered. The study enrolled a random sample of 5,209 children ,5 years old from 3,039 households that had at least one child ,24 months at the beginning of the study. A random subsample of 1,150 children ,24 months at enrollment were tested for soil transmitted helminth and protozoan infections in stool. The randomization successfully balanced intervention and control groups, and we estimated differences between groups in an intention to treat analysis. The intervention increased percentage of households in a village with improved sanitation facilities as defined by the WHO/UNICEF Joint Monitoring Programme by an average of 19% (95% CI for difference: 12%– 26%; group means: 22% control versus 41% intervention), decreased open defecation among adults by an average of 10% (95% CI for difference: 4%–15%; group means: 73% intervention versus 84% control). However, the intervention did not improve child health measured in terms of multiple health outcomes (diarrhea, HCGI, helminth infections, anemia, growth). Limitations of the study included a relatively short follow-up period following implementation, evidence for contamination in ten of the 40 control villages, and bias possible in self-reported outcomes for diarrhea, HCGI, and open defecation behaviors. Conclusions: The intervention led to modest increases in availability of IHLs and even more modest reductions in open defecation. These improvements were insufficient to improve child health outcomes (diarrhea, HCGI, parasite infection, anemia, growth). The results underscore the difficulty of achieving adequately large improvements in sanitation levels to deliver expected health benefits within large-scale rural sanitation programs. Trial Registration: Clinicaltrails.gov NCT01465204 Please see later in the article for the Editors’ Summary. Citation: Patil SR, Arnold BF, Salvatore AL, Briceno B, Ganguly S, et al. (2014) The Effect of India’s Total Sanitation Campaign on Defecation Behaviors and Child Health in Rural Madhya Pradesh: A Cluster Randomized Controlled Trial. PLoS Med 11(8): e1001709. doi:10.1371/journal.pmed.1001709 Academic Editor: Paul R. Hunter, University of East Anglia, United Kingdom Received January 30, 2014; Accepted July 17, 2014; Published August 26, 2014 Copyright: ß 2014 World Bank. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC-BY 3.0 IGO), which permits unrestricted use, distribution, and reproduction in any medium, provided the original authors, source and license are attributed and any changes made are indicated. Data Availability: The authors confirm that all data underlying the findings are fully available without restriction. The study protocol, questionnaires, and access to data collected in the study are available upon registration at http://microdata.worldbank.org/ Funding: This work was supported by a grant from the Bill & Melinda Gates Foundation (www.gatesfoundation.org) to the Water and Sanitation Program of the World Bank. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. Competing Interests: BB was part of the research team and regular staff of the World Bank (under employment contract), which administered the funding through the Water and Sanitation Program. All other authors have declared that no competing interests exist. Abbreviations: BPL, below poverty line; CLTS, community led total sanitation; HCGI, highly credible gastrointestinal illness; IHL, individual household latrine; ITT, intention to treat; JMP, Joint Monitoring Programme; MUAC, mid-upper arm circumference; TSC, Total Sanitation Campaign; TSSM, Total Sanitation and Sanitation Marketing; WSP, Water and Sanitation Program (the World Bank). * Email: srpatil@neerman.org PLOS Medicine | www.plosmedicine.org 1 August 2014 | Volume 11 | Issue 8 | e1001709 RCT to Evaluate Health Effects of a Rural Sanitation Program in India Introduction support from the WSP under the TSSM project using a cluster- randomized controlled trial in 80 rural villages. The practice of open defecation is thought to be a major cause We hypothesized that the program would increase availability of of the persistent worldwide burden of diarrhea and enteric parasite individual household latrines (IHLs) and reduce the practice of infection among children ,5 years old [1]. Reducing open open defecation in a community through use of IHLs. On the basis defecation requires access to and use of improved sanitation of previous research [4–8], we further hypothesized that less open facilities, which are defined as facilities that prevent human feces defecation would: (i) reduce the quantity of feces in the from re-entering the environment [2]. In 2010, an estimated 47% environment that could contaminate shallow groundwater aqui- of the world’s population did not have access to improved fers, water distribution networks, and soil in the community, and sanitation facilities. India alone accounts for a third of those (ii) also reduce enteric pathogen transmission through flies, which without improved sanitation (814 million), nearly 60% of those are well-established vectors for transmission [9–11]. Conditional who practice open defecation (626 million) [2], and a quarter of on improvements in these intermediate outcomes, we hypothe- the world’s deaths from diarrheal diseases among children aged sized that children ,24 months at enrollment in intervention less than 5 years [3]. villages would have a lower prevalence of diarrhea, HCGI, enteric Observational studies of interventions that prevent human feces parasite infections, and anemia when measured after the from entering the environment have been shown to reduce intervention. Finally, we hypothesized that the program would diarrheal diseases [4,5] and enteric parasite infections [6–8]. Most improve average weight-for-age and height-for-age in these young of this research, however, has focused on the provision of sewerage children as a result of fewer symptomatic and asymptomatic systems in urban centers. Few studies have been conducted in enteric infections over longer exposure periods to improved rural areas of low-income countries where the provision and sanitation [12–16]. The above hypothesized causal chain between maintenance of networked sewerage is prohibitively expensive. the intervention and health outcomes is depicted in Figure 1. Consequently, most government and donor financing in the rural sanitation sector focuses on the provision of non-networked toilets. Methods Despite the wide scale deployment of such programs, to our knowledge there have been no published randomized trials to Ethics Statement measure the effect of rural sanitation programs on diarrheal The study is a part of larger six-country study commissioned by diseases, intestinal parasite infections, anemia, or growth in young the WSP. The study protocol was approved by the Western children. Institutional Review Board, Olympia, Washington, USA (study The objective of this study was to measure the effect of India’s number 1095420) and the Independent Ethics Committee, Total Sanitation Campaign (TSC) in rural Madhya Pradesh on Mumbai, India (IEC/09/11). The survey respondents provided household availability of improved sanitation facilities as defined a verbal consent after enumerators apprised them of study by WHO/UNICEF Joint Monitoring Programme (JMP) for water objectives, use of collected information, confidentiality, risks, and sanitation [2], open defecation behaviors of household benefits, and respondent rights. Written consent was not obtained members, water quality, and child health (diarrheal diseases, because of lower literacy and social norms that would deter highly credible gastrointestinal illness [HCGI], enteric parasite women (child caregivers were the main respondents) from signing infections, anemia, and growth). The TSC, scaled up to all districts any document without her husband’s or elders’ permission. in India and deployed to hundreds of millions of people, is possibly The protocol for the broader study was formally registered after the largest rural sanitation program in the world. As a part of their the completion of fieldwork because at the time the study was Total Sanitation and Sanitation Marketing (TSSM) project, the conceived pre-registration was not a well-known convention in the Water and Sanitation Program (WSP; the World Bank) provided field of development economics [17] (the study was originally capacity building support to ten districts of Madhya Pradesh to conceived by PJG and colleagues at the World Bank). The team strengthen the implementation of the program. In two of these ten agreed that a late registration was better than none at all. The districts, we studied the effects of the TSC implemented with original study protocol—established in 2008; before the baseline Figure 1. Hypothesized causal pathways for intervention impact and measurements. doi:10.1371/journal.pmed.1001709.g001 PLOS Medicine | www.plosmedicine.org 2 August 2014 | Volume 11 | Issue 8 | e1001709 RCT to Evaluate Health Effects of a Rural Sanitation Program in India survey—included another two districts from the state of Himachal In each of the 80 study villages, the field team listed and Pradesh where the WSP provided support under their TSSM mapped 200 households and randomly selected 25 households project as well [18]. However, the study in Himachal Pradesh was with at least one child ,24 months of age at enrollment. If a discontinued because it was impossible to retain a control group village had multiple sub-villages, then to avoid spreading the for the duration of the study period. The only substantive change sample too thin, the survey team selected the most populous two to to the protocol in Madhya Pradesh after the start of the trial was three sub-villages for the listing purposes. From the numbered list the increase in sample size planned for follow-up; we provide of eligible households, a random starting number was chosen and details in the section on Sample Size. The CONSORT checklist thereafter every nth household number was selected where n was (Text S1) and the follow-up study protocol (Text S2) are provided determined by dividing eligible number of households by 25. For as supplemental information. the follow-up survey we increased the sample size of households per village from 25 to 38 (see section on Sample Size). Additional Trial Design 100 to 150 households were listed and mapped before the follow- The study design was a cluster randomized controlled trial with up survey to select additional households. Figure 2 summarizes randomization at the village level and equal allocation to the two loss to follow-up in the original cohort and recruitment of new treatment arms. The study population included 80 villages from households in the follow-up survey. Because we conducted the two neighboring districts in Madhya Pradesh: Dhar and follow-up survey 21 months after baseline, the eligibility criteria for Khargone. The villages randomized to the intervention group newly enrolled households was that they had at least one child received the TSC program and villages in the control group did between the ages of 21 months and 45 months and were living in not receive the TSC until after the study. As a demand driven the village at the time of the baseline survey to be commensurate program, the district administration was duty bound to provide the with the eligibility criteria for the original cohort. Child caregivers program to the villages in the control group if they requested it were the main survey respondents, but household heads or other and if the funding was available. The district administration elders occasionally answered questions related to household agreed to provide the program to all control villages after the characteristics. completion of the study. The study measured outcomes, antici- pated confounders, and covariates at household and child levels Intervention Program both before and after the intervention in two survey waves. The India’s TSC, initiated in 1999, was an ambitious program with follow-up survey was administered to the same households who a goal to eliminate the practice of open defecation in India by participated in baseline data collection and additional households 2012. In 2012, the government transformed TSC into a new were included at follow-up (see the section on Sample Size for program named Nirmal Bharat Abhiyaan (Clean India Cam- details). paign). The TSC included subsidies for and promotion of IHLs that can safely confine feces (similar to JMP defined improved Study Population sanitation facilities), school sanitation and hygiene education, Table 1 describes population characteristics for the study region Anganwadi (preschool) toilets, and community sanitation com- relative to the state and national population on the basis of India’s plexes. The TSC also supported rural sanitary marts and 2011 Census. Overall, Madhya Pradesh is one of the less production centers to provide good quality but affordable material developed states of India, including its water and sanitation for toilet construction. Additionally, the TSC included several infrastructure. The study districts are more agricultural, with features such as ongoing social mobilization and behavior change higher proportion of marginalized population groups and lower activities at state, district, and village levels, flexible technology literacy than the state average, but with better water supply and options for toilets, and a community award called the Nirmal drainage infrastructure. IHL coverage (percentage of households Gram Puraskar (NGP) given to communities that were determined with access to IHL) in rural areas of study districts (19.2% in Dhar to be ‘‘open defecation free’’—defined as a community where all and 13% in Khargone) is comparable to the state average (13.1%) households have and use IHLs that can safely confine feces—and but much worse than the country average (30.7%). The IHLs are meet all of the other ‘‘total sanitation’’ requirements defined by the predominantly the types included in the JMP definition of Indian government. The NGP awards ranged from Rs 50,000 improved sanitation [2]. On average the IHL coverage across (US$1,000) to Rs 500,000 (US$10,000) for villages, up to Rs India increased by approximately 10% between 2001 and the 2,000,000 (US$40,000) for Blocks, and Rs 5,000,000 (US$100,000) 2011 Census. However, the change in the IHL coverage between for districts. 2001 and 2011 varied widely between states and between districts In Madhya Pradesh, the TSC was implemented with a within each state [19]. concurrent program named Nirmal Vatika (Clean House) under Study villages were selected in collaboration with the Madhya the National Rural Employment Guarantee Scheme to provide Pradesh state government. Madhya Pradesh is divided into 50 additional financial and material subsidies to households. TSC and districts, 313 Blocks, and 23,040 Gram Panchayats (referred to as Nirmal Vatika together provided at least Rs 4,200 (US$84) to ‘‘villages’’ in this manuscript). A Gram Panchayat is the smallest below poverty line (BPL) households in the village. The Indian Indian administrative unit and has a local elected body. The 80 Ministry of Rural Development classifies households as BPL using study villages were the independent units selected in three steps. characteristics such as land holdings, house type, consumer First, through a series of meetings and site visits, the state durables, and literacy [20]. BPL households were identified in government and the WSP selected two of 50 districts in Madhya this study by their ration card color (a document used to access Pradesh: Dhar and Khargone. Second, 11 of 13 Blocks from Dhar public food and grain distribution system). While the TSC and eight of nine Blocks from Khargone were selected for the provided subsidy of Rs 2,200 (US$44) to BPL households, Nirmal study. The remaining Blocks were excluded from the sample frame Vatika provided additional at least Rs 2,000 (US$40) to BPL and because all villages from these Blocks were earmarked for the TSC non-BPL households both to support IHL construction. These program, precluding the enrollment of control villages. Third, in costs were determined by the government to be adequate to each administrative Block the government identified villages where construct an offset two-pit latrine with water sealed squat plate and they were amendable to randomizing the TSC program. a brick walled room (which will be a JMP defined improved PLOS Medicine | www.plosmedicine.org 3 August 2014 | Volume 11 | Issue 8 | e1001709 RCT to Evaluate Health Effects of a Rural Sanitation Program in India Table 1. Descriptive statistics for India, Madhya Pradesh, and study districts, Census 2011. Indicators India Madhya Pradesh Dhar District Khargone District Population and occupations Total population 1,210,569,573 72,626,809 2,185,793 1,873,046 Rural population 833,463,448 52,557,404 1,772,572 1,574,190 Percent rural population 68.80 72.40 81.10 84.00 Percent 0–6 years children 14.60 15.80 16.90 16.60 (of rural population) Percent SCST (of rural population) 29.70 42.90 70.60 55.80 Percent literates (of .6 years 67.80 63.90 54.10 58.90 rural population) Percent of cultivators 33.00 38.30 42.90 38.30 (of rural workers) Percent of agriculture laborers 39.30 47.30 47.70 52.20 (of rural workers) Percent of other occupations 27.70 14.40 9.40 9.50 (of rural workers) Water and sanitation Number of rural households (RHHs) 167,826,730 11,122,365 339,844 309,603 Percent RHHs with permanent/good 45.90 33.40 38.90 31.50 house construction Percent RHHs with improved 84.30 74.10 79.90 84.20 drinking water sourcea Percent RHHs with access to tap 30.80 9.90 19.70 41.10 water (on premise or away) Percent RHHs with on premise 35.00 13.00 13.50 24.60 water source (any type) Percent RHHs with bathing rooms 45.00 34.00 38.10 50.40 Percent RHHs with closed drainage 5.70 2.10 3.20 4.20 Percent RHHs with open drainage 31.00 23.10 24.00 43.30 Latrine availability Percent RHHs with on-premise latrineb 30.73 13.12 19.17 13.00 Flush toilet connected to piped 2.20 0.80 1.43 1.15 sewer system Flush toilet connected to septic tank 14.70 8.32 12.91 9.07 Flush toilet connected to 2.53 1.26 1.25 0.70 other system Pit latrine with slab/ventilated 8.19 1.79 2.23 1.50 improved pit Pit latrine without slab/open pit 2.34 0.76 1.12 0.41 Toilets disposing waste to open drain 0.22 0.10 0.13 0.07 Serviced toilets where waste is 0.35 0.03 0.03 0.02 removed by humans Serviced toilets where waste is 0.19 0.07 0.06 0.08 removed by animals Percent RHHs with access to public toilets 1.94 0.46 0.68 0.36 Percent RHHs with no toilet/ 67.33 86.42 80.15 86.65 open site (2011) Percent RHHs with no toilet/ 78.10 91.10 86.40 91.10 open site (2001) a Improved drinking water sources include tap water, covered well, hand pump, and tube well as defined by Census of India, 2011. b On premise latrines are also referred to as IHLs. The first four types of toilets—flush toilets connected to sewer system, septic tank or other systems, and pit latrine slab and/or ventilated improved pit—are a subset of latrine types included in the definition of improved sanitation by WHO/UNICEF JMP for water and sanitation [2]. SCST, Schedule Caste or Schedule Tribe (marginalized population group); RHH, rural household. doi:10.1371/journal.pmed.1001709.t001 PLOS Medicine | www.plosmedicine.org 4 August 2014 | Volume 11 | Issue 8 | e1001709 RCT to Evaluate Health Effects of a Rural Sanitation Program in India Figure 2. CONSORT Flowchart: enrollment, intervention allocation, attrition, and addition of participants. doi:10.1371/journal.pmed.1001709.g002 sanitation facility), and this type of latrine was actively promoted in change methods [21]. The CLTS methodology involves a series of the study districts. community ‘‘triggering’’ exercises, led by an external facilitator Beginning in 2006, the WSP India office supported the TSC after building rapport with the community in the pre-triggering program under the TSSM project in ten districts in Madhya phase, which highlight the magnitude of the practice of open Pradesh. The WSP worked with local authorities to create an defecation, elicit shame and disgust, and mobilize community enabling environment for the TSC activities, to develop local action to end open defecation [21]. These triggering activities are implementation capacities at the district level, and to support the followed by community follow-up actions that are supported by use of monitoring systems to assess progress towards the TSC facilitators. Although the intervention used CLTS based tools for goals. WSP promoted and provided capacity building support to behavior change, it cannot be considered as a classical CLTS implement community-led total sanitation (CLTS) based behavior intervention. CLTS principles require that no hardware subsidies PLOS Medicine | www.plosmedicine.org 5 August 2014 | Volume 11 | Issue 8 | e1001709 RCT to Evaluate Health Effects of a Rural Sanitation Program in India be provided to individual households and specific latrine models using a previously published instrument [24]. HCGI—a more not be prescribed [21], whereas the intervention provided inclusive measure of enteric infection—was defined as any of the hardware subsidies to individual households to build offset pit following four conditions: (1) diarrhea; (2) vomiting; (3) soft or latrine designs approved under the Nirmal Vatika program. watery stool and abdominal cramps occurring together on any Provision of hardware subsidy as a post-construction incentive was day; or (4) nausea and abdominal cramps occurring together on advocated by the WSP, but the mechanisms of the convergence of any day [25–28]. We measured respiratory symptoms (constant Nirmal Vatika and the TSC essentially meant that the subsidies cough, pulmonary congestion, difficulty breathing, breaths per were released before and during but rarely after IHL construction. minute) and defined acute lower respiratory illness (ALRI) as The TSC program in the study areas was implemented by the constant cough or difficulty breathing and a raised respiratory rate village government (Gram Panchayat) with support from district [29]. We also measured bruising/abrasions and itchy skin/scalp to and block administration personnel or consultants. The study serve as negative control outcomes [30] to check for differential investigators and staff were not involved in program implemen- reporting bias in this unblinded trial [31,32]. tation. Anthropometry. We measured children ,24 months at enrollment for height, weight, and mid-upper arm circumference Outcome Definition and Measurement (MUAC) using a standardized anthropometry protocol [33,34]. The study measured outcomes using a combination of Pairs of trained anthropometrists measured child length/height to structured questionnaires and observations, sampling and testing the nearest 0.1 cm using a portable stadiometer (manufacturer: of drinking water, child anthropometry and specimen (stool and Seca); children ,24 months were measured in the recumbent blood) testing. GfK Mode Pvt Ltd. was contracted to conduct the (lying) position and older children (at follow-up) were measured fieldwork. Training and all field activities were overseen by the standing. Weight was measured to the nearest 0.1 kg using an study investigators (SRP, ALS). The baseline survey was conduct- electronic scale (manufacturer: Tanita); children unable to stand ed between 25 May and 18 July, 2009, and the follow-up survey were weighed in their caregiver’s arms and the caregiver’s weight was conducted between 23 February and 25 April, 2011. measured separately. MUAC was measured to the nearest 0.1 cm Questionnaires used in the follow-up survey were the same as using a pediatric measuring tape. All measurements were collected those used in the baseline survey with some additional questions to in duplicate and we used the average of the two measurements in measure program exposure and outcomes. The household the analysis. We excluded observations if the two measurements questionnaire collected information about household socioeco- differed by .10% (n = 21 [0.48%] for height, n = 85 [1.93%] for nomics, demographics, exposure to the TSC activities, water and weight, n = 23 [0.52%] for MUAC). We converted the anthropo- sanitation infrastructure, sanitation- and hygiene-related behav- metric measurements into Z-scores using the WHO’s 2006 growth iors, and health/diseases. Interviewers conducted standardized standards and the WHO publicly available Stata algorithm [35]. spot-check observations of dwelling sanitation and hygiene Anemia. If the caregiver provided informed consent, trained facilities. Defecation behavior was reported by adults during field staff conducted an in-field test for anemia for children private, in-home interviews. Main outcomes were defined as between the ages of 6 and 60 months using HemoCue (HemoCue follows. Ltd). We classified children as severely anemic if their hemoglobin Toilets, open defecation, hygienic conditions. We classi- concentration was ,7.0 g/dl, moderately anemic if their hemo- fied household sanitation facilities using questions and definitions globin concentration was 7.0–9.9 g/dl, and mildly anemic if their proposed by the JMP [2]. JMP-defined improved sanitation hemoglobin concentration was 10.0–11.9 g/dl [36]. Parents of includes flush/pour flush toilet connected to piped sewer, to septic children who were severely anemic were advised to visit the tank or to offset pit, ventilated improved pit latrine, on-pit latrine nearest health facility for medical attention. with slab and composting toilet that can hygienically separate Water quality. We collected 100 ml stored drinking water human excreta from human contact. However, it is possible that samples from a random sample of 404 households in the the households build rudimentary latrines that are not included in intervention and 403 households in the control groups, and also the JMP definition of improved sanitation. For example, in collected paired samples from the water source from which the addition to no facility or open defecation, the JMP defined households collected their drinking water (511 source samples). unimproved sanitation facilities include flush toilets disposing The water samples were collected in sterile containers, labeled, waste elsewhere, pit latrine without a slab (open hole), bucket and individually packed in a sterile plastic zip-lock cover provided latrine, and hanging latrine. We also report availability of all types by the laboratory. The sample collectors were provided with sterile of IHLs whether improved or unimproved to assess whether the gloves and trained to avoid cross-contamination of water and households moved up the sanitation ladder from no facility to containers. Water samples were stored and transported in ice some type of latrine even if unimproved. To assess defecation boxes and tested for Escherichia coli using membrane filtration behavior for men, women, and children (,5 years), interviewers (100 ml volume filtered) within 36 hours of collection at Envir- asked households separately for each group whether they openly ocare Laboratories Pvt Ltd, Mumbai. The laboratory used defecate daily/always, occasionally/seasonally, or never. Inter- HiCrome Agar (M1466) by HiMedia. Each incubation batch viewers also asked about child feces disposal using the standard included positive and negative control plates. Positive colonies of JMP question [2]; disposal in a toilet, a confined pit, or buried was E. coli were further confirmed with Triple Sugar Iron (TSI) agar classified as hygienic. Field staff also observed whether the IHLs (of test and group of Indole, Methyl red, Voges-Proskauer, and any type if present) were being used on the basis of worn path, Citrate tests (IMViC). Samples below the lower limit of detection closable door, odor, anal cleaning material, and water to flush. were imputed at 0.5 colony forming units (CFU) per 100 ml (half Field staff also recorded any observed human or animal feces in the limit of detection [37]), and samples beyond the upper limit of the household living area. detection were imputed at the limit of detection (200 CFU/ Caregiver reported illness. The study’s primary outcome 100 ml). was diarrhea and HCGI among children ,5 years old. We Child stool parasitology. At the follow-up survey, we defined diarrhea as $3 loose or watery stools in 24 hours, or a selected a random subsample of 1,150 households from 3,039 single stool with blood/mucus [22] with a 7-day recall period [23] households and collected a stool specimen from the oldest child PLOS Medicine | www.plosmedicine.org 6 August 2014 | Volume 11 | Issue 8 | e1001709 RCT to Evaluate Health Effects of a Rural Sanitation Program in India between 21 and ,60 months of age. All stool samples were evaluate any differential effect of attrition (loss to follow-up) preserved in 10% formalin and analyzed at the National Institute between baseline and follow-up, we compared baseline charac- for Cholera and Enteric Diseases in Kolkata. Lab technicians teristics of those present at follow-up with those lost to follow-up. tested the samples for soil transmitted helminthes (Ascaris We also compared the balance of baseline characteristics across lumbricoides, Trichuris trichiura, Ancylostoma duodenale, and treatment groups for individuals who were present at both baseline Necator americanus) and tapeworm helminthes (Hymenolepis and follow-up to determine whether attrition was differential by nana, Taenia sp., Diphyllobothrium latum) using the Kato-Katz treatment group. technique [38]. Our parameter of interest for all outcomes was the mean A separate aliquot was analyzed to test for protozoan infections difference between randomized groups. We conducted the analysis (Giardia lamblia, Cryptosporidium sp., Entamoeba histolytica) using using households and individuals as they were randomized a commercially available ELISA kit (TechLab) [39,40]. All (intention to treat [ITT]). We estimated differences between specimens were tested with a combination of microscopy, ELISA, groups using the following linear regression model: and PCR to achieve high levels of sensitivity and specificity. If a child tested positive for one of the protozoan infections using either Yijk ~azbTj zdXij zbk z"ijk ð1Þ microscopy or ELISA, the result was confirmed using isolated DNA from the ELISA positive samples followed by PCR- Where, Yijk is the outcome for individual i in village j and Block restriction fragment length polymorphism (RFLP) methods for k; Tj is the intervention indicator (1 for intervention, 0 for control); genotyping local isolates of giardia (b-giardin) [32], Cryptosporid- Xij are individual, household, and village level characteristics used ium (18s rRNA) [33], and E. histolytica (SSU rRNA) [34]. If a in adjusted analyses; bk are indicator variables for Blocks since sample tested positive by microscopy or ELISA but was not randomization was stratified at the Block level; and eijk is the error confirmed by molecular methods then the sample was classified as term. The parameter b estimates the ITT difference between the negative. randomized groups. In the adjusted analyses, we included the following covariates to improve precision: whether the household Sample Size head had attended school; whether the government categorized The study was originally designed to have 80% power to detect the household as Scheduled Caste or Tribe; child age; and child a 4.5 percentage point reduction in diarrhea prevalence among sex. Additionally, the adjusted models included three baseline children ,5 years old assuming 15% prevalence in the control characteristics found to be slightly imbalanced between groups group (or a 30% relative reduction) with a two-sided alpha of 5%, despite randomization. These included: percentage of households and an intra-class correlation of 0.105 [41]. These assumptions led in the village that used improved water sources; percentage of to a design with 40 clusters (villages) per arm and 25 households households in the village that were observed to have soap and with children ,24 months per cluster. After the commencement of water at the hand-washing place used after defecation; and mean the study but without knowledge of any study outcomes, we height-for-age Z-score of children in the village. To further assess decided to additionally power the trial to detect differences differential impacts of the program by important population between groups in height-for-age Z-scores on the basis of a subgroups, we re-estimated the effect of the intervention for hypothesis published on the possible effects of improved hygiene households with and without IHL (any type) at baseline, and and sanitation on child growth [14]. We reviewed measures of households below the official poverty line and the other variability and within-cluster correlation of height-for-age Z-scores households. (SD = 2.09, intra-class correlation = 0.17), and chose to increase Since we would expect behaviors and child health outcomes to be the within-cluster sample sizes from 25 to 38 households to ensure correlated within villages, all estimates used Huber-White robust the study had 80% power to detect differences of +0.2 Z in height- standard errors for the parameter b clustered at the village level [42] for-age. and reported p-values for the two sided t-test. Following guidance from Schulz and Grimes [43], we did not adjust p-values or Randomization confidence intervals for multiple comparisons because many of the The village-level randomization was stratified at the adminis- outcomes were highly correlated with one another (for example, trative Block level because the TSC implementation was correlation between primary outcomes diarrhea and HCGI = 0.78); coordinated at the Block level and we wanted to ensure that the nominal p-values should be interpreted with this in mind. All analyses treatment arms were evenly allocated between districts and were conducted using Stata v12 (Statacorp), and all primary analyses geographically stratified within districts. The randomization took were independently replicated by two investigators (SRP, BFA) from place in a public lottery led by study investigators. The Block TSC untouched datasets to final estimates. coordinators or their representatives picked the lottery ticket that assigned villages to treatment groups. Overall, we allocated a total Access to Protocol and Data of 20 villages in each district to the intervention and 20 to control The study protocol, questionnaires, and access to data collected (40 villages per arm). The program implementers and researchers in the study are available upon registration at http://microdata. were not blinded to the group assignment. Field interviewers were worldbank.org/. not informed of group assignment, but it was possible for them to identify intervention villages during interviews of Block officers or Results the village secretary. Enrolment, Baseline Balance, and Attrition Statistical Methods Figure 2 depicts the study participants flow. The baseline survey We checked the baseline balance in the observable character- enrolled a sample of 3,390 children ,5 years from 1,954 istics of the randomized groups. Due to highly comparable groups households from 80 villages. In the follow-up survey the sample at baseline and the large increase in our within-cluster sample size was increased to 5,209 children ,5 years from 3,039 between baseline and follow-up, our analysis focused on group households. As reported in Table 2, baseline covariates in comparisons post-intervention (using follow-up measures only). To intervention and control groups were well balanced with four PLOS Medicine | www.plosmedicine.org 7 August 2014 | Volume 11 | Issue 8 | e1001709 RCT to Evaluate Health Effects of a Rural Sanitation Program in India exceptions. First, 89% of the households in the intervention group IHL Coverage and Sanitation-Related Behaviors had access to improved water sources—tap/piped water, tube well Table 3 reports the intervention’s effect on IHL availability and protected dug wells—compared to 80% of households in the (JMP defined improved sanitation facilities and any type of IHLs) control group. In contrast, a larger proportion of control and open defecation behaviors by household members. The households (54%) were observed to have soap and water at intervention increased the coverage of JMP defined improved hand-washing locations used after defecation than in intervention sanitation facility by average 19 percentage points (95% CI 12%– households (44%). On average, more children were found to be 26%; p-value,0.001) in intervention villages compared to control anemic in the control group (93%) than in the intervention group villages (41.4% intervention versus 22.6% control). The interven- (88%). Finally, average height-for-age Z-scores were also slightly tion increased the coverage of any type of IHL facility by 20 imbalanced (21.38 intervention versus 21.81 control). percentage points (95% CI 13%–27%; p-value,0.001) in inter- Attrition was not differential by randomized group on the basis vention villages compared to control villages (44.1% intervention of observable characteristics (see Table S1). Of the 1,954 versus 24.2% control). These results indicate that available IHLs households enrolled at the baseline, 1,655 were located at the were predominantly JMP defined improved sanitation facilities 21-month follow-up survey (15% attrition) without any significant and very few rudimentary latrines or latrines defined as difference between the intervention (16%) and the control (15%) unimproved by the JMP were built. These results are consistent groups. Characteristics remained balanced between intervention with the TSC design that promoted latrine models that can safely and control groups in remaining households. contain the feces. Although on average fewer households in intervention villages Compliance to Randomization were likely to report daily open defecation compared to control The study measured intervention implementation in multiple villages for adult men (75% intervention versus 84% control; mean ways because of the complexity of the TSC program. These difference: 9.5%; p-value = 0.001), adult women (73% intervention measures included: reported implementation by Block coordina- versus 83% control; mean difference: 10%; p-value,0.001), and tors, expenditure of funds documented by official program children ,5 years (84% intervention versus 89% control; mean records, and interviews with local village officials. Out of 40 difference: 5%; p-value = 0.014), these reductions in reported open intervention villages, staff collected administrative information on defecation behaviors were smaller than the gains in IHL 39 villages from the TSC Block coordinators (government officers). availability. Amongst the 630 households in intervention villages The coordinators reported that 15/39 intervention villages that had JMP defined improved sanitation facilities at follow-up, received some CLTS activities, 33/39 villages applied for a 41% reported that adult men or women still practiced daily open NGP award prior to the follow-up survey. According to Block defecation; this same figure was 28% among the 339 control coordinators’ records, 25/39 villages had 100% households with village households at follow-up (not reported in results table). A IHLs, 11/39 villages had 80%–99% households with IHLs, and follow-up debriefing question to households who had IHL three of 39 villages with 37%–68% households with IHLs. Block identified that the main reasons for daily open defecation in spite coordinators also reported that 21/39 villages received 100% of of having IHL were culture, habit, or preference for defecating in the funds allocated under the TSC program, 12/39 villages open followed by inadequate water availability. received between 50% and 99%, and six of 39 villages received , 50% of their allocated funds. The latest disbursement of the TSC Drinking Water Quality funds was given to 36/39 intervention villages at least 4 to 5 In control villages, 82% (331/403) of household drinking water months before the follow-up survey, which would offer sufficient samples tested positive for E. coli compared to 77% (310/404) of time for IHLs to be constructed and used for 3 or more months. samples in intervention villages (mean difference: 5.5%; p- The study review meetings with Block coordinators also value = 0.050) (Table 3). Table S2 lists the distribution of positive identified that some control villages were contaminated during household samples by different E. coli contamination level the study period: TSC activities were initiated in eight control categories. villages within a few months of baseline survey and possibly in two Of 511 water source samples tested, 74% (208/280) of the additional control villages a few months prior to the follow-up sources in control villages and 70% (162/231) in intervention survey; official records were not available for control villages to villages tested positive for E. coli but the difference was not ascertain this information objectively. As per the follow-up survey statistically significant ( p-value = 0.143). in these ten contaminated villages, the household level coverage of JMP defined improved sanitation facilities increased from 17.4% at baseline to 41.4% at the follow-up, which is similar to the Caregiver Reported Illness program effect we observed in the intervention group. The Diarrhea prevalence did not differ between groups (7.4% household level coverage of JMP defined improved sanitation intervention versus 7.7% control; p-value = 0.687) (Table 4). facilities in uncontaminated control villages increased from 10.7% HCGI prevalence also did not differ between groups (11.5% to 16.2% in the same period. The study’s long follow-up period (21 intervention versus 12.0% control; p-value = 0.692) (Table 4). We months) and the highly publicized and politicized nature of the observed no significant differences between groups in negative TSC program may have contributed to this contamination. control caregiver-reported outcomes including bruising/abrasions Information from additional sources (village secretaries, school (1.4% intervention versus 1.3% control) and itchy skin/scalp teachers, Anganwadi [pre-school] workers in the village, and the (2.5% intervention versus 2.2% control) suggesting that differential rapid assessment from random sample of households) confirmed outcome reporting bias for diarrhea and HCGI was unlikely. that TSC activities translated into a higher recollection and knowledge of the TSC program in the intervention villages Enteric Parasite Infections compared to the control villages. We also found that households in In the subsample of 1,150 children with stool collection, 5.7% intervention villages were more aware of CLTS activities, had (66/1150) had helminth infections and the majority (50/66) were higher knowledge of TSC, and experienced more personal visits to Ascaris infections. All remaining infections were tapeworms; no convince them to build and use IHLs (Table 3). children were infected with T. trichiura or hookworm. We PLOS Medicine | www.plosmedicine.org 8 August 2014 | Volume 11 | Issue 8 | e1001709 RCT to Evaluate Health Effects of a Rural Sanitation Program in India Table 2. Baseline characteristics by randomized intervention groups, 2009. Characteristics Intervention (I) Control (C) a N Mean or Percent Na Mean or Percent Household characteristics Age in months for children ,5 yearsb 1,683 21.89 1,707 22.12 Age of HH head in years 976 45.34 978 43.18 Whether HH head went to school 954 49.90% 952 52.73% Government category of HH as BPL 976 34.53% 978 38.96% Government category of HH as schedule caste/tribe 935 69.73% 905 71.38% Pucca (better quality) HH construction 976 57.07% 978 60.43% Monthly HH income (Rupees) 976 11,293 978 11,022 WASH infrastructure and behaviors HH access to improved water source 976 89.24% 978 79.65% Reported drinking water treatment at home 976 68.34% 978 66.26% Interviewer observed soap and water at hand-washing 969 44.48% 972 54.22% place used post defecation PCG reports hand washing w/soap after fecal contact in last 978 61.76% 985 64.16% 24 hours Child nutrition Child ever breastfedc 1,026 99.03% 1,037 98.55% Child still breastfeedingc 1,013 91.21% 1,021 89.52% Iron pills, syrup givenc 1,019 7.36% 1,033 5.91% c Drugs for intestinal worms given in past 6 months 1,025 19.12% 1,033 15.97% Did receive VitA dose last 6 monthsc 1,013 37.41% 1,032 36.14% Sanitation Reported main sanitation facility is JMP defined improved 975 13.64% 978 12.37% sanitation facility Reported main sanitation facility is any type of IHL/is not open 975 18.36% 978 20.96% defecation Reported correct disposal of child feces 976 15.98% 978 13.39% Interviewer did not observe feces in living area around HH 973 41.11% 976 38.11% Water microbiology HH drinking water is contaminated with E. coli 172 95.93% 174 97.70% Health status Diarrhea 7-day prevalenceb 1,683 13.19% 1,707 12.13% HCGI 7-day prevalenceb 1,683 15.27% 1,707 15.06% ALRI 7-day prevalenceb 1,683 11.47% 1,707 10.13% Weight-for-age Z-scorec 957 22.20 943 22.18 Length/height-for-age Z-scorec 932 21.38 933 21.81 Arm circumference-for-age Z-scorec 921 21.31 895 21.33 c Weight-for-height Z-score 895 21.68 879 21.43 Anemic: Hb,110 g/lc 293 88.05% 329 92.71% a N is the base number of observations (the denominator) for the reported percentages or the sample size used to estimate the reported means. N is the number of households except for the variables measured at the child level (as indicated by b and c) where N is the number of children. N varies across different variables because of measurements in only a subset of the sample by design, non-response/refusal, and the loss due to measurement errors. b For children less than 60 months of age. c For children less than 24 months of age. ALRI, acute lower respiratory illness; CFU, colony forming units; Hb, Hemoglobin; HH, household; PCG, primary care giver; VitA, vitamin A; WASH, water, sanitation and hygiene. doi:10.1371/journal.pmed.1001709.t002 observed no difference in helminth prevalence between interven- intervention villages (18%) compared to children in control villages tion and control groups. Giardia infection was common, and (23%) (mean difference: 4.8%; p-value = 0.047). We detected no consistent with slightly improved water quality in the intervention Cryptosporidium infections in the study children, and a low group, we found lower Giardia prevalence among children in prevalence of E. histolytica (33 out of 1,150; 2.9%). PLOS Medicine | www.plosmedicine.org 9 August 2014 | Volume 11 | Issue 8 | e1001709 RCT to Evaluate Health Effects of a Rural Sanitation Program in India Table 3. Effect of the intervention on program outputs, behavioral outcomes, and water quality, 2011. Outputs and Outcomes Control Groupa Intervention Groupa ITT Unadjustedb ITT Adjustedc N Mean N Mean Difference [95% CI]d Difference [95% CI]d Program exposure HH received WASH message 1,511 0.272 1,523 0.295 0.023 [20.033 to 0.080] 0.000 [20.048 to 0.048] from mass media HH received WASH message 1,472 0.099 1,479 0.240 0.140 [0.097–0.183]*** 0.127 [0.081–0.172]*** from personal visits HH participated or is aware of 1,514 0.157 1,525 0.291 0.135 [0.083–0.186]*** 0.140 [0.089–0.191]*** CLTS activities HH knew of TSC/NGP 1,514 0.211 1,525 0.273 0.062 [0.011–0.114]** 0.053 [0.004–0.103]** Drinking water supply and hand-washing Infrastructure HH access to improved water 1,514 0.949 1,525 0.970 0.021 [20.001 to 0.043]* 0.007 [20.014 to 0.027] source Interviewer observed soap and 1,269 0.436 1,334 0.494 0.056 [20.006 to 0.118]* 0.052 [20.002 to 0.105]* water at hand-washing place used post defecation IHL access and sanitation behaviors HH with JMP defined improved 1,512 0.226 1,522 0.414 0.188 [0.118–0.258]*** 0.177 [0.107–0.246]*** sanitation facilities HH with any type of IHL 1,514 0.242 1,525 0.441 0.198 [0.126–0.270]*** 0.189 [0.116–0.263]*** Interviewer assessed that HH is 1,504 0.167 1,520 0.272 0.104 [0.047–0.161]*** 0.093 [0.042–0.144]*** using IHL (any type) Reported daily OD by men 1,514 0.841 1,525 0.746 20.095 [20.152 to 20.039]*** 20.087 [20.135 to 20.038]*** Reported daily OD by women 1,514 0.835 1,525 0.732 20.102 [20.159 to 20.045]*** 20.091 [20.141 to 20.041]*** Reported daily OD by children 1,514 0.892 1,525 0.839 20.053 [20.095 to 20.011]** 20.054 [20.088 to 20.020]*** Reported correct child feces 1,514 0.184 1,525 0.271 0.087 [0.045–0.129]*** 0.075 [0.036–0.113]*** disposal Interviewer did not observe 1,500 0.398 1,512 0.404 0.006 [20.045 to 0.057] 0.019 [20.026 to 0.065] human/animal feces in HH living area Drinking water quality E. coli present in household 403 0.821 404 0.767 20.055 [20.111 to 0.000]* 20.032 [20.101 to 0.036] drinking water E. coli present in the source 280 0.743 231 0.701 20.115 [20.269, 0.040] 20.016 [20.180, 0.149] from where household collected drinking water a The number of observations used to estimate means of the intervention and the control groups is the same as the number of observations used in ITT-unadjusted analysis. b Explanatory variables in the unadjusted model include the treatment assignment and indicator variables for Blocks. Therefore, the ITT effects for outcomes may not be exactly the difference between the listed mean in intervention and control groups in previous columns. c Because of missing adjustment variables data, the observations used in adjusted analysis are fewer than those used in unadjusted analysis. The number of observations used is seven to 113 less than that in unadjusted analysis. d Following the commonplace norms, statistical significance is indicated as: ***significant at a = 0.01; **significant at a = 0.05; *significant at a = 0.10. Please note that p-values are not adjusted for multiple comparisons following guidance from Schulz and Grimes [43]. CFU, colony forming units; HH, household; OD, open defecation; TSC/NGP, Total Sanitation Campaign/Nirmal Gram Puraskar; WASH, water, sanitation and hygiene. doi:10.1371/journal.pmed.1001709.t003 Anemia and Anthropometry Subgroup Results Anemia was prevalent in the study children (54%) and Table 5 presents the results of subgroup analyses of the effect of children were small according to international growth stan- the intervention on households with or without any type of IHL at dards (Table 4). However, we found no differences between baseline and BPL or non-BPL households. As expected, the the randomized groups in anemia prevalence or growth program had the largest improvements on JMP defined improved outcomes. sanitation facilities, IHL use as assessed by enumerators, and PLOS Medicine | www.plosmedicine.org 10 August 2014 | Volume 11 | Issue 8 | e1001709 RCT to Evaluate Health Effects of a Rural Sanitation Program in India Table 4. Effect of the intervention on health outcomes, 2011. Health Outcomes Control Groupa Intervention Groupa ITT Unadjustedb ITT Adjustedc N Mean N Mean Difference [95% CI]d Difference [95% CI]d Caregiver reported illness in the last 7 dayse Diarrhea 2,609 0.077 2,600 0.074 20.003 [20.019 to 0.013] 20.002 [20.019 to 0.015] HCGI 2,609 0.120 2,600 0.115 20.004 [20.026 to 0.017] 20.002 [20.024 to 0.020] Acute lower respiratory illness 2,609 0.128 2,600 0.163 0.038 [0.003–0.073]** 0.049 [0.009–0.089]** Enteric parasite infectionsf Any protozoan present 569 0.257 581 0.217 20.040 [20.089 to 0.008] 20.027 [20.082 to 0.029] Entamoeba histolytica present 569 0.025 581 0.033 0.008 [20.009 to 0.024] 0.009 [20.009 to 0.028] Giardia lamblia present 569 0.232 581 0.184 20.048 [20.096 to 20.001]** 20.036 [20.088 to 0.015] Any helminth present 569 0.056 581 0.059 0.001 [20.021 to 0.023] 20.005 [20.028 to 0.018] Ascaris lumbricoides present 569 0.044 581 0.043 20.002 [20.021 to 0.017] 20.011 [20.031 to 0.010] Any enteric parasite present 569 0.309 581 0.270 20.040 [20.087 to 0.006]* 20.032 [20.083 to 0.020] Anemia and anthropometrye Anemic: Hb,110 g/l 1,922 0.508 1,919 0.562 0.050 [20.011 to 0.110] 0.033 [20.030 to 0.096] Child weight (to 0.1 kg) 2,161 10.277 2,154 10.069 20.229 [20.492 to 0.033]* 20.130 [20.345 to 0.085] Child height (to 0.1 cm) 2,185 82.312 2,175 81.682 20.678 [21.362 to 0.006]* 20.242 [20.789 to 0.304] Child arm circumference (to 0.1 cm) 2,191 13.805 2,197 13.783 20.004 [20.145 to 0.138] 20.022 [20.167 to 0.123] Weight-for-age Z-score 2,161 21.833 2,154 21.921 20.095 [20.253 to 0.063] 20.094 [20.246 to 0.058] Length/height-for-age Z-score 2,185 22.155 2,175 22.189 20.034 [20.195 to 0.127] 20.040 [20.223 to 0.144] MUAC-for-age Z-score 2,191 21.337 2,197 21.337 0.020 [20.115 to 0.155] 20.022 [20.151 to 0.108] Weight-for-height Z-score 2,054 20.834 2,054 20.847 20.018 [20.195 to 0.160] 0.029 [20.142 to 0.199] BMI Z-score 2,052 20.604 2,052 20.664 20.062 [20.241 to 0.117] 20.019 [20.191 to 0.153] a The number of observations used to estimate means of the intervention and the control groups is the same as the number of observations used in ITT-unadjusted analysis. b Explanatory variables in the unadjusted model include the treatment assignment and indicator variables for Blocks. Therefore, the ITT effects for outcomes may not be exactly the difference between the listed mean in intervention and control groups in previous columns. c Because of missing adjustment variables data, the observations used in adjusted analysis are fewer than those used in unadjusted analysis. d Following the commonplace norms, statistical significance is indicated as: **significant at a = 0.05; *significant at a = 0.10. Please note that p-values are not adjusted for multiple comparisons following guidance from Schulz and Grimes [43]. e For children less than 60 months of age. f For children less than 60 months of age. The eldest child less than 60 months of age selected from a household. BMI, body mass index; Hb, hemoglobin; HH, household; OD, open defecation. doi:10.1371/journal.pmed.1001709.t004 reduced reported open defecation by household members in of JMP-defined improved sanitation facilities by a modest 19 households that did not have IHL (any type) at baseline and in percentage points in intervention villages compared to control BPL households. This finding is consistent with the TSC design (41% intervention versus 22% control; p-value,0.001). How- that targeted households without IHLs and offered larger IHL ever, the reductions in reported open defecation by adults were construction subsidies for BPL households. Among BPL house- even more modest: falling 9 to 10 percentage points (among holds, the intervention increased JMP defined improved sanitation men: 75% intervention versus 84% control; p-value = 0.001; facilities coverage by 30 percentage points (48% intervention among women: 73% intervention versus 83% control; p- versus 18% control; p-value,0.001) and it reduced open value,0.001), while reports of correct child feces disposal defecation among women by 17 percentage points (73% increased because of intervention by 9 percentage points (27% intervention versus 90% control; p-value,0.001). Despite larger intervention versus 18% control; p-value,0.001). The avail- improvements in these intermediate outcomes among BPL ability of IHL and the reductions in open defecation were households or households without IHL at baseline, we did not higher in the BPL household or households without any IHL at observe consistent improvement in health outcomes in these the time of baseline but we did not find consistent improve- subgroups (Table 5). ments in the multiple health outcomes in these subgroups. The less than universal or very high levels of IHL coverage in the intervention villages combined with relatively small behavior Discussion changes are consistent with our finding of no improvements in The TSC program, implemented with support of the WSP in child health outcomes including: diarrhea, enteric parasite Dhar and Khargone districts, increased household level coverage infection, growth, and anemia. PLOS Medicine | www.plosmedicine.org 11 August 2014 | Volume 11 | Issue 8 | e1001709 RCT to Evaluate Health Effects of a Rural Sanitation Program in India Table 5. Differential effect of the intervention by population subgroups, 2011. Characteristics Control Groupa Intervention Groupa ITT Unadjustedb ITT Adjustedc N Mean N Mean Difference [95% CI]d Difference [95% CI]d HH with JMP defined improved sanitation facilities All HH 1,512 0.224 1,522 0.414 0.189 [0.119–0.259]*** 0.178 [0.108–0.247]*** HH with IHL (any type) at 190 0.979 212 0.967 20.018 [20.056 to 0.020] 0.001 [20.027 to 0.029] baseline HH with no IHL (any type) at 1,319 0.114 1,297 0.318 0.202 [0.139–0.264]*** 0.209 [0.142–0.277]*** baseline BPL HH 551 0.181 452 0.476 0.307 [0.227–0.388]*** 0.320 [0.234–0.406]*** Non-BPL HH 961 0.249 1,070 0.388 0.135 [0.059–0.210]*** 0.108 [0.027–0.189]*** Reported daily OD by women All HH 1,514 0.835 1,525 0.732 20.102 [20.159 to 20.045]*** 20.091 [20.141 to 2 0.041]*** HH with IHL (any type) at 191 0.105 214 0.103 0.000 [20.078 to 0.077] 0.005 [20.070 to 0.080] baseline HH with no IHL (any type) at 1,320 0.941 1,297 0.837 20.101 [20.140 to 20.062]*** 20.097 [20.140 to 2 baseline 0.054]*** BPL HH 551 0.902 453 0.733 20.178 [20.241 to 20.115]*** 20.169 [20.233 to 2 0.105]*** Non-BPL HH 963 0.796 1,072 0.732 20.061 [20.129 to 0.006]* 20.029 [20.097 to 0.040] E. coli present in household drinking water All HH 403 0.821 404 0.767 20.055 [20.111 to 0.000]* 20.032 [20.101 to 0.036] HH with IHL (any type) at 54 0.796 60 0.817 0.004 [20.183 to 0.192] 20.003 [20.137 to 0.131] baseline HH with no IHL (any type) at 347 0.827 340 0.765 20.064 [20.121 to 20.006]** 20.055 [20.135 to 0.026] baseline BPL HH 147 0.803 111 0.739 20.069 [20.169 to 0.031] 20.076 [20.198 to 0.047] Non-BPL HH 256 0.832 293 0.778 20.054 [20.125 to 0.017] 20.042 [20.128 to 0.043] Diarrhea in the past 7 dayse All HH 2,609 0.077 2,600 0.074 20.003 [20.019 to 0.013] 20.002 [20.019 to 0.015] HH with IHL (any type) at 302 0.063 343 0.035 20.034 [20.072 to 0.003]* 20.037 [20.083 to 0.010] baseline HH with no IHL (any type) 2,302 0.079 2,231 0.080 0.001 [20.016 to 0.018] 0.003 [20.015 to 0.021] at baseline BPL HH 949 0.085 783 0.078 20.005 [20.031 to 0.021] 0.004 [20.022 to 0.029] Non-BPL HH 1,660 0.072 1,817 0.073 0.000 [20.019 to 0.019] 20.001 [20.023 to 0.021] f Ascaris lumbricoides infection All HH 569 0.044 581 0.043 20.002 [20.021 to 0.017] 20.011 [20.031 to 0.010] HH with IHL (any type) at 82 0.037 92 0.043 20.004 [20.051 to 0.043] 20.005 [20.087 to 0.078] baseline HH with no IHL (any type) at 487 0.045 482 0.041 20.004 [20.025 to 0.017] 20.013 [20.033 to 0.006] baseline BPL HH 221 0.045 160 0.044 0.008 [20.030 to 0.046] 0.023 [20.026 to 0.072] Non-BPL HH 348 0.043 421 0.043 20.001 [20.027 to 0.026] 20.022 [20.046 to 0.001]* f Giardia lamblia infection All HH 569 0.232 581 0.184 20.048 [20.096 to 20.001]** 20.036 [20.088 to 0.015] HH with IHL (any type) at 82 0.232 92 0.185 20.115 [20.221 to 20.008]** 20.060 [20.206 to 0.086] baseline PLOS Medicine | www.plosmedicine.org 12 August 2014 | Volume 11 | Issue 8 | e1001709 RCT to Evaluate Health Effects of a Rural Sanitation Program in India Table 5. Cont. Characteristics Control Groupa Intervention Groupa ITT Unadjustedb ITT Adjustedc N Mean N Mean Difference [95% CI]d Difference [95% CI]d HH with no IHL (any type) at 487 0.232 482 0.185 20.041 [20.094 to 0.011] 20.036 [20.094 to 0.023] baseline BPL HH 221 0.226 160 0.144 20.073 [20.141 to 20.005]** 20.059 [20.139 to 0.020] Non-BPL HH 348 0.236 421 0.200 20.041 [20.098 to 0.016] 20.027 [20.088 to 0.035] a The number of observations used to estimate means of the intervention and the control groups is the same as the number of observations used in ITT-unadjusted analysis. b Explanatory variables in the unadjusted model include the treatment assignment and indicator variables for Blocks. Therefore, the ITT effects for outcomes may not be exactly the difference between the listed mean in intervention and control groups in previous columns. c Because of missing adjustment variables data, the observations used in adjusted analysis are fewer than those used in unadjusted analysis. d Following the commonplace norms, statistical significance is indicated as: ***significant at a = 0.01; **significant at a = 0.05; *significant at a = 0.10. Please note that p-values are not adjusted for multiple comparisons following guidance from Schulz and Grimes [43]. e For children less than 60 months of age. f For children less than 60 months of age. The eldest child less than 60 months of age selected from a household. BPL, based on verification of household’s food ration card; HH, household; non-BPL, households who do not have/show BPL ration card; OD, open defecation. doi:10.1371/journal.pmed.1001709.t005 The study’s findings should be viewed as a measure of effectiveness levels of IHL coverage and use were reported in past evaluations of for this specific implementation of India’s TSC program in rural pilot programs and early implementations of India’s TSC [48,49]. Madhya Pradesh. By the end of the study in the intervention group, Within the broader water-sanitation-hygiene sector, the diffi- coverage of JMP defined improved sanitation facilities in a village culty of scaling up interventions that are efficacious when widely ranged between 5% and 79% households and percentage of adopted and properly used across a community is not unique households in a village reporting daily open defecation by adult to rural sanitation. Evaluations of large-scale hand-washing men ranged between 32% and 97% and that by adult women ranged promotion campaigns in Peru and Vietnam—part of the broader between 34% and 97%. It is unknown whether enteric pathogen risk research effort that included the present trial—found almost no is linearly or non-linearly related to the level of improved sanitation in improvements in hand-washing behavior and thus no downstream a community, and the intervention did not achieve the goal of impacts on child health [50,51]. Furthermore, the interim universal availability of IHLs or universal elimination of open evaluation of the national-level Sanitation Hygiene Education defecation during the study period. Therefore, our findings cannot and Water supply in Bangladesh program found very small speculate the child health outcomes for universal or higher levels of improvements in hygiene and sanitation outcomes, with no IHL availability or larger open defecation reductions that may be impacts on child health [52]. feasible under different contexts, program designs, or implementation The present evidence from the sector suggests that with few efficacy. Additional, forthcoming cluster randomized sanitation exceptions [53] scaled up sanitation and hygiene programs in intervention trials [44,45] may generate such evidence if they can rural settings have had difficulty in delivering the health benefits achieve adequately high latrine coverage and proportional reductions measured in small efficacy studies. Typically, the well-controlled in open defecation. efficacy trials can result in high enough levels of sanitation and This study presents a cautionary tale of how difficult it can be to hygiene infrastructure and behaviors necessary to deliver the achieve universal IHL coverage or elimination of open defecation health benefits, but the same levels of infrastructure or behavior for scaled up rural sanitation programs. The study documented change are not guaranteed to accrue to large-scale programs. clear evidence of more social mobilization, exposure to behavior From a public health perspective, these findings call into question change activities, and IHL construction in intervention villages the likelihood of the TSC in its current form to improve child compared to control villages. However, these intermediate outputs health. Still, the program may be valuable from the policy and of the TSC could not translate into high enough levels of IHL development perspective for reasons beyond public health, such availability and reductions in open defecation practice to deliver as the social benefits of sanitation (dignity, privacy, safety, and the health impacts. This evaluation was a part of a broader six- reduced burden of coping especially for women) accrued to country effort to also study large-scale sanitation promotion households that have and use IHLs, and the obligation of the programs in rural Indonesia and Tanzania, as well as large-scale government to provide access to sanitation as a recently hand-washing promotion programs in Peru, Vietnam, and recognized human right by the United Nations General Assembly Senegal. While the Tanzania results are forthcoming, the (Resolution number 64/292). As the next iteration of the TSC Indonesia study found even smaller increases in availability of program—named Nirmal Bharat Abhiyaan (Clean India Cam- JMP defined improved sanitation facilities and reductions in open paign)—continues, research efforts that focus on how to defecation following a large-scale sanitation campaign [46] that significantly increase the access to and use of IHLs would be was similar in design to the classical CLTS approach [21]. A particularly valuable to guide future program refinement. High recent cross-sectional survey in Orissa found more optimistic levels of IHL coverage and use should be demonstrated in pilot results—72% IHL availability following the TSC [47]—but programs before these program refinements are taken to national implementation was heterogeneous. Much less than universal scale. PLOS Medicine | www.plosmedicine.org 13 August 2014 | Volume 11 | Issue 8 | e1001709 RCT to Evaluate Health Effects of a Rural Sanitation Program in India Limitations Indian states had large growth in IHL availability between 2001 Like other effectiveness studies that measure the impact of large- and 2011 when the TSC program was active across India. In scale government programs, we faced the challenges typically not Madhya Pradesh, the TSC program was combined with Nirmal encountered in well-controlled efficacy trials such as imperfect Vatika that served to increase the IHL construction subsidies compliance with treatment assignment and poor fidelity of available to all eligible households. Additionally, the districts intervention implementation. We found that by 21 months of enrolled in this study received support from the WSP’s TSSM follow-up, none of the intervention villages achieved the program project to build capacity for creating an enabling environment, goal of 100% households having and using IHLs that can safely record keeping and monitoring, and implementing CLTS-based confine feces; the average household level coverage of JMP defined behavior change approaches. Therefore, the behavior change improved sanitation facilities was 40% (range: 5%–79%). The approaches in the study districts were arguably more intensive reasons for the gap between the officials monitoring records of the than those in the rest of Madhya Pradesh. However, this study TSC and the actual status are discussed elsewhere [54]. The Block should be not viewed as an evaluation of the CLTS approach as coordinators also identified that at least eight and possibly ten advocated by its practitioners [21] because the intervention only control villages received the TSC program. ITT estimates of used CLTS behavior change tools and did not follow the key program impacts with imperfect compliance will underestimate principles of CLTS such as not providing hardware subsidy and the effect possible under perfect compliance. not prescribing latrine models. Another challenge in trials where study investigators have limited control over the program implementation, is significant Conclusions deviations in the actual implementation timeline compared to the This 80 village study in rural Madhya Pradesh represents the timeline on which the evaluation study is based. While the planned first published large-scale, randomized evaluation of India’s follow-up period from the baseline was 18 months in this study, the TSC to measure and report outcomes at all stages of the causal actual follow-up measurement at 21 months was the latest possible chain (Figure 1). While the TSC program in rural Madhya point we could measure outcomes under the possibility of program Pradesh implemented with support from the WSP increased expansion into control villages and contractual constraints with the the household level availability of JMP defined sanitation evaluation funding. Although it was possible that impacts on facilities (+19%) and to a lesser extent reduced open defecation diarrheal diseases could begin relatively soon after intervention, as (210%), these improvements were insufficient to improve child documented in short-duration efficacy trials [55], we would expect health outcomes (diarrhea, parasite infections, anemia, impacts on enteric parasite infection, anemia, and growth to growth). Despite the limitations of the present study, including potentially accrue more slowly. short follow-up and evidence for contamination in the control The limited length of follow-up could have also influenced our group, the results underscore the challenge of achieving estimates of the program’s effect on IHL availability and use. adequately large levels of improvements in sanitation to deliver Longer follow-up could have led to potentially higher levels of the expected health benefits within the scaled-up rural IHL coverage or, conversely, lower levels of use (if IHLs are not sanitation programs. maintained). Despite this limitation, our estimates of IHL coverage and reported use are broadly consistent with other independent measures following rural sanitation programs in Supporting Information India [47–49]. For example, Barnard and colleagues [47] found Table S1 Analysis of balance in the baseline panel after that 4 to 6 years after TSC implementation in Orissa that 53% of attrition, Madhya Pradesh, 2009 households with an IHL reported some individuals still practiced (DOC) open defecation. In the present study, 41% of men and 38% of women from the intervention group who have JMP defined Table S2 Distribution of E. coli contamination in improved sanitation facilities reported practicing daily open household drinking water defecation. (DOCX) Self-reported outcomes can be subject to differential, biased Text S1 CONSORT checklist reporting in unblinded trials [31,32]. Therefore, in addition to self- (DOCX) reported illnesses, we included several objective child health measurements in this study (parasite infections, anemia, anthro- Text S2 Follow-up study protocol pometry). However, we did not include objective measures of (DOC) sanitation behaviors (disposal of child feces, IHL use, and open defecation). To the extent that our measurements of reported Acknowledgments outcomes were subject to courtesy bias, we may have over- We thank members of WSP, Eduardo Perez, Ajith Kumar, Upneet Singh, estimated IHL use or under-estimated open defecation prevalence K. Arokiam, and Christopher Juan Costain, for information about the in the study population. Furthermore, if the bias was differential by program and support throughout the study. We are grateful to the treatment group, then we would expect the study to have over- Department of Rural Development of Madhya Pradesh and the District estimated the improvements due to intervention because we would Collectorates (Dhar and Khargone) for implementing the TSC as per the expect the intervention households to be more sensitized to the study design, sharing their monitoring records and other secondary data, stigma of open defecation. Measures of IHL use could be and providing necessary administrative support. Lia Fernald and Patricia improved in future sanitation studies through the use of passive Kariger provided support developing the anthropometry and anemia sensors mounted in the latrine [56,57]. measurement protocols and training the field teams. Nilesh Amritkar (Envirocare Laboratories Pvt Ltd) supervised water sample analysis. Jack Molyneaux provided guidance on initial study design. Members of GfK Generalizability Mode, Urmil Dosaj, Prem Talwar, Subhabrata Guha, and Ashutosh Rana, There is wide variation in TSC implementation within India, coordinated and supervised the field activities. Finally, we are indebted to and it remains possible that the TSC program was more or less the many field staff who administered the surveys and families who successful in other states [19]. We note, however, that very few participated in the study. PLOS Medicine | www.plosmedicine.org 14 August 2014 | Volume 11 | Issue 8 | e1001709 RCT to Evaluate Health Effects of a Rural Sanitation Program in India Author Contributions criteria for authorship read and met: SRP BFA ALS BB SG JMC PJG. Agree with manuscript results and conclusions: SRP BFA ALS BB SG Conceived and designed the experiments: SRP BFA ALS SG JMC PJG. JMC PJG. Designed and supervised laboratory analysis of stool samples: Performed the experiments: SRP ALS. Analyzed the data: SRP BFA. SG. Wrote the first draft of the manuscript: SRP BFA. Contributed to the writing of the manuscript: SRP BFA ALS BB SG JMC PJG. ICMJE References 1. Mara D, Lane J, Scott B, Trouba D (2010) Sanitation and health. PLoS Med 7: consumption of drinking water meeting current microbiological standards. e1000363. Am J Public Health 81: 703–708. 2. UNICEF, WHO (2012) World Health Organization and United Nations 26. Colford JM, Wade TJ, Sandhu SK, van der Laan MJ, Brookhart MA, et al. Children’s Fund Joint Monitoring Programme for Water Supply and Sanitation (2005) A randomized, controlled trial of in-home drinking water intervention to (JMP). Progress on Drinking Water and Sanitation: 2012 Update. New York and reduce gastrointestinal illness. Am J Epidemiol 161: 472–482. Geneva: UNICEF and WHO. 27. Colford JM, Hilton JF, Wright CC, Arnold BF, Saha S, et al. (2009) The 3. Unicef, WHO (2009) Diarrhoea: Why children are still dying and what can be Sonoma Water Evaluation Trial: a randomized drinking water intervention trial done. New York and Geneva: Unicef/WHO. to reduce gastrointestinal illness in older adults. Am J Public Health 99: 1988– 4. Clasen TF, Bostoen K, Schmidt W-P, Boisson S, Fung IC-H, et al. (2010) 1995. Interventions to improve disposal of human excreta for preventing diarrhoea. 28. Arnold B, Arana B, Ma ¨ usezahl D, Hubbard A, Colford JM (2009) Evaluation of Cochrane Database Syst Rev 6: CD007180. a pre-existing, 3-year household water treatment and handwashing intervention 5. Norman G, Pedley S, Takkouche B (2010) Effects of sewerage on diarrhoea and in rural Guatemala. Int J Epidemiol 38: 1651–1661. enteric infections: a systematic review and meta-analysis. Lancet Infect Dis 10: 29. Gove S (1997) Integrated management of childhood illness by outpatient health 536–544. workers: technical basis and overview. The WHO Working Group on 6. Barreto ML, Genser B, Strina A, Teixeira MG, Assis AMO, et al. (2010) Impact Guidelines for Integrated Management of the Sick Child. Bull World Health of a city-wide sanitation programme in Northeast Brazil on intestinal parasites Organ 75 Suppl 1: 7–24. infection in young children. Environ Health Perspect. Available: http://www. 30. Lipsitch M, Tchetgen ET, Cohen T (2010) Negative controls: a tool for detecting ncbi.nlm.nih.gov/pubmed/20705544. Accessed 20 November 2010. confounding and bias in observational studies. Epidemiology 21: 383–388. 7. Mascarini-Serra LM, Telles CA, Prado MS, Mattos SA, Strina A, et al. (2010) 31. Wood L, Egger M, Gluud LL, Schulz KF, Jni P, et al. (2008) Empirical evidence Reductions in the prevalence and incidence of geohelminth infections following of bias in treatment effect estimates in controlled trials with different a city-wide sanitation program in a Brazilian Urban Centre. PLoS Negl Trop interventions and outcomes: meta-epidemiological study. BMJ 336: 601–605. Dis 4: e588. 32. Schmidt W-P, Cairncross S (2009) Household water treatment in poor 8. Ziegelbauer K, Speich B, Ma ¨ usezahl D, Bos R, Keiser J, et al. (2012) Effect of populations: is there enough evidence for scaling up now? Environ Sci Technol sanitation on soil-transmitted helminth infection: systematic review and meta- 43: 986–992. analysis. PLoS Med 9: e1001162. 33. Cogill B (2003) Anthropometric indicators measurement guide. Washington 9. Chavasse DC, Shier RP, Murphy OA, Huttly SR, Cousens SN, et al. (1999) (D.C.): Food and Nutrition Technical Assistance Project, Academy for Impact of fly control on childhood diarrhoea in Pakistan: community- Educational Development. randomised trial. Lancet 353: 22–25. 34. De Onis M, Onyango AW, Broeck JV den, Chumlea WC, Martorell R (2004) 10. Emerson PM, Lindsay SW, Alexander N, Bah M, Dibba S-M, et al. (2004) Role Measurement and standardization protocols for anthropometry used in the of flies and provision of latrines in trachoma control: cluster-randomised construction of a new international growth reference. Food Nutr Bull 25: S27– controlled trial. Lancet 363: 1093–1098. S36. 11. Emerson PM, Simms VM, Makalo P, Bailey RL (2005) Household pit latrines as 35. WHO (2006) WHO Child Growth Standards: length/height-for-age, weight- a potential source of the fly Musca sorbens–a one year longitudinal study from for-age, weight-for-length, weight-for-height and body mass index-for-age: The Gambia. Trop Med Int Health 10: 706–709. Methods and developments. Geneva: WHO. 12. Checkley W, Buckley G, Gilman RH, Assis AM, Guerrant RL, et al. (2008) 36. WHO (2011) Haemoglobin concentrations for the diagnosis of anaemia and Multi-country analysis of the effects of diarrhoea on childhood stunting. assessment of severity. Vitamin and Mineral Nutrition Information System. Int J Epidemiol 37: 816–830. Geneva: WHO. 13. Schmidt W-P, Boisson S, Genser B, Barreto ML, Baisley K, et al. (2010) Weight- 37. Helsel DR (1990) Less than obvious - statistical treatment of data below the for-age z-score as a proxy marker for diarrhoea in epidemiological studies. detection limit. Environ Sci & Technol 24: 1766–1774. J Epidemiol Community Health 64: 1074–1079. 38. Katz N, Chaves A, Pellegrino J (1972) A simple device for quantitative stool 14. Humphrey JH (2009) Child undernutrition, tropical enteropathy, toilets, and thick-smear technique in Schistosomiasis mansoni. Rev Inst Med Trop Sao handwashing. Lancet 374: 1032–1035. Paulo 14: 397–400. 15. Prendergast A, Kelly P (2012) Enteropathies in the developing world: neglected 39. Garcia LS, Shimizu RY (1997) Evaluation of nine immunoassay kits (enzyme effects on global health. Am J Trop Med Hyg 86: 756–763. immunoassay and direct fluorescence) for detection of Giardia lamblia and 16. Checkley W, Gilman RH, Black RE, Epstein LD, Cabrera L, et al. (2004) Effect Cryptosporidium parvum in human fecal specimens. J Clin Microbiol 35: 1526– of water and sanitation on childhood health in a poor Peruvian peri-urban 1529. community. Lancet 363: 112–118. 40. Den Hartog J, Rosenbaum L, Wood Z, Burt D, Petri, Jr WA (2013) Diagnosis of 17. Miguel E, Camerer C, Casey K, Cohen J, Esterling KM, et al. (2014) Promoting multiple enteric protozoan infections by enzyme-linked immunosorbent assay in transparency in social science research. Science 343: 30–31. the guatemalan highlands. Am J Trop Med Hyg 88: 167–171. 18. Salvatore AL, Patil SR (2011) Scaling up rural sanitation: findings from the 41. Luby SP, Agboatwalla M, Painter J, Billhimer WL, Keswick BH, et al. (2006) impact evaluation baseline survey in Madhya Pradesh, India. The Water and Combining drinking water treatment and hand washing for diarrhoea Sanitation Program, The World Bank. Washington (D.C.): The World Bank. prevention, a cluster randomised controlled trial. Trop Med Int Health 11: 19. Ghosh A, Cairncross S (2014) The uneven progress of sanitation in India. 479–489. J Water, Sanitation & Hygiene for Devel 4: 15–22. 42. Freedman DA (2006) On the so-called ‘‘Huber sandwich estimator’’ and ‘‘robust 20. Report of the expert group to recommend the detailed methodology for standard errors.’’ Am Stat 60: 299–302. identifcation of families living below poverty line in the urban areas (2012). 43. Schulz KF, Grimes DA (2005) Multiplicity in randomised trials I: endpoints and Planning Commission, Perspective Planning Division, Government of India. treatments. Lancet 365: 1591–1595. Available: http://planningcommission.nic.in/reports/genrep/rep_hasim1701. 44. Clasen T, Boisson S, Routray P, Cumming O, Jenkins M, et al. (2012) The effect pdf. Accessed 14 April 2014. of improved rural sanitation on diarrhoea and helminth infection: design of a 21. Kar K, Chambers R (2008) Handbook on Community-Led Total Sanitation. cluster-randomized trial in Orissa, India. Emerg Themes Epidemiol 9: 7. Plan International (UK). Available: http://www.communityledtotalsanitation. 45. Arnold BF, Null C, Luby SP, Unicomb L, Stewart CP, et al. (2013) Cluster- org/resource/handbook-community-led-total-sanitation. Accessed 12 February randomised controlled trials of individual and combined water, sanitation, 2013. hygiene and nutritional interventions in rural Bangladesh and Kenya: the 22. Baqui AH, Black RE, Yunus M, Hoque AR, Chowdhury HR, et al. (1991) WASH Benefits study design and rationale. BMJ Open 3: e003476. Methodological issues in diarrhoeal diseases epidemiology: definition of 46. Cameron L, Shah M, Olivia S (2013) Impact evaluation of a rural sanitation diarrhoeal episodes. Int J Epidemiol 20: 1057–1063. behavior change project in Indonesia. Washington (D.C.): The World Bank. 23. Arnold BF, Galiani S, Ram PK, Hubbard AE, Bricen ˜o B, et al. (2013) Optimal 47. Barnard S, Routray P, Majorin F, Peletz R, Boisson S, et al. (2013) Impact of recall period for caregiver-reported illness in risk factor and intervention studies: Indian Total Sanitation Campaign on latrine coverage and use: a cross-sectional a multicountry study. Am J Epidemiol 177: 361–370. study in Orissa three years following programme implementation. PLoS One 8: 24. Goldman N, Vaughan B, Pebley AR (1998) The use of calendars to measure e71438. child illness in health interview surveys. Int J Epidemiol 27: 505–512. 48. Pattanayak SK, Yang JC, Dickinson KL, Poulos C, Patil SR, et al. (2009) Shame 25. Payment P, Richardson L, Siemiatycki J, Dewar R, Edwardes M, et al. (1991) A or subsidy revisited: social mobilization for sanitation in Orissa, India. Bull randomized trial to evaluate the risk of gastrointestinal disease due to World Health Organ 8: 580–587. PLOS Medicine | www.plosmedicine.org 15 August 2014 | Volume 11 | Issue 8 | e1001709 RCT to Evaluate Health Effects of a Rural Sanitation Program in India 49. Arnold BF, Khush RS, Ramaswamy P, London AG, Rajkumar P, et al. (2010) 53. Biran A, Schmidt W-P, Varadharajan KS, Rajaraman D, Kumar R, et al. (2014) Causal inference methods to study nonrandomized, preexisting development Effect of a behaviour-change intervention on handwashing with soap in India interventions. Proc Natl Acad Sci U S A 107: 22605–22610. (SuperAmma): a cluster-randomised trial. Lancet Glob Health 2: e145–e154. 50. Galiani S, Gertler P, Orsola-Vidal A (2012) Promoting handwashing behavior in 54. Hueso A, Bell B (n.d.) An untold story of policy failure: the Total Sanitation peru: the effect of large-scale mass-media and community level interventions. Campaign in India. Water Policy 15: 1001–1017. World Bank Policy Research Working Paper 6257. Washington (D.C.): The 55. Waddington H, Snilstveit B (2009) Effectiveness and sustainability of water, World Bank. sanitation, and hygiene interventions in combating diarrhoea. J Dev Eff 1: 295– 51. Chase C, Do Q-T (2012) Handwashing behavior change at scale: evidence from 335. a randomized evaluation in Vietnam. World Bank Policy Research Working 56. Clasen T, Fabini D, Boisson S, Taneja J, Song J, et al. (2012) Making sanitation Paper 6207. Washington (D.C.): The World Bank. count: Developing and testing a device for assessing latrine use in low-income 52. Huda TMN, Unicomb L, Johnston RB, Halder AK, Sharker MAY, et al. (2012) settings. Environ Sci Technol 46: 3295–3303. Interim evaluation of a large scale sanitation, hygiene and water improvement 57. Thomas EA, Barstow CK, Rosa G, Majorin F, Clasen T (2013) Use of remotely programme on childhood diarrhea and respiratory disease in rural Bangladesh. reporting electronic sensors for assessing use of water filters and cookstoves in Soc Sci Med 75: 604–611. rwanda. Environ Sci Technol 47: 13602–13610. PLOS Medicine | www.plosmedicine.org 16 August 2014 | Volume 11 | Issue 8 | e1001709 RCT to Evaluate Health Effects of a Rural Sanitation Program in India Editors’ Summary Background. Diarrheal diseases are linked with the deaths implementation was over in the intervention villages. A of hundreds of thousands of young children each year in random subsample of children was also tested for infection resource-limited countries. Infection with enteric pathogens with enteric parasites. The intervention increased the (organisms such as bacteria, viruses, and parasites that percentage of households in a village with improved infect the human intestine or gut) also affects the health sanitation facilities by 19% on average. Specifically, 41% of and growth of many young children in these countries. A households in the intervention villages had improved major contributor to the transmission of enteric pathogens latrines on average compared to 22% of households in the is thought to be open defecation, which can expose control villages. The intervention also decreased the propor- individuals to direct contact with human feces containing tion of adults who self-reported open defecation from 84% infectious pathogens and also contaminate food and to 73%. However, the intervention did not improve child drinking water. Open defecation can be reduced by health measured on the basis of multiple health outcomes, ensuring that people have access to and use toilets or including the prevalence of gastrointestinal illnesses and latrines. Consequently, programs have been initiated in intestinal parasite infections, and growth. many resource-limited countries that aim to reduce open defecation by changing behaviors and by providing What Do These Findings Mean? These findings indicate technical and financial support to help households build that in rural Madhya Pradesh, the TSC implemented with improved latrines (facilities that prevent human feces from support from the WSP only slightly increased the availability re-entering the environment such as pit latrines with sealed of individual household latrines and only slightly decreased squat plates; an example of an unimproved facility is a the practice of open defecation. Importantly, these findings simple open hole). However, in 2011, according to the show that these modest improvements in sanitation and in WHO/UNICEF Joint Monitoring Programme for Water defecation behaviors were insufficient to improve health Supply and Sanitation, more than 1 billion people (15% of outcomes among children. The accuracy of these findings the global population) still defecated in the open. may be limited by various aspects of the study. For example, several control villages actually received the intervention, Why Was This Study Done? Studies of sewerage system which means that these findings probably underestimate the provision in urban areas suggest that interventions that effect of the intervention under perfect conditions. Self- prevent human feces entering the environment reduce reporting of defecation behavior, availability of sanitation diarrheal diseases. However, little is known about how rural facilities, and gastrointestinal illnesses among children may sanitation programs, which usually focus on providing stand- also have biased these findings. Finally, because TSC alone sanitation facilities, affect diarrheal disease, intestinal implementation varies widely across India, these findings parasite infections, anemia (which can be caused by parasite may not apply to other Indian states or variations in the TSC infections), or growth in young children. Governments and implementation strategies. Overall, however, these findings international donors need to know whether large-scale rural highlight the challenges associated with achieving large sanitation programs improve child health before expending enough improvement in access to sanitation and corre- further resources on these interventions or to identify an spondingly large reductions in the practice of open urgency to improve the existing program design or defecation to deliver health benefits within large-scale rural implementation so that they deliver the health impact. In sanitation programs. this study, the researchers investigate the effect of India’s Total Sanitation Campaign (TSC) on the availability of Additional Information. Please access these websites via individual household latrines, defecation behaviors, and the online version of this summary at http://dx.doi.org/10. child health in rural Madhya Pradesh, one of India’s less 1371/journal.pmed.1001709. developed states. Sixty percent of people who practice open defection live in India and a quarter of global child deaths N This study is further discussed in a PLOS Medicine from diarrheal diseases occur in the country. India’s TSC, Perspective by Clarissa Brocklehurst which was initiated in 1999, includes activities designed to change social norms and behaviors and provides technical N A PLOS Medicine Collection on water and sanitation is available and financial support for latrine building. So far there are no published studies that rigorously evaluated whether the TSC N The World Health Organization (WHO) provides informa- tion on water, sanitation, and health (in several languages), improved child health or not. on diarrhea (in several languages), and on intestinal parasites (accessed through WHO’s web page on neglect- What Did the Researchers Do and Find? A cluster ed tropical diseases); the 2009 WHO/UNICEF report randomized controlled trial randomly assigns groups of ‘‘Diarrhea: why children are still dying and what can be people to receive the intervention under study and done’’, is available online for download compares the outcomes with a control group that does not receive the intervention. The researchers enrolled 5,209 N The WHO/UNICEF Joint Monitoring Programme for Water children aged under 5 years old living in 3,039 households in Supply and Sanitation monitors progress toward improved 80 rural villages in Madhya Pradesh. Half of the villages (40), global sanitation; its 2014 update report is available online chosen at random, were included in the TSC (the interven- tion). Field staff collected data on sanitation conditions, N The children’s charity UNICEF, which protects the rights of children and young people around the world, provides defecation behaviors, and child health from caregivers in information on water, sanitation, and hygiene, and on each household at the start of the study and after the TSC diarrhea (in several languages) PLOS Medicine | www.plosmedicine.org 17 August 2014 | Volume 11 | Issue 8 | e1001709