Using Behaviour Change Messaging to Improve Communal Toilets in Dhaka, Bangladesh

Hypothesis to be tested: A behavior change communication intervention can improve the quality and cleanliness of sanitation services available to low-income residents of Dhaka.

Specific Objectives:

1. To deliver communication messages (selected as the most promising for changing target behaviors influencing the quality and cleanliness of shared sanitation facilities) to randomly selected low-income communities in Dhaka.
2. To evaluate the effectiveness of a behavior change communication intervention designed to change targeted behaviors related to the quality and cleanliness of shared sanitation facilities.

Background The investigators contend that important barriers to improved sanitation in low-income communities in Dhaka include: 1) Low-income community residents lack the financial resources to purchase a better option. 2) The cost of land in Dhaka, which has increased 600 fold since the 1970s (Hossain, 2011), makes space for toilets expensive 3) A public sector characterized by weak institutions that lack funding and management capacity to implement and maintain effective sanitary infrastructure (Hossain, 2011) and 4) a pervasive informal local political system that maximizes the financial interests of politically connected elites over the communal interests of low-income community residents (Hackenbroch and Hossain, 2012, Hossain, 2011). More than one third of Dhaka residents live in informal settlements where government authorities do not accept responsibility for providing urban utilities including water and sanitation services (World Bank, 2007). A central motivation for politicians to avoid formalizing settlements is that informality keeps residents in a constant state of dependency on local party associates, a state which these local elites can use to extract a continuous stream of resources (Anjaria, 2006, Hossain, 2011).

Behavior change communication messages cannot directly resolve these major barriers to improving community sanitation in low-income communities in Dhaka. Simple general messages, such as encouraging people to clean keep their sanitary facility clean to reduce the risk of disease transmission, are unlikely to be effective because they do not address the determinants of the poor availability, cleanliness and maintenance of shared sanitary facilities. By contrast, by developing behavior change messages that are informed by the experience and perspective of the target population this study can identify some opportunities where new information or a new social norm could alter some key behaviors.

The pilot work found considerable interest in cleaner better maintained latrines. Inclusion of 2 low-cost hardware accessories, a waste bin so that materials such as condoms or sanitary pads can be discarded in ways that do not block the toilet and bucket to assist with rinsing the toilet after defecation makes it easier for communities to follow behavior change recommendations to support a cleaner shared latrine.The central question for this research project is: Can a strategically designed behavior change campaign supplemented with low-cost hardware accessories shift the current equilibrium of unhygienic shared sanitation facilities towards cleaner facilities in low-income communities of Dhaka?

Research Design and Methods:

SUMMARY:

• Deliver behavior change communication messages using a behavior change communication intervention to randomly selected low-income communities in Dhaka. (Package of behavior change interventions for targeted behaviors: waste disposal, rules for latrine use and repair and maintenance)
• Evaluate the effectiveness of the behavior change communication intervention in changing the targeted behaviors.

DETAILS:

• Deliver messages using a behavior change communication intervention to randomly selected low-income communities in Dhaka.

The study team will add intervention components onto existing NGO interventions to increase impact and decrease cost. Thus, through extensive piloting informative work, the investigators have developed behavior change communication materials, based on specific behaviors targeted to improve latrine cleanliness and quality. The materials and messages will be used to train community health promoters (CHPs) that are currently working for the selected NGOs. Some materials will include sign boards for latrines that will be fixed to those areas randomly selected to receive the intervention. The intervention will also include enabling hardware, a waste bin for trash collection and the bucket for carrying water to flush the latrine. The intervention: these will be delivered by partner NGO community hygiene promoters. This intervention approach is aimed to leverage existing intervention delivery partners and thus add value across the sector. The investigators have identified an NGO partner with large urban programs including sanitation activities in Mirpur, Mohakhali and Mohammadpur.

The survey team will work with partner NGOs and identify the coverage areas that are eligible for participation in the study. These include:

1. Some larger slums with 60 to 180 communal toilets and sanitation interventions provided by 3 to 12 community health promoters whose activities collectively cover the entire slum area.
2. Some smaller slums with 6-10 communal toilets available for around 150 to 300 households. These communal toilets were situated very close to one another, and this type of slum is served by a single community health promoter.
3. Some scattered slums which are very small compared to other slums. This type of slum consists of 15-50 households sharing one communal toilet. One community health promoter covers 4-8 scattered slums.

The team will divide those selected slums into distinct clusters primarily based on community health promoter coverage area. If the area is very large, the CHP coverage areas will be sub-divided using geographic boundaries such as roads, bazaars, water bodies that clearly distinguish between distinct clusters. Each cluster will comprise:

1. an area covered by a single CHP
2. between 1 and 30 communal toilets

A communal toilet consists of a structure with approximately 2 to 16 chambers that drain into a common septic tank. The clusters will be the unit of randomization for this study.

At the beginning of the study, the study team will conduct a communal toilet listing to collect information on numbers of toilet users for each communal toilet. Once listing is done, a statistician who is independent of this project will randomly assign clusters to either the intervention or comparison arms. Clusters are separated geographically by streets, bazaars, fields, drainage or other settlements, and therefore investigators do not expect intervention messages or activities to spill over into control toilet clusters even when they are located very close by.

For each intervention cluster, the field will also enroll a matched control cluster, matched on the number of toilet chambers per cluster. Investigators will use the following categories for matching: <5 chambers, 5-10 chambers, 11-15 chambers, 16-20 chambers, and >20 chambers assuming that based on the number of chambers, the numbers of users could vary. A statistician, who is not a resident of Bangladesh and who is not a member of the study project, will randomly assign clusters to intervention versus control. WSUP, on behalf of the intervention team, will sign contracts with the partner NGOs to implement the behavior change package. The behavior change messages and enabling hardware will be delivered by the partner NGOs through their community health promoters (CHPs). Investigators will help partner NGOs train CHPs in the components of the intervention to ensure that CHPs conduct group meetings and inter-personal communication that will effectively convey our intervention messages.

The partner NGO will receive final behavior change communication materials from the study team and it will be responsible for printing and distribute these materials only to intervention communities. The research team will be involved with the implementation team to ensure that rollout occurs in the specified intervention clusters at the planned frequency and timing and that there is no messaging in the control clusters. Investigators will explain the plan for the fidelity assessment with the implementers and share the results as soon as they are available.

• Evaluate the effectiveness of the behavior change communication intervention in changing the targeted behaviors.

Behavior change campaigns that provoke discussion between viewers of the campaign and their social network are more likely to result in behavior change (Boulay et al., 2002, Hafstad and Aaro, 1997). The survey instrument includes objective uptake indicators, a measurement of exposure to the intervention, awareness of the behavior change package, frequency of discussion of the contents of the behavior change communication information with the respondent's social network, access to sanitation facilities and attitudes regarding the acceptability and cleanliness of these facilities. Investigators will conduct a pre-intervention baseline survey by drawing a systematic random sample of one adult user per communal toilet in both intervention and control clusters. The field team will sample equal numbers of users from intervention and control arms. The field team will visit the communal toilet and enroll the third household nearest to that communal toilet and administer the toilet user survey to one adult resident. If the selected user doesn't want to participate in the survey, the field team will approach a subsequent household for enrollment and also record the non-response. The field team will administer a toilet quality/cleanliness observation instrument to assess the primary objective indicators for all study toilets. To ensure inter-rater reliability of toilet observations among observers, a subset of communal toilet will be evaluated by two separate observers. Observer pairs will conduct their assessments during a 5-10 minutes interval, and their independent evaluations will be compared to make sure that our observers measure cleanliness consistently and objectively. Such research reliability observations will be conducted for initial 50 communal toilets, or until observer credibility can be effectively established and discrepancies minimized.

Along with quantitative survey, field investigators will also explore local elite or landlord's view on their role and roles of others in constructing, maintaining, cleaning and emptying the communal toilet facilities by conducting qualitative interviews for the previously mentioned three types of slums. Field investigators will conduct in-depth interview with 4 different landlord or elite from mentioned each types of slums or they will interview until they have reached data saturation. Field investigators will interview those elites or landlords before the intervention commences and at the end of the intervention.

The study team will conduct a baseline evaluation in intervention and control communities as well as an endline after the 6 months of intervention roll out. In addition, two months after initiation of the intervention, the field team will conduct a rapid fidelity assessment using the same survey instrument in a random subset of 50 intervention toilets. The purpose of this rapid assessment is to ensure that the behavior change intervention exposure is occurring at the expected level. If not, then the partner NGOs will be contacted and the problems addressed.

The study team will compare the change in targeted behaviors observed in the intervention communities compared with changes observed in the control communities and so measure the amount of behavior change attributable to the intervention.