The impact of pre-emptive home delivery of ORS + zinc on treatment for child diarrhea: a randomized controlled in Bauchi, Nigeria

The intervention will involve pre-emptive distribution of two co-packs of ORS+zinc per child to every household with at least one child under the age of five. Each co-pack includes two sachets of ORS and one strip of 10 zinc tablets, enough to be able to treat two cases of child diarrhea. This research will evaluate the impact of one round of deliveries.

2024-06-01

2025-08-31

The primary outcome for the study is self-reported ORS use for a case of child diarrhea that occurred within the last 4 weeks. We will measure this outcome average over 6 months, 12 months, and monthly.

This will be measured through a series of survey questions which ask caretakers who reported having a child who had a diarrhea episode in the past 4 weeks the following set of questions:
1. “Did you give [child name] anything to treat the Diarrhea?”
2. If yes to Q1, “can you tell me or show me what treatments you gave [child name] either home-prepared or from outside of home?” Do not prompt responses.
3. If ORS not mentioned in Q2, “did you give the child any ORS to treat the diarrhea?”
Our main outcome variable will be a binary variable that is set to 1 if the respondent reports that they used ORS in 2 or 3 and to 0 if they reported that they did not use ORS in 2 and 3 or if they reported that they gave no treatment to the child in (1). We will conduct a sensitivity analysis where we use only the unprompted response from (2) to create our ORS use variable.

Zinc + ORS use
Antibiotic use
Zinc use alone
Time to ORS initiation
Exposure to unsafe drinking water
Willingness to purchase new ORS packets
Child mortality; all cause and from diarrhea
Hospitalization from diarrhea
Number of diarrhea episodes
Care seeking outside the home
Wastage of ORS and zinc packets
Receipt of ORS and zinc delivery

We will randomize wards to treatment or control groups before baseline data are collected. We will randomly assign wards to either of the two groups listed below, and let the randomization dictate which of the two groups end up with an extra ward:
• Group 1 – Delayed-Start Control (cN = ~162): No intervention will take place until after endline data collection in completed, after which the wards in the Control group will receive the intervention. During the study period, the caregivers in this group will have standard access to ORS and zinc at local health facilities and pharmacies. Some community health workers (called Community Health Influencers and Promoters, and Services or CHIPS(16)) in control villages could make household visits; however, we do not expect any delivery of ORS or zinc in the control group as CHIPS are generally not the source of diarrhea treatment.(12)
• Group 2 – Household Visit + Information + Free pre-emptive distribution of ORS (cN = ~162): Campaigners recruited by CHAI will be paid NGN40 for each household visited in their catchment area that contain a child under 5 years old at the beginning of the study. During the distribution visits, the campaigners will train caregivers on the dangers of diarrhea and the importance of ORS and zinc use, among other things (including how to prepare, use, and store ORS/zinc, benefits of the treatments, recommended health behaviors such as seeking care, and encouraging basic handwashing and hygiene practices). The caregivers will also receive a flyer describing the same information in their local language, for future reference. Campaigners will then distribute 2 ORS and zinc co-packs (each co-pack contains 2 sachets of ORS and 10 tablets of zinc) for free for each child under the age of five in the household.

The participant and the intervention implementer will not be blinded to the random assignment due to the nature of the intervention.

in office by computer

Ward (sub-state geographic regions)

Yes

323

68000 households. We expect 13% diarrhea prevalence in the previous 4 weeks which would give ~9000 diarrhea cases.

161 clusters per arm

Assuming a cRCT design with two-point data collection, a diarrhea prevalence rate of 13% (for a conservative MDE estimation) with an expected number of diarrhea cases per cluster of roughly 14 at endline, a control endline ORS use rate of 40%,(13) and a conservative intra-class correlation (ICC) of 0.2, we will be sufficiently powered to detect a minimum difference of 7.8 percentage points in ORS use between control and intervention groups over months 1-6.(19) Under the same assumptions, the study will be sufficiently powered to detect a minimum difference of 7.3 p.p. in ORS use between the two groups over the study period (12 months). The pooled estimate of the impact over six months and twelve months will be our primary analysis, but we will also conduct analyses to assess the impact in each month. However, since we will sample only 288 EAs per month, and to maximize power, we will expect to sample 288 wards, these parameters will allow for a minimum detectable effect (MDE) in each of the twelve months following the deliveries of 11.2 p.p. per month. Table 2 shows what MDEs will be if the ICC is 0.15 or 0.1 instead of 0.2.

Protocol and pre-analysis plan

Protocol and pre-analysis plan (adding strata FE)