Kilkari Impact Evaluation

The Johns Hopkins Global mHealth Initiative, Oxford Policy Management, and University of Cape Town proposes supporting BBC Media Action and the Ministry of Health and Family Welfare (MOHFW) to conduct an external evaluation of Mobile Academy (MA) and Kilkari. In the below text, we outline 6 proposed activities which comprise the evaluation, which are planned from mid 2018 to April 2020.

These studies will help inform program modification and scale-up plans, and will help better understand the effects of maternal mobile messaging in India and globally.

1. Secondary analysis of cross-site trends in key performance indicators (Duration: throughout study): Mobile Academy (MA): We will use system generated data to estimate the proportion of ASHAs who have completed MA training across 12 states. We will additionally aim to understand linkages between ASHA's exposure to the MA training content and performance on MA knowledge assessments. Kilkari: We estimate that Kilkari users will have differing levels of exposure to the program. For example, some women may not be present when the Kilkari calls come and thus miss out on key content. Analyses of system generated data will aim to measure exposure to Kilkari messages by content area (e.g. nutrition, family planning, etc) across 12 states where implementation has occurred to date. As part of this analysis we'll seek to understand reasons why calls are not received, including phones being out of network, changes in SIM cards, etc.
2. Summative evaluation of Kilkari (Period: April 2018 to December 2019) Objective: To determine the effectiveness of mobile maternal health information messages on key health outcomes in 4 districts of Madhya Pradesh (MP). This objective will be carried out through an individually randomized controlled trial initiated in early 2018. We propose conducting household surveys in all 4 study districts among the cohort of women identified during conception, starting in March 2018. Women identified within the 5-7 month of their pregnancy, who have access to a mobile phone, and consent to participate will be interviewed as part of baseline survey and recruited for participation into Kilkari by 8 months of conception (this is intended to mirror timings observed in Kilkari outreach in other states). Among individuals not randomized to receive messages, we would propose delaying implementation for Kilkari for a 12-month window to allow for evaluation activities to take place and additionally glean information which is anticipated to improve the efficiency of implementation / rollout in other areas.
3. ASHA Survey Objective: Identify factors underpinning differences Reproductive Maternal Newborn and Child Health (RMNCH) knowledge and motivation among frontline health workers, including exposure to the Mobile Academy training program. As part of the baseline surveys described under activity 2 in MP, we will conduct face to face interviews with all ASHAs (n=1,200) across the four study districts in MP. A sub-sample will also be contacted for a follow-on phone survey a week later, to pilot modules and validate a phone survey tool on knowledge and motivation, which can then be applied to other ASHAs to measure key outcomes as part of MA deployments in other states, including Rajasthan.
4. Post-natal care phone survey Objective: Measure respectful maternity care (RMC) during childbirth, essential newborn care, and infant feeding practices within 0-3 months postpartum As part of baseline survey activities in MP, we propose additionally identifying women in 1-2 districts who are 0 to 3 months postpartum (n=880). Face to face interviews with postpartum women will provide a forum to measure key infant feeding and essential newborn care outcomes. Phone surveys, a week later, with a subset will help develop a validated phone survey tool for their measurement amongst individuals enrolled into Kilkari.
5. Qualitative interviews of both Kilkari and MA in MP and Rajasthan Objectives: Mobile Academy

• Understand stakeholder (ASHA, ASHA supervisor and beneficiary) perceptions of MA, including barriers / facilitators underpinning uptake and completion of MA training and enacting its content, and how it affects interpersonal communication between stakeholders. Kilkari
• Understand stakeholder (ASHA, ASHA supervisor and beneficiary) perceptions of Kilkari, including different message content areas, barriers / facilitators to the adoption of Kilkari's recommended RMNCH practices and careseeking behaviors
• Identify barriers / facilitators for pregnancy and birth registration The qualitative research component will complement the summative evaluation and secondary data analysis components. Qualitative research will include in-depth interviews and focus group discussions in both MP and Rajasthan. Differentials in uptake of MA across these settings provides a unique opportunity to understand the characteristics of ASHAs and the larger health systems environment which contribute to MA completion. Among Kilkari beneficiaries, qualitative research can help better understand user perception of content, the extent and mechanism of impact of the messages on health-related knowledge, behavior and relationships.

1. Economic evaluation of Kilkari Objective: To determine the incremental cost-effectiveness of Kilkari as compared to a null scenario of no messages To estimate the value for money of Kilkari, we will estimate the costs of program to all key stakeholders involved-including BBC Media Action, MOHFW, and end-users. To estimate the effects of the program on health outcomes, we will draw from household survey data at baseline and endline amongst individuals exposed and not exposed to messages. Data on changes in careseeking for antenatal care as well the adoption of practices around breastfeeding and family planning will be inputted into the Lives Saved Tool to generate a measure of lives saved associated with program activities. Lives saved will be used to yield an estimate of the incremental Disability-Adjusted Life Years (DALY) averted and ultimately a cost per DALY averted.