Impact of Improved "Injera" Baking Stove Use on Childhood Acute Respiratory Infection Prevention in Northwest Ethiopia

1. Study area: The implementation of the study project has been started among rural and urban communities of Mecha Health \& Demographic Surveillance System (MHDSS) Site in Northwest Ethiopia. MHDSS site is a field research center established by Bahir Dar University (BDU) to conduct field researches \& support major graduate level research projects principally. The field research center is located at 525 km far away from the capital city of Ethiopia, Addis Ababa, towards Northwest. This field research center has been selected as study area in light of the following particular rationales:

• Relevance of the project to the community in terms of current biomass fuel use, climatic diversity \& local acceptability of the improved stove intervention
• Strategic research importance of the locality to achieve the mission's of BDU to be one of the ten premiere research universities in Africa by the year 2025.
• Opportunities to be benefited from the existing field research infrastructures of MHDSS
• Availability of local collaborators with experience of improved stove technology transfer
• Availability of improved baking stove initiatives already operating in the study area
• Good field experience of investigators at the study area

1. Study design: Cluster randomized controlled study will be employed to test the effectiveness of improved "Injera" baking biomass stove intervention in reducing both household air pollution (HAP) \& childhood acute respiratory infection (ARI) incidences through comparing children living in households (HHs) with different baking stoves over one year follow-up period. With this study design, HAP \& childhood ARI outcomes will be measured before the improved baking stove is installed and again after part of the HHs received an improved baking stove. The overall study procedures will involve the following three major phases:

• Base line data collection using HH interview and air monitoring methods.
• Installing improved "Injera" baking biomass stove in the main cooking area of the intervention HHs to be used as primary baking stove at least for one year.
• Consecutive data collection using the HH interview method every 2 weeks and measure the new amounts of indoor particulate matter in the main cooking area using standard HAP monitoring device every 3 months for one year.

1. Sample size: For the primary outcome, the sample size was calculated by applying the two-sample comparison of proportions formula using STATA with equal participants in both arms considering 21% proportion of childhood ARI, detectable difference of 20%, \& two-tailed alpha of 0.05, power of 80%, intra-cluster correlation coefficient (Ƿ) value of 0.01, coefficient of variation of 0.6 for cluster sizes, 55 average number of children within each cluster (m') and over sampling of 10%; the required sample size with cluster randomization becomes about 2750 within 50 clusters per arm. Similarly, for the secondary outcome, the sample size was calculated by considering detectable difference of 60% in mean indoor PM2.5 concentration (µg/m3), correlation coefficient (Ƿ) of 0.5 for indoor PM2.5 measurement variations in biomass using HHs, 2-sided alpha of 0.05, power of 80% and 55 average number of children within each cluster (m'), 0.6 coefficient of variation for cluster sizes and over sampling of 10%; the required sample size becomes about 990 within 18 clusters per arm. Thus, among the HHs included in the primary outcome study, only part of the HHs will be included for the secondary outcome study.
2. Sampling method: Using cluster sampling technique, clusters/"Gotes" will be selected randomly to represent the total population \& then all HHs with under 4 years old child with in the selected clusters will be included. A cluster is defined as the small village, termed as "Gote" for rural or "Ketena" for urban settings in Amharic (both national \& local language), used as the smallest unit of enumeration area by Ethiopian national census. According to the update data report of MHDSS at the end of 2017, each cluster/"Gote" includes about 55 average number of children. To set the randomisation list in advance based on eligibility criteria, the sampling frame (list of HHs with children less than 4 years old) is established form the MHDSS update report data. Finally, the selected HHs will be identified for data collection using the specific MHDSS house number \& the youngest child equal or less than 4 years old will be recruited for study from each HH. In situations where there are 2 or more under-four children living in the same HH, only the youngest child will be included in the study.
3. Participant recruitment: To facilitate the participant enrolment process, community sensitization will be conducted through public communications during local gatherings \& regular health development army (HDA) meetings as well as through home visit by health \& agricultural extension workers to let people know that childhood ARI \& HAP studies will be carried out. The actual participants will be recruited at HH level by field health workers during baseline survey after ensuring whether the HHs meet the eligibility criteria. When HH meets the eligibility criteria, the study will be explained to parent/s of the index child \& they will be asked whether the HH is willing to participate in the studies or not. When both parents agreed to be involved, the field staff will administer a written consent form in local language.
4. Participant allocation: An independent epidemiologist was performed the allocation to control \& intervention arms which will be revealed after all baseline measurements have been completed as well as all study HHs recruited \& assigned to their respective arm to ensure allocation concealment.
5. Intervention adherence and compliance monitoring strategies: To avoid the potential detrimental effects of non-adherence, a variety of strategies will be used. In view of that, timely response to trial-related difficulties such as rapid maintenance or replacing of defective stoves will be used as strategy to improve adherence to the intervention protocol. Regarding trial protocol compliance monitoring, visual inspection through unannounced visit of HHs will be conducted to enhance validity of data.
6. Data collection methods: For the primary objective, data collection will be carried out by local nurses through face-to-face interview with index child mothers/caregivers using questionnaire \& through direct verification whenever essential \& possible. After baseline HH survey, the follow up survey will be carried out for one year through HH visits every 2 weeks. Concerning the duration, since seasonal factors typically have a major effect on HAP level in Ethiopia the duration of the follow-up period will be one year to cover the major Ethiopian periods of annual weather changes. Data collection manual containing a detailed description of the questions \& other administrative issues will be used to facilitate data collection process. For the secondary objective, baseline HAP exposure (indoor PM2.5 concentration) will be measured in the main cooking area using standard HAP monitoring device and Standard Operating Procedures (SOPs) by trained household air monitoring team. In addition, independent variables data will be collected at baseline through face-to-face interview with index child mothers/caregivers using structured questionnaire. After baseline survey \& implementation of improved stove intervention, subsequent indoor PM2.5 concentration measurements \& other covariates data collection will be carried out every 3 months for one year.
7. Data handling and analysis: All original participant files will be stored in a secured manner for a period of at least 2 years after completion of the study. Data integrity will be enforced through a variety of mechanisms such as double data entry; valid values checks \& consistency checks against data already stored in the database (i.e., longitudinal checks). Considering the underlying design of the study, assumptions regarding the inter-relationship between the independent \& outcome variables under investigation as well as the repeated \& dependent nature of the data, Generalized Estimating Equation modeling method of data analysis will be applied to determine the effect of improved baking stove intervention on both HAP \& childhood ARI incidence.
8. Ethical consideration and trial registration: The proposal had been reviewed \& approved on May 08/2018 by Institutional Ethical Review Board of Bahir BDU. Written informed consent will be obtained from all participants. The informed consent process emphasizes that participation in the study is voluntary \& consent to participation can be withdrawn at any time, without giving a reason as well as without affecting their current or future benefits to which the participants are entitled. Moreover, to maintain justice control HHs will be offered improved baking biomass stove at the end of their participation. Whenever ARI illness is occurred, field health workers will refer the child to the nearest health institution for proper treatment.
9. Close-out procedures:

The trial will be terminated at the planned target of one year after the last participant has been randomized or at an earlier or later date if the circumstances necessitate. Regardless of the timing \& circumstances of the end of the study, close-out will proceed in 2 stages: the first stage covers the interim period for analysis \& documentation of study results which is expected to take about 3 months. The second stage will comprise debriefing of participants \& dissemination of study results.