Study design summary

The SABABU (Sakafu Bora Afya Bora - Utafiti; Kiswahili for Better Floors, Better Health - Research) trial is a two-arm parallel, cluster-randomised trial to measure the effect of the intervention 12 months post delivery. We consider households as clusters and, in each site, 220 households meeting the inclusion criteria are randomised (1:1) to either control or intervention arms. The intervention involves replacement of earthen floors with an improved cement-stabilised floor, accompanied by a behaviour change component (floor-care guide). In both study arms, all household residents over 1 year are offered a baseline treatment for STH (200 mg albendazole 12–23 months and 400 mg albendazole ≥24 months).

In each household, prior to the STH treatment, stool samples are collected from all residents immediately before and 12 months post installation of floors (figure 1). Stool samples are assessed via PCR for enteric pathogens (children <5 years) and via Kato-Katz for STH (individuals >1 year), and clinical examinations are performed for tungiasis (children 1–14 years). During tungiasis assessments, those found to be affected are treated with benzyl benzoate (BB), provided by the study, at the point at which they are found, and the BB lotion is provided to the relevant community health volunteer to follow-up with two further treatments.15 29

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Study design. STH, soil-transmitted helminth.

In both study arms, quality of life measures in enrolled children and their caregivers are recorded immediately before and 12 months post installation of floors. Soil/dust sampling from all enrolled house floors is performed 6 months and 12 months post installation to screen for off-host Tunga penetrans stages. Environmental sampling for enteric pathogens is conducted on floors of a subsample of enrolled households at 12 months post installation. Alongside the trial, a process evaluation is undertaken to investigate intervention acceptability, durability, practicality and cost.

This protocol is reported in accordance with the Standard Protocol Items: Recommendations for Intervention Trials (SPIRIT) checklist (online supplemental file 1).

Study settings

To improve generalisability and examine impact heterogeneity across environmental and cultural contexts, two sites are selected for the trial. Initially, formative work was carried out in 2021 in Kwale, Bungoma and Narok Counties to: identify housing conditions, map daily routines, and explore home improvement priorities to inform intervention development.23 Working with County Departments of Health, seven nearby villages per county were selected based on epidemiological profiles (reported STH and tungiasis endemicity) and housing conditions (high proportion of households with earthen floors). In Kwale and Bungoma, communities demonstrated a strong demand for sealable, washable floors, whereas in Narok this was less explicit with stakeholders citing other priorities. Moreover, a large majority of the traditional dwellings in Narok were deemed unsuitable for retrofitting of new floors. Subsequently, only Kwale County on the south Kenyan coast and Bungoma County in western Kenya (figure 2) are included in the trial.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2

Map of study settings.

Within Kwale County, the study villages are located in Dzombo ward, Lunga Lunga Sub-County. STH prevalence (predominantly hookworm) was 20% in 2017,30 reported diarrhoea prevalence in children under 5 was 3.2% in 202231 32 and tungiasis prevalence has been found to be greater than 50% in some villages.33 Formative work in the seven villages censused 5241 individuals living across 812 households.23 76% of house floors were earthen and 13% had unimproved (without slab) latrines. Within Bungoma County, the study villages span South Bukusu and Kabula wards in Bumula Sub-County. Prevalence of any STH in school-aged children is 7.3%,34 reported diarrhoea prevalence in children under 5 was 18% in 202231 32 and tungiasis has been reported as present. Formative work among the seven study villages censused 4560 individuals living across 906 households, the majority of which (75%) had earthen floors and 16% of which had unimproved latrines.23

Intervention

Our theory of change is described in detail elsewhere,28 with intervention design guided by three principles: ownership; community norms and standards; and awareness of existing routines. Briefly, installation and ongoing maintenance of improved household flooring are expected to reduce transmission of enteric and parasitic infections, by preventing direct exposure to pathogen and parasite contamination and through an intermediate effect of improved domestic hygiene. This is predicated on the assumptions that household members spend their time on the improved floor; it is a sealed surface limiting the opportunity for STH and sand flea off-host stages; it is washable and is regularly cleaned to reduce pathogen build-up and has limited animal contact to minimise faecal contamination. Improved household flooring is also intended to improve caregiver well-being through a number of routes including reduced health anxiety,25 improved comfort35 and increased levels of satisfaction, pride and self-efficacy.36 37 Child well-being could be impacted through increased energy levels resulting from decreased exposure to enteric pathogens, and reduced pain and itching as well as improved mobility, sleep patterns and concentration resulting from reduced exposure to skin parasites such as Tunga penetrans.38–40 The intervention consists of two components: (1) provision of a low-cost, sealed, washable floor throughout all eligible buildings in the dwelling and (2) a simple behaviour change component to support keeping these floors clean and well-maintained.

A low-cost, cement-stabilised earthen floor is installed in each eligible building of the dwelling (excluding unimproved pit latrines, dedicated animal sheds and stores) to meet the following requirements: (1) non-absorbent, durable and smooth; (2) good wear resistance; (3) acceptable appearance; and (4) affordable. The decisions on flooring material and construction were informed by and built on a recent feasibility study exploring locally appropriate and affordable flooring solutions to create a sealed, washable floor for the prevention of tungiasis in coastal Kenya.24 The existing soil is broken up, levelled and compacted; a damp-proof plastic membrane is added to prevent dampness rising up through the stabilised floor; a cement-stabilised murram mix (1 part cement:9 parts murram) is added and compacted to a depth of 100 mm over the membrane; and finally, a screed mix (1 part cement:4 parts sand) is prepared and applied (25 mm), with a smooth cement and water slurry smeared on the screed to give a smooth finish (figure 3). During curing, water is sprinkled over the floor surface twice a day for 7 days.

• Download figure
• Open in new tab
• Download powerpoint

Figure 3

Technical drawing of the improved (cement-stabilised) floor installation.

All materials are provided by the study. Installation is performed by locally recruited masons, trained by the study team, with the support of local labourers. The study team provides day-to-day supervision and quality checks using a standardised checklist for all flooring activities, and the process is overseen by the same supervisors across sites. Additionally, to ensure consistency across sites, the masons from Bungoma travel to Kwale to observe the Kwale masons at work, and the Kwale masons travel to Bungoma to partake in the training and offer advice based on lessons learnt. Household members are not expected to contribute to labour or costs of laying floors but are required to vacate their dwellings for up to 9 days while floors are laid and cured. They are also requested to assist with water provision for the floor installation and conduct the curing process and are encouraged to contribute labour during excavation where practical.

The behaviour change component, developed in consultation with the community, comprises a floor-care guide and household-level floor-care meetings. The guide is a double-sided, laminated visual aid demonstrating the ‘dos’ and ‘don’ts’ of floor maintenance including regularly sweeping and washing the floor, avoiding fire directly on the floor, and keeping animals off, where possible (online supplemental file 2). Following floor installation (and before households move back into their dwelling), trained field officers visit to address any questions and to provide advice on floor-care.

Trial outcomes

Primary outcomes are (1) enteric infections among children under 5; (2) tungiasis infection among children 1–14 years old; (3) and STH infection among residents of the same households who are at least 12 months of age; all of which are measured in both study arms at 0 months and 12 months (figure 1 and online supplemental file 2). A binary indicator of tungiasis infection is based on clinical examination and a binary indicator of any STH infection is assessed via Kato-Katz. For enteric infections, PCR of stool samples is used to identify prespecified enteric pathogens selected on the basis of a review of the literature:5 9 41–43 including Giardia lamblia, Cryptosporidium spp., Campylobacter jejuni, enteroaggregative E. coli (EAEC), Shigella and adenovirus, with the primary outcome based on infection with any pathogen versus none. Secondary outcomes include the following: caregiver-reported gastrointestinal illness in the last 7 days in children under 5 years; intensity of tungiasis and severity of tungiasis-associated pathology in children 1–14 years; well-being and quality of life in caregivers and children aged 8–14 years; enteric pathogen-specific prevalence among children under 5 years; and STH infection prevalence and intensity by species in individuals over 1 year.

To quantify reduction in contamination, secondary outcomes measured in both arms through environmental surveys of soil/dust samples include the following: contamination of floors with the same prespecified enteric pathogens and STH eggs/larvae (measured by PCR) and with eggs, larvae, pupae and adults of T. penetrans (measured using entomological screening).

Well-being in caregivers and children aged 8–14 years is measured using a range of standardised well-being and quality of life tools.

The differential effect of the intervention on the three primary outcomes (prevalence of enteric infections, STH and tungiasis) by context is examined via subgroup analysis stratified by study site and household contextual factors.

Process evaluation

Formative research indicated potential for heterogeneity in how household members interact with the floor and adapt behaviours such as cooking, animal husbandry and sleeping, all of which may play an important role in mitigating intervention success.23 A process evaluation therefore explores implementation practicality, acceptability, durability and cost.

The practicality of retrofitting a floor and the consistency of delivery across sites are assessed using internal project reporting on the following: procurement, training, flooring installation, and implementation timelines. Installation quality checklists, observations and interviews with masons, study teams and households are also used. Consistency of the delivery of the behaviour change component is measured using internal project reporting, spot-checks and interviews with intervention households. The durability of the improved floor is observed in relation to environmental, installation and use factors at three structured spot-check visits, during which performance characteristics such as crazing, cracking and blistering are assessed by engineers.

Intervention acceptability is assessed through use of space within the dwelling measured preintervention and postintervention using self-reported measures, and the practice of target behaviours measured postintervention using spot-checks and direct semistructured observations. Caregiver satisfaction is measured using semistructured in-depth interviews and structured Likert scale questionnaires administered at endline to intervention households. Costs are measured using semistructured interviews and internal project records to assemble a cost breakdown for delivering each of the intervention components.

Sample size calculation

Sample size and power calculations are based on primary outcomes and have been informed by data from Kenyan populations, including from the national school-based deworming programme,34 community-based tungiasis surveys44 and the Global Enteric Multicenter Study.45 Calculations are based on the principles of cluster-randomised trials, assuming an ICC of 0.1 based on small cluster size (household). Effect sizes for enteric pathogens are based on previous WASH efficacy studies.46 47 Effect sizes for tungiasis and STH are based on the research team’s expert opinion of the smallest meaningful public health effect. Tungiasis prevalence will be evaluated per site, while data on STH and enteric infections will be pooled across sites.

Expecting a postintervention STH infection prevalence of 15% in the control arm and 10% in the intervention arm, with an ICC of 0.1, and assuming five enrolled participants per household with a 15% loss to follow-up, 220 households per arm in total across the two sites provides 80% power at 0.05 significance. This sample size is also sufficient to detect at 80% power and 0.05 significance: (1) the expected difference in enteric infection risk in children <5 years old—assuming one <5 year old per household, and an expected prevalence postintervention of 70% in the control arm and 56% in the intervention arm; (2), the expected difference in tungiasis prevalence in children 1–14 years at a site level—assuming two children aged 1–14 per household and an expected prevalence postintervention of 30% in the control arm and 15% in the intervention arm. Based on these estimates, we aim to enrol 220 clusters (households) per arm—thus ensuring 220 children (aged under 5 years) and 440 children (aged 1–14 years) per arm across both sites.

Sensitisation and engagement

Stakeholders at national, county and subcounty levels have been involved from conceptualisation. Prior to formative research, meetings were held at county, subcounty and ward levels (with village representatives) to introduce the study. Following formative research, consultative meetings were held at the ward and village levels, presenting research findings and compiling input on acceptability and ownership of the proposed intervention.28 This input informed development of care guides, and how to undertake engagement, given not all households would be eligible. Following selection of study villages and identification of eligible households, meetings are held with ward-level and village-level leadership to discuss plans for research participation and intervention roll-out. Additionally, preparation meetings are held with participants several days prior to the installation of the flooring intervention.

Eligibility and enrolment

Inclusion criteria for households are as follows: at least one child under 5 years; home ownership; building where children under 5 sleep meets structural criteria (unimproved earthen flooring throughout, structurally sound); members providing informed consent; willing to temporarily relocate and provide water for installation. Exclusion criterion for households is as follows: intending to move within the next 12 months.

To assess eligibility, a two-stage census and eligibility assessment is conducted in study villages. Households are visited and a questionnaire is administered including questions on members’ sex, age and living arrangements, as well as direct observations of the dwelling’s building characteristics. Households meeting the core eligibility criteria are visited by a mason to assess the building for structural eligibility using a survey tool developed by structural engineers together with local masons. Eligible households are offered the opportunity to enrol and have baseline assessments undertaken.

Randomisation

Site-level randomisation of households to control or intervention arms is stratified by village and done through a ceremony conducted in each site upon completion of baseline assessments. The ceremony is attended by members of the study team, village-level and ward-level leadership, and representatives from participating households, who each select a sealed card, indicating group A or B. Allocation to control or intervention arms is announced after all households have drawn cards.

Timeline

The census and eligibility assessments were carried out in March 2023 in Kwale and June 2023 in Bungoma. Baseline assessments (0 months) are undertaken with all households prior to flooring installation in intervention households. Installation occurs following baseline assessments. Midline assessments are carried out 6 months post midpoint of floor installation. Three unannounced spot-checks are performed throughout the implementation period. Endline assessments are carried out 11–12 months after floor installation.

Assessments

The assessments undertaken are described below and outlined in online supplemental file 2 . Once enrolled, any household which moves house is followed up in their new location, with a note made of this, providing they are still within the study sites.

A household questionnaire is administered to the primary caregiver at 0 and 12 months, including questions on demographics, assets, livestock and poultry ownership and husbandry practices, WASH infrastructure and practices, and building and room uses. WASH facilities and building conditions are directly observed and GPS coordinates of the dwelling collected.

A caregiver questionnaire conducted with primary caregivers at 0 and 12 months includes prevalidated question modules based on the WHO Five Well-being Index (WHO-5)48 and WHO Quality of Life BREF (WHOQOL-BREF)49 50 to evaluate respondent psychological well-being, and to ask about episodes of gastrointestinal illness in the last 7 days for all children under 5 years.

A child questionnaire is conducted with children aged 8–14 years at 0 and 12 months comprising questions about psychological well-being and perceived quality of life drawn from the WHO-5 and the EuroQol 5-Dimension Health-related Quality of Life for Youth (EQ-5D-Y) tools.51 52 The modified dermatological quality of life index for tungiasis is administered to children with tungiasis.17 53

During the stool survey, all household members are invited to provide a single stool sample at 0 and 12 months for screening for enteric infections (children under 5 years) and STH infections (all residents over 1 year). Participants are provided with sample collection kits and instructed how to collect an early morning sample for collection and transport to the laboratory by the study team.

A tungiasis assessment is conducted with children aged 1–14 years, supported by a parent/guardian if appropriate, at 0 and 12 months by trained field workers. Diagnosis of tungiasis is made on the basis of the presence of embedded fleas and manipulated lesions in participant’s feet. The viability and number of lesions are recorded.

Entomological sweeps undertaken at 6 and 12 months include collection of sand, dust and fine debris from floors in three or more locations in intervention and control households: (1) food preparation areas, (2) sitting/resting locations indoors and/or outdoors, and (3) child sleeping areas. Improved floors are swept and dust removed from crevices and gathered into zip-lock bags. Soil from unimproved floors is gathered with small brushes or trowels. Zip-lock bags are shipped in cool-boxes to the laboratory.

Environmental surveys examine for the presence of enteric pathogens and STH at 12 months. Floor dust samples are collected in three locations within dwellings (75 per arm per site): (1) 50cm in from the exterior entrance to the primary building, (2) 50 cm inside the room used for cooking, and (3) 50 cm inside the room where the youngest child sleeps. Adapted from soil sampling protocols,19 54 55 study personnel demarcate surface areas using a sterilised 50×50 cm stencil and sweep the enclosed surface with a sterilised paintbrush. This is repeated until, at each location, 2 g dust has been collected into sterile 2 mL sample collection tubes.

Unannounced spot-checks are performed at three points during the implementation period in all intervention households. A structured checklist is used to assess the following: floor structural condition (cracking, abrasion, water resistance); levels of debris; and location and duration of daily activities. An adapted checklist is conducted twice in control households, assessing levels of debris and location and duration of daily activities.

Semistructured observations are undertaken in 20 intervention and 20 control households in each study site at 12 months to quantify time spent by children, caregivers and animals in different parts of the dwelling and to identify where different activities are undertaken.

In-depth interviews are conducted at 12 months with caregivers from a random sample of intervention houses in which observations are performed (12 per study site). These cover user satisfaction with the floor, cleaning regimens, animal husbandry practices, food preparation, relaxation, child caregiving and perceptions on social status and the economic value of the floor. Interviews take 30–60 min and are conducted by trained interviewers using prepiloted question guides.

Cost data collection: Cost data are collected using an ingredients approach and a standardised costing framework, based on semistructured questionnaires and consultation of the project accounts.

Laboratory work

Stool samples from children <5 years are aliquoted into cryovials for storage in 95% ethanol at −80°C. DNA is extracted from the stool using the QIAGEN QIAamp DNA Mini Kit56 and real-time quantitative PCR is used for the detection of prespecified enteric pathogens: G. lamblia, Cryptosporidium spp., C. jejuni, enteroaggregative E. coli (EAEC), Shigella and adenovirus.

Stool samples from individuals >1 year are processed and examined in duplicate (41.7 mg template) for presence and quantity of STH eggs using the Kato-Katz method.57 Duplicate slides are prepared from the single day stool samples and are read by independent microscopists with a 10% quality control check performed.

Samples collected during the entomology sweeps undergo heat extraction of soil arthropods using a Berlese-Tullgren apparatus.58 59 The presence and abundance of eggs, larvae, pupae and adults T. penetrans and other flea species as well as other arthropods, such as bedbugs, are recorded using a stereo microscope.

DNA is extracted from soil samples collected during the environmental surveys according to specifications outlined in the QIAGEN DNeasy PowerSoil Pro documentation (QIAGEN, Hilden, Germany) and is assessed for the preselected enteric pathogens and STH egg/larvae DNA using the established real-time PCR panel.

Patient and public involvement

Intervention design, including logistics of fitting floors and development of floor-care guides, was shaped by findings from formative research and direct feedback from community members during community engagement activities. Floors are installed by local masons from study communities, with participant households providing water for curing. Community members provided feedback into the range of assessment methodologies used and proposed for the trial.