An estimated 2.2 million children under the age of 5 years die from diarrheal disease each year. Most of the burden of diarrheal disease is thought to be preventable with improvements in sanitation, water quality, and hygiene. Large scale interventions promoting these behaviours have either not been rigorously tested or have not produced sufficient change to warrant being rolled out at scale.

Research into the determinants of hand washing behaviour has identified disgust and shame as key motivators. Evidence supports the theory that disgust is a natural behavioural reaction to objects carrying disease risk, thus it may act as a key motivator for other health related behaviours such as water treatment. Whether this knowledge can be harnessed to increase the efficacy of hand washing and safe water campaigns in Bangladesh or elsewhere has yet to be rigorously tested.

The investigators will develop an intervention that utilizes disgust and shame eliciting messages to promote hand washing with soap and point of use water treatment in low income housing compounds of urban Dhaka. The investigators will test the efficacy of this intervention against a more traditional public health intervention based on increasing knowledge of health risks and germ transmission using a randomized controlled trial. The study sample will be broken into the following four arms.

1. Standard Public Health Intervention with Water Treatment
2. Standard Public Health Intervention with Water Treatment & Hand Washing
3. Disgust and Shame Based Intervention with Water Treatment
4. Disgust and Shame Based Intervention with Water Treatment & Hand

This design will allow us to compare outcomes for hand washing and water treatment between both standard public health interventions and disgust and shame based interventions as well as test the overall efficacy of the program comparing with the control. Data will be collected from all compounds at baseline, three month midline and at the six month endline giving us the practical and analytical benefits of a longitudinal dataset.

Compounds will participate in interactive, educational safe water and/or hygiene promotion meetings. For the Disgust and Shame group, these meetings will emphasize disgust and shame related to unsafe water and/or hygiene practices, whereas the Standard group's meetings will resemble a more typical public health intervention explaining the risks and methods of contamination.

At the first meeting, compounds will receive a one month free trial of the latest compound based chlorine dispenser model to treat their drinking water. A randomly selected half will also receive a one month free trial of the latest compound based handwashing station. At the end of the month, there will be a sales meeting in which the investigators will measure compound members' willingness to pay for the trialled products by giving them the opportunity purchase and keep the hardware in a Becker-DeGroot-Marschek (BDM) style auction.

In assessing the impact of the interventions, the investigators are primarily interested in whether the prevalences of safe water and hygiene behaviours differ by treatment arm and over time. The best measurements for approximating behaviour prevalence are physical observations (presence of residual chlorine, hand cleanliness inspections), structured observation of behaviour, rapid physical observations (physical state of hardware/drinking water), self-report of water treatment and hand washing behaviour and willingness to pay for necessary products. The investigators will also attempt to measure and track changes in personal determinants of behaviour such as feelings of disgust and shame related to hand washing and water treatment behaviours.

Study Design To rigorously test the main study hypotheses the investigators are using a stratified, cluster randomized trial. Sample compounds will be divided into four strata based on compound size and presence of gas and will then be randomly assigned to one of four study arms . Thus the proportion of compounds in each arm will be the same within each stratum.

Strata

1. Gas & Fewer than 8 households per compound
2. Gas & 8 households per compound or more
3. No Gas & Fewer than 8 households per compound
4. No Gas & 8 households per compound or more

Study Arms

1. Standard Public Health Intervention with Water Treatment

2. Standard Public Health Intervention with Water Treatment & Hand Washing

3. Disgust and Shame Based Intervention with Water Treatment

4. Disgust and Shame Based Intervention with Water Treatment & Hand Washing

   Via randomization, half of compounds will be assigned to standard treatment groups one and two and half to disgust and shame groups three and four. Within these halves, one third will receive water treatment only and two thirds will receive water treatment and hand washing.

   Randomization will allow us to infer causality when analyzing the differences in outcomes and stratification will increase the statistical power of our inference.

   Field Staff The investigators will have three field staff designations, Field Research Assistants (FRAs), Field Intervention Specialists (FISs) and Field Assistants (FAs). FRAs will be responsible for baseline, midline and endline data collection as well as structured observation. FAs will be responsible for fee collection visits. FISs will be responsible for all remaining visits.

   Site Selection Based on the literature review and hypotheses to be tested, the investigators determined that our study would be most feasible and would provide the most benefit in poor housing compounds of Dhaka city. To identify specific sites where the intervention would take place, the investigators created a list of feasibility criteria (see appendix 2) and received recommendations from ICDDR,B Research Investigators with experience in urban Dhaka. The investigators then sent two FRAs to evaluate the recommended sites using the list. With their evaluation the investigators selected six communities: Mohammedpur, Mirpur, Badda, Korail, Khilgaon and Bashabo.

   Piloting Phase The first five of our specific aims, all involving development of new techniques, are the subject of our piloting phase.

   Aims One and Two To develop our disgust and shame intervention for water treatment and for water treatment and hand washing the investigators will refine a long list of potential intervention script contents. Our Qualitative Research Investigator and members of his/her team will field test and review all items from this list, systematically changing, removing and adding items until the list has become a feasible intervention script that is culturally appropriate and appears to be effective. The Qualitative team will use focus group discussions and in depth interviews as primary tools to assess field tests.

   Aim Three To develop a group version of the Becker-DeGroot-Marsckek (BDM) auction our staff Economist along with the Qualitative Research Investigator and a team of FRAs will field test and review different variations of the auction to determine which is most feasible and which appears to produce the most accurate measurement of true willingness to pay in our communities. The Qualitative team will use focus group discussions and in depth interviews as primary tools to assess field tests.

   Aim Four To develop a survey instrument which allows us to measure behavioural determinants of hand washing and water treatment, the investigators will perform multiple pilots and revisions of our questionnaire on small sample groups similar to our sample. Revisions will be made based on qualitative review of subjects' impressions and a quantitative review of gathered data. In the quantitative review the investigators will perform principal components analysis and factor analysis of data to refine scales for behavioural determinants already defined and/or to investigate whether there are other determinants worth measuring.

   Aim Five The investigators have a short list of new measurement ideas that the investigators will field test and review prior to implementation. For more details on this list, see the "Data Collection" section below.

   Implementation

   Visit 1: Sample Enrollment, Baseline Data Collection and Set-up of Promotion Meeting

   Within the chosen field sites, FRAs will search for compounds which, match all the essential criteria (see below), match at least 2 of the preferred criteria (see below) and which are located at least 75 footsteps apart from each other. Our field team will make several visits during the six months of the study to each of the enrolled compounds. During each of these visits the investigators will conduct different activities including a household survey, structured observation, and physical observations. The investigators will ask for consent from all the compound members verbally during a courtyard meeting, and ask the members to come to a consensus of who can sign as the compound head or manager. In the written consent form all the study activities have been described briefly to make it simple and easy to administer and understand. For each specific activity the investigators will again ask for an individual's verbal consent, which will serve as a reminder that individuals are still willing to take part in the activity. Taking written consent for each activity during each visit could make the visit even longer. Signing paper documents several times could make the participants confused and concerned because most of them cannot t read or write. For any household members who miss the first courtyard meeting, the investigators will repeat all the information and process.

   If they do not agree, then FRAs will move to the next suitable compound. If they agree and sign the form, then FRAs will record stratification data and perform baseline data collection which consists of two household questionnaires and six household physical observation forms (see "Data Collection" below for more details on questionnaire and physical observation). Later, FROs will use stratification data to determine the stratum and randomly assign the treatment arm for each compound.

   Household representatives who will participate in baseline data collection will be randomly selected from the pool of compound members present at visit 1. To ensure random selection, the FRA will first arbitrarily assign numbers to the households with members present. He/she will then write the numbers on small plastic balls and drop the balls into a cup. He/she will then shake the cup and pour out one ball. The number on the ball corresponds to the first household to approach for data collection. The FRA will ask the present members of the household who is responsible for collecting water. This person will be asked to participate, if he/she declines, the FRA will pour another ball out of the cup and repeat the procedure. These steps will be repeated until six consenting compounds have been selected. The first two will be given the questionnaire and physical observation form and the last four, only the physical observation.

   Essential compound criteria

   • Between 3 and 15 House'holds
   • Shared water source
   • Physical space exists to hold a compound meeting here or nearby
   • No other interventions going on at this time
   • Population is all Bangali

   Preferred criteria

   • Use of water source is visible to others
   • Shared kitchen
   • Shared toilets visible from common area Visit 2: First Meeting- Promotion and Free Trial FISs will conduct a three hour integrated behaviour change/product promotion program taking place in the compound or at a convenient meeting place nearby. The content of the program, which will vary by treatment arm, will be developed during piloting and is briefly described below under "Meeting Content and Hardware".

   At the end of the program, compounds will be offered a one month free trial of the hardware specific to their study arm (see "Meeting Content and Hardware").

   Visit 3: Reminder Visit Roughly two weeks after the first meeting compounds will receive a check-up and reminder visit. FISs will meet with available individuals in two smaller groups. The FIS will reiterate messages from the first meeting and discuss understanding and agreement with these messages in order to further understanding and reduce perceived barriers to the new behaviours.

   Visit 4: Setup of Second Meeting and BDM Auction Coaching Roughly three weeks after the first meeting, FISs will visit compounds to set up a time and date for the second compound meeting and to prepare the compound members for the auction. FISs will meet with groups of two to four prospective bidders for an hour each and will explain in detail the process and hold multiple mock auctions.

   Visit 5: Second Meeting- Promotion and Sales Roughly one month after the first meeting, as the free trial is ending, FISs will return to treatment compounds to conduct the promotion and sales meetings.

   The promotion segment will be a one and a half hour behaviour change/product promotion program similar to the first meeting but containing fresh material.

   In the second one and a half hour segment of the meeting, FISs will measure compound members' willingness to pay for the trialled products by giving them the opportunity purchase and keep the hardware in a Becker-DeGroot-Marschek (BDM) style auction. Compounds not willing to pay the offered price will be offered other options to keep the hardware assuring that at least 75% of compounds keep the products until endline (more details on this in Auction section below).

   Fee Collection Visits Every month for a year after the sales meeting, compounds will be visited by an FA who will collect fees and take rapid observation data (detailed under "Data Collection" below).

   Visit 6: Midline Data Collection Three months after baseline data collection, FRAs will return to compounds to collect midline data which consists of two household questionnaires and six household physical observation forms (see "Data Collection" below for more details on questionnaire and physical observation). FRAs will attempt to use the same households as before, but if they are not present after returning to the compound for a second attempt, then FRAs will use the randomization strategy from the first visit to select new households.

   Visit 7: Structured Observation Half way between midline and endline data collection a separate team of FRAs will perform 5 hour structured observations of hand washing and water treatment behaviour in a representative subgroup of our sample. More on this under "Data Collection" below.

   Visit 8: Endline Data Collection Six months after baseline data collection, FRAs will return to compounds to collect endline data which consists of two household questionnaires and six household physical observation forms (see "Data Collection" below for more details on questionnaire and physical observation). FRAs will attempt to use the same households as before, but if they are not present after returning to the compound for a second attempt, then FRAs will use the randomization strategy from the first visit to select new households.

   Meeting Content and Hardware:

   Hardware Water Treatment One wall mounted liquid chlorine dispenser One 40 liter reservoir with tap Hand Washing One reused 1.5 liter water bottle to be filled with soapy water (laundry detergent mixed with water)

   Meeting Content

   All compounds:

   All will receive water treatment promotion interventions, whether standard or disgust and shame based. Two thirds will be randomly selected to receive hand washing promotion along with water treatment.

   Standard Public Health Intervention:

   The standard public health intervention meetings will be modeled after typical pre-existing high quality water treatment and/or hand washing interventions. The content will include explaining/demonstrating how germs can enter into our body via untreated water and/or unwashed hands, how they can make us sick and lead to death, and how these risks can be reduced by practicing safe water and/or hand hygiene behaviour.

   Disgust and Shame Based Intervention:

   The disgust and shame meetings will contain similar explanations/demonstrations of contamination mechanisms and risk, but will place most emphasis on the presence of fecal matter in or on contaminated objects. FIS' will use more vivid or harsh local terms equivalent to "crap" or "shit" to try to elicit a stronger reaction from participants. They will try to communicate the information that we sometimes unknowingly serve feces to our family by not washing hands with soap or by not treating our drinking water. They will emphasize how fecal matter can spread between people, especially neighbours, to try to encourage people to care about others behaviours and what others think of their behaviours. Pending piloting, the meetings will also contain promotion of positive identities-being a good mother, a strong father or a good Muslim-to stand in contrast with shameful and disgusting activities. The content and terms used will be determined by qualitative investigation with community members.

   Data Collection Questionnaire All compounds will be surveyed at baseline, a three month midline and a six month endline. The questionnaire for each of these instances will be largely the same excepting corrections made during implementation and small changes for administering at different times. The questionnaire will measure self-reported usage, reports of neighbours' usage, knowledge and practice pertaining to hand washing and safe water, perceptions of risk and severity of diarrhoea, norms/ beliefs/ feelings/ reactions relating disgust and shame, social networking, and demographic information. These measurements will allow us to study the differences in behaviour and of behavioral determinants over time and between groups.

   Physical Observation One of each of the following is taken in 6 households per compound each time the questionnaire is administered and also during the structured observation.

   Chlorine residual testing of stored drinking water (not taken at baseline) Hand cleanliness inspections of child (see scoring sheet in appendix 4) Hand washing demonstration by mother The investigators will ask her to demonstrate how she washes her hands after defecation. The investigators will evaluate her performance as outlined in appendix 4.

   Others the investigators may include pending piloting:

   Hand rinse water testing via H2S testing (see appendix 13 for previous principle and interpretation) Other finger cleanliness checks Wiping fingers on a white material and evaluating darkness Wiping fingers on oil blotter paper Drinking water quality testing via H2S testing

   Rapid Physical Observation Rapid observations will take place in all compound visits and consist of identifying the location and usability/usage status of hand washing and water treatment hardware.

   Structured observation There will be one structured observation visit halfway between the midline and endline data collection visit. FRAs will conduct 5 hour structured observations of hand washing and water treatment behaviours in a sub-group of compounds which is representative of our sample. Hand hygiene and water treatment behaviours will be assessed using the score sheet in appendix 6.

   Auction/Willingness to Pay After the courtyard meeting, the investigators will conduct an experimental auction to elicit individual households willingness to pay (WTP)for a monthly rent-to-own subscription to the chlorine dispenser and the reservoir. The investigators will use Becker-DeGroot-Marschek (BDM) procedure to elicit WTP, whereby participants bid against a predetermined randomly assigned price confidentially kept in an envelope. The investigators will collect individual bids and list them to find the lowest of these bids, which will then be compared with the price in the envelope. In this auction, if the lowest of participants bid exceeds the preassigned price, participants as a group win the auction but each pay the price in the envelope. The investigators assume the mechanism is incentive compatible for participants to bid truthfully as their bid does not affect the transaction price. The investigators aim to keep the actual sales price very low so that most compounds have the opportunity to keep the dispenser.

   Payment data After the sales meeting the investigators will measure continued enrollment and maintenance of the safe water (or safe water and hand-washing) station on a monthly basis.

   Qualitative Data Collection:

   During implementation our qualitative team will regularly visit a representative sub-group of our sample to gather data on how the intervention is going, how it is perceived and to give us an idea of what needs to be changed in our meetings or our survey. The standards the investigators are using to determine the frequency of qualitative work are as follows: Roughly one in every five compounds will participate in in-depth interviews and one in every 18 compounds will take place in focused group discussions. These numbers may vary depending on data redundancy.

Inclusion Criteria:

Essential compound criteria

• Between 3 and 15 Households
• Shared water source
• Physical space exists to hold a compound meeting here or nearby
• Population is all Bangali

Exclusion Criteria:

• No other interventions going on at this time