Community engagement through supportive feedback and non­monetary incentives: a randomised controlled trial in Ethiopia

Authors: Abebual Demilew, Besfat Wodajo, Margaret McConnell and Saugato Datta

Project Description

Zerihun Associates is partnering with Marie Stopes International Ethiopia (MSIE) and ideas42, a non-profit firm that uses behavioral science to improve social outcomes, to increase child immunization rates in Ethiopia. Despite substantial economic progress, only 24% of Ethiopian children aged 12-13 months are fully immunized, with rates for the lowest three wealth quintiles below 20%.[1] Oromia region, the largest region in the nation has even levels of immunization coverage. The 2011/2 Ethiopia Demographic and Health Survey, the latest data available on immunization, reported full immunization rate for the region at 15%, 40% below the national average. Dropouts between the first and the third rounds of major vaccines is as high as 20% in some districts of Oromia region (HMIS[2], 2014).

To improve full immunization coverage and reduce dropouts we developed multiple designs and through an iterative process have refined the design based on implementation feasibility, how well the design addresses the identified bottlenecks, and costs associated with both piloting and scale-up. User testing in Oromia in April 2015 will allow us to refine and finalize the design.

The design has two parts that work in tandem: giving HEWs the ability to easily track dropouts and a system for making follow-up of drop-outs easy while providing recognition and social proof to both the HEW and caregivers. Utilizing a planning prompt to create a moment of action, the “HEW Outreach movement” prompts HEWs to think about outreach for immunization at the right time and to perform targeted outreach to dropouts. Utilizing social proof and a salient tracking system, the “Stamp System” component makes it simple for HEWs to track dropouts and for mothers to take additional ownership of and track their child’s progress and see that it is both normal and desirable to immunize their child.


This is a randomised evaluation that also has a qualitative research component. Local health posts (HPs) will be used as the unit of clustering and individual households as the unit of randomisation. A total of 90 HPs will be equally and randomly assigned to treatment and control arms. From the catchment area near each HP, 30 households with children less than 13 months of age will be randomly selected. The total sample size is estimated to be 2,760 children. Quantitative data, such as number and timing of home visits and immunisation status of children, will be collected during the baseline surveys. Knowledge among households about immunisation will be tested using a Likert scale. Their intention to immunise will also be measured using quantitative indicators like ranking child health activities in the order of importance to the household. The study team will also carry out stakeholder interviews at both the baseline and endline stage. The intention to vaccinate will be measured through subjective questions presented in the baseline and endline surveys.

[1]  Ethiopia Demographic and Health Survey 2011, Central Statistical Agency, Addis Ababa, Ethiopia.

[2] Facility based Health Management Information System

Leave a Comment

Your email address will not be published.