2016 ICT4D Conference

Delivering interventions for acute malnutrition in low-resource settings is an arduous endeavor. Beyond the effort of managing the sheer caseloads, other challenges faced include logistics, stock management, development and enforcement of protocols, high turnover of service-providers, limited visibility into the magnitude of the problem and, subsequently, of their impact. Having a tool that could provide and centralized solutions for several of these challenges could and has proven useful to several organizations. This aim of this abstract is to share the experiences of Save the Children in using mobile health to begin to address some of the aforementioned challenges in Kenya.
Project implementation: Starting in 2014, World Vision Canada and Save the Children partnered with Dimagi an organization involved in developing mobile health applications designed as job-aides for health workers for Integrated Management of Acute Malnutrition (IMAM). This happened in a consortium and is being undertaken in four countries namely Kenya, Niger, Chad and Mali. In this agreement Dimagi would integrate the World Health Organization (WHO) protocol into an algorithm that enforces the global malnutrition treatment protocols that health centers should follow. Reflecting a decision-tree, the application guides clinic staff through screening for malnutrition, classifies the degree of the illness, prompts recommendations for treatment and counseling, and automatically lists beneficiaries into groups for follow-up. On the back-end, viewable by program staff, the application generates both granular data about each visit for every beneficiary (pregnant or lactating women, or children under five) as well as global data, for trends across the intervention areas.
In Kenya deployment started in January 2015, the IMAM application was assessed and adapted to the Kenyan Context in line with IMAM guidelines from the application made for Niger. Following a scoping visit made in Wajir, the original application was modified and contextualized. The modification factored inputs from local health protocols, Z-Score calculations auto-generated by the mobile platform, locally-vetted, personalized counseling messages in multiple languages and dialects, as well as culturally-relevant images. The prototype was then tested with select users and refined before training and launching in selected health clinics. The entire process took about 10 months and actual use of the application was in November 2015. So far using the application health workers have been able to reach 131 Pregnant and lactating women and 118 (60M, 58F) children in the Outpatient Treatment Program (OTP) and 156 (75M, 81F) children in Supplementary Feeding Center (SFP).
Methodology: In order to generate evidence, an evaluation using the Randomized Control Cluster Trial method is being done to run concurrently with the pilot implementation. An initial baseline assessment was undertaken, which was followed by a random selection of 20 health facilities from three sub counties (Wajir County (Wajir East, Wajir South and Tarbaj). The health facilities were stratified by sub-county and mobile signal strength and provided with mobile devices with the application. The remaining 25 health facilities in the three sub-counties continue to use the paper based system so as to act as comparison sites. In order to monitor the project, periodic observations will be conducted to compare quality of care in intervention and non-intervention sites together with prospective and historical data collection completeness, timeliness and quality.
Findings: The project implementation is still underway but the following have are the key learning's in so far;
- Field test and retesting of the application with end users (front line health workers) is vital
- Working in a consortium is more beneficial than working as a single entity, as lessons learn between contexts and implementing partners can be maximized.
- Its beneficial to work with a technology partner specialized in working in mHealth in low resource contexts since then you leverage on their experience
- While sensitization and involvement of Ministry of Health (MoH) is critical, it is a time consuming process.
- Evidence generation for learning and informing program implementation is crucial however, it is complex to implement and integrate with the pilot
Conclusions and implications: The quality of care and data potential of the mHealth application has not yet been sufficiently exploited since the program is still on going. Findings from the evaluation will answer questions on the applications potential for improvement of quality of care and empowerment of health workers in better management of acute malnutrition.
Acknowledgement: The project was funded by the USAID Office of U.S. Foreign Disaster Assistance (USAID/OFDA) through World Vision.