Design process
The intervention design process combined the ABS expertise of the London School of Hygiene and Tropical Medicine (LSHTM) researchers, the local knowledge and experience of the team from the Centers for Infectious Disease Research in Zambia (CIDRZ), and the creativity of the local creative agency. The agency provided a series of five proposed versions of the campaign concept, two of printed materials, and two of live action videos for feedback; pre-tested campaign concepts and content with the target audience; and conducted a full pilot of the intervention with a single group of landlords. LSHTM and CIDRZ provided feedback on materials, pre-tests, and piloting based on research design considerations, formative research findings, and pre-test/piloting participant responses. They also streamlined the campaign manuals and lesson content based on the surprise/revaluation/performance paradigm and created emo- and exo-demos for behaviors that were lacking them.
The creative agency initially suggested several intervention ideas that had to be discarded as either logistically infeasible in the time allotted (creating a toilet evaluation system featured in a publication designed for tenants to browse homes for rent) or compromising the individually randomized evaluation plan (a “talking toilet” installed in the central market that drew attention to its own poor quality). Additional ideas, such as financial literacy training for landlords, were discarded when pre-testing revealed that most landlords viewed their plots as businesses, but this discovery led to a profit and conflict-reduction focused campaign focus.
Some concern was also raised among the team about profit-motive-related interventions and their potential to displace tenants through higher rental fees, resulting only in more income for landlords, or in the potential for increased marginalization caused by negative messaging [
61]. We hypothesized that simply constructing high-quality toilets would inevitably lead to the displacement of many tenants. However, we established that many landlords did not think that tenants would pay for differences in sanitation quality [
59] and that tenants were willing to pay [
60], and thus focused on increasing the functioning of the market. There were few plots with low-quality housing and high-quality toilets, denying tenants the opportunity to access a plot with a higher-quality toilet, and while some tenants might not choose to or be able to pay more, the negative impact of allowing a larger range of choices was thought to be much smaller in this case. We also determined not to use any negatively-framed messages, instead focusing on the benefits that had been underestimated by landlords previously.
There was some tension between the expectations and processes of the creative agency and the researchers. These groups had very different perspectives on methodological rigor and expectations for the degree to which a campaign could change based on information from field testing. In addition, despite their “local” status, the intervention targeted areas of town and socio-economic classes a bit distant from some of the creative agency staff, and ensuring an adequate knowledge of and interaction with the target population was challenging. Though we prefer to involve local creative agencies, the BCD framework provided sufficient guidance to allow the researchers to generate additional intervention components and evaluate these components from both theoretical and practical perspectives, meaning that some components of the final intervention were produced by the creative agency, while others were generated by academic researchers.
The overall process from framing workshop to intervention delivery took 8 months, including 4 months to develop campaign concepts and materials and 4 months to complete video production. Several reasons for delays from the original 5-month timeline could be eliminated by other teams using this process to easily cut that timeline in half—we experienced slow administrative processes at both institutions and local government levels, procurement delays, and included an extended period for generating concepts based on the formative research data due to the lack of previously available information or interventions. Delivery to 20 groups of up to 25 landlords, with four meetings occurring over 2 months, was done by four pairs of presenters (one community health worker and one actor each), with tablets to show videos and some reusable printed materials. Four research assistants also worked as monitors. Materials provided to participants were only small printed cards and a durable plastic rota symbol, and rooms in local venues (churches and schools) were rented for delivery. Further details about study costs are available in the Additional file
1.
Intervention logic
The main intervention delivery mechanism was the creation of a “secret society,” which selected landlords would be invited to join so they could receive “insider knowledge” of how to build wealth and reduce conflict by improving sanitation on their plots. Meetings took place at a location near where they lived (either a school hall or church) rented for a small price by the project. Campaign manuals provided guidance for how to present “secret” information to landlords collected from tenants, some that allowed landlords to experience the emotions of their tenants, and some that gave practical tips. These were sorted by the researchers into the overall structure of surprise (videos), revaluation (games and demonstrations), and performance (practical guidance) for each meeting. These meetings were led by paid facilitators (often actors) trained in the intervention content as well as activation styles. High status was associated with meeting attendance and behavioral performance, as invitations were made using high-quality, branded materials and name badges displayed stars during the meetings to indicate the degree to which landlords made the targeted improvements.
The four target improvements identified were:
1.
Regular cleaning of the toilet interface to reduce direct user exposure to pathogens
2.
Installation of a lock on the inside of the door to increase safety and privacy
3.
Installation of a lock on the outside of the door to allow access to plot residents while excluding outsiders
4.
Installation of a water-sealed pan or cover to reduce smell and the spread of pathogens through vector contact with fecal material
The promoted locks, water sealed pans, and covers were already common in the local market. Regular latrine cleaning was encouraged by replacing a daily, verbal cleaning rotation with each household cleaning for a week, with a plastic decal hung above the door of the responsible household to bring accountability.
The opportunity for social reward, learning, and reinforcement was identified as a key behavior change mechanism, corresponding to surprise and revaluation in the BCD theory of change. Formative research indicated that landlords did not interact socially much with their tenants or even nearby landlords. In piloting, landlords praised even the opportunity to talk in an undirected manner about common challenges they faced as a helpful activity that rarely occurred otherwise. Hence, the intervention created a novel “social environment,” integrating aspects of learning by observing others from social learning theory [
32]. In these meetings, landlords could learn from the successes of others, ask for advice from others in dealing with barriers faced, and in cases where few successes were reported in one group, stories from other group meetings could be used to provide additional insights. Landlords worked together to solve problems and to help each other to make improvements (the Affiliate motive), but were also given name badges with the stars indicating the quality of their toilet to bring a sense of hierarchy (Status) [
62].
Another purpose of the four special-purpose group meetings was to facilitate behavioral performance through encouragement and monitoring. Paid monitors, distinct from the facilitators, conducted home visits to observe if improvements were made and to troubleshoot barriers faced. This information was given to the facilitators who used it for discussion at the start of the subsequent meeting. In addition, cards describing the main improvement were distributed to participants at the end of each meeting, which they were supposed to get a tenant to sign, indicating that they have taken relevant action after each meeting. These cards served as a tangible indicator of behavioral performance that could be monitored in the group setting, and visits by monitors to plots provided additional verification.
These cards also demonstrated the final purpose of the intervention structure—encouraging increased interaction between landlords and tenants to reveal unexpressed demand (again, surprise and revaluation). Each card represented a particular improvement that was made, and the required signature by a tenant designed to lead to an increasing number of discussions about working together in additional ways to improve sanitation on the plot. In particular, the card related to regular cleaning of the toilet (available at the project website [
63]) required a signature verifying that a meeting had taken place between the landlord and his or her tenants for the explicit purpose of discussing a system for toilet cleaning.
Specific messages and activities were developed for each target behavior within the overall surprise, revaluation, and performance framework of each meeting’s content as well (Table
1). For “Surprise,” we chose to create live action, “hidden camera”-style videos, surprising participants with both edgy content that they may rarely observe (such as a man failing to aim properly while using a toilet due to his concern about holding a door closed) and information that is not generally communicated to them as landlords (such as tenants admitting that a poor toilet has scared them away from renting a room).
Table 1
Key Messages and Segment Content for Each Landlord Meeting
Surprise | Key Message | An improved rota keeps the toilet clean and makes your tenants happy. | Without an inside lock on your toilet, your tenants are robbed of their privacy. | A toilet without an outside lock will be abused by others and anger your tenants. | A smelly toilet full of flies will scare away paying tenants. |
Video Description | Tenants gossip about who doesn’t clean the toilet, and this boils over into full-blown conflict and blaming the landlord for not handling the problem. | Tenants struggle to keep the toilet door closed, culminating with a man walking in on a woman using the toilet. An argument ensues, and both end up blaming the landlord for the lack of a lock. | Drunk men stumble in to use the toilet at night, but when the landlord finds it dirty in the morning and yells at a tenant, she turns it back on him for not securing the toilet from outsiders. | A series of potential tenants come to look at a room for rent, ask to see the toilet, and then abruptly leave, confusing the landlord about what the problem was. The tenants privately discuss that they will go rent a more expensive place with a better toilet. |
Revaluation | Key Message | Your toilet stays clean when the rota is simple and visible | A lack of privacy will drive good tenants away. | Asking tenants to do disgusting things will drive good tenants away. | A toilet is a wise investment that brings you more money quickly. |
Activity Description | Two teams were chosen with a landlord and 3 tenants each, and the tenants are assigned numbers—one team in blocks (i.e., 1–10) that are visible, and another in a more complicated, unwritten manner (i.e., every 3rd number). Landlords take turns identifying the tenant with a given number. | The facilitator asks for a chosen landlord to open their handbag and reveal every detail of the items inside and emphasizes the discomfort this lack of privacy causes. | Several participants were asked to come up one at a time to hold a tissue while the facilitator pretends to blow their nose loudly and messily. The facilitator translates this into the disgust tenants feel in having to clean up after outsiders who are messy and aren’t responsible to clean. | Two participants are assigned to invest either in improving the toilet or building a new room to rent. The toilet generates income sooner, rental gains are multiplied by the number of tenants, and a scenario where income is reduced shows that this is a more reliable and way to generate wealth. |
Performance | Key Message | Give your tenants the power to remind one another of their responsibilities. | It is easy to install an inside lock by yourself or with your lock-buddy. | Remember to call on your ‘landlord lock-buddy’ to help you install an outside lock. | Invest in a decent cover pan (or a pour-flush toilet) to keep your plot full of tenants and build your wealth. |
Activity Description | Landlords are given a badge to hang outside the door of the tenant responsible for cleaning the toilet that week and asked to have a meeting with all tenants to institute the new system. | A handy man demonstrates installing a lock and then has landlords take turns practicing. “Lock buddies” are paired up to purchase and install locks. | Same as performance for inside lock. | A handyman describes how to build a simple cover and the process and cost of installing various flushing toilet options. Merry-go-rounds suggested to spread out the cost over time and encourage accountability to each other. |
For revaluation, the SanDem intervention used “emo-demos,” or emotional demonstrations designed to revalue behaviors through emotional responses [
64]. It also used “exo-demos,” an extension developed for this intervention, of “executive-level [i.e., cognitive] demonstrations” aimed at revaluing behaviors through activities requiring conscious group deliberation in areas such as calculating potential profits from toilet investments [
65]. The overall theme of the revaluation sections was that a poor quality toilet leads a landlord to lose good tenants and give up a steadier, higher monthly rental income. Specific revealed desires of the tenants (e.g., privacy, cleanliness) were always translated directly into motivations for landlords (e.g., reduced plot conflict, more rental income). Exact details of the intervention can be found in the facilitator guide [
63].
Behavior-specific performance facilitation varied by the kind of behavior. For cheaper, one-time actions (installing outside and inside locks), a buddy system was used where pairs of landlords helped each other purchase and install the locks at a set time following the meeting. For cheaper, ongoing actions (initiating an improved cleaning rota), an initial meeting with tenants was reported, and use of a visible, durable symbol of the cleaning system was verified during monitoring visits described below. For more expensive, one-time actions (improving the seal of the toilet or building a door, perhaps necessary prior to lock installation), a handyman already working in the community provided information about products available and the range of installation costs based on existing infrastructure. “Merry go rounds,” a common local mechanism where each participant contributed money each round and one participant received the contributions (rotating each round), were also suggested to landlords to ease the amount of one-time savings required and to provide peer accountability for making pledged improvements.