763 - Supporting community health workers as vaccine educators: Effectiveness of digital training and a chatbot communication aid to promote routine childhood immunization in Kenya

Publication Number: 763.6935

Victoria Ward, Stanford University School of Medicine, Palo Alto, CA, United States; Nadia Hemmat, University of California, San Diego & San Diego State University, San Diego, CA, United States; Y. Xian Ho, Dimagi, Inc., Cambridge, MA, United States; Nophiwe Job, Stanford University School of Medicine, Cape Town, Western Cape, South Africa; Jamie Sewan Johnston, Stanford Center for Health Education, Stanford, CA, United States

Background: Immunization is among the most important, cost-effective measures to reduce child morbidity and mortality, yet hesitancy is rising worldwide. Community Health Workers (CHWs) are vital to promoting vaccine acceptance, but face barriers as vaccine educators, including the spread of misinformation, lack of reliable information, and limited training in behavior change communication. Digital tools can support CHWs' capacity to address vaccine hesitancy. We developed a vaccine promotion package including training videos, job aids, and a patient-facing chatbot to promote childhood immunizations in rural Migori County, Kenya.

Objective: We aimed to test the effectiveness of digital tools to support CHWs' impact on vaccination uptake in rural Kenya.

Design/Methods: We conducted a cluster RCT of 781 CHWs randomized to receipt of the digital training and chatbot tools and compared them to a control arm receiving standard training. Follow-up surveys were conducted with CHWs 2 months after the intervention, assessing CHW knowledge, beliefs, and preparedness to address vaccine hesitancy. Follow up surveys and health records were then collected 8 months following rollout from 1600 households to investigate the interventions' impact on patient knowledge, beliefs, intent to vaccinate, and vaccine uptake.

Results: Overall, the tools had high acceptability: 91% of intervention CHWs reported the digital training videos “very helpful” for sharing vaccine information with patients, 95% found videos “very helpful” for communicating with patients, and 70% of CHW respondents would recommend the chatbot. Differences in CHW beliefs were not observed, except that control arm respondents were more likely to think children received too many vaccines (p=0.05). We did not observe a difference in CHW knowledge in the 15-item assessment: The mean knowledge score of the treatment group was slightly higher than the control group (10.04 vs. 9.92) but was not statistically significant (p=0.32).

Conclusion(s): CHW survey findings reflect a well-trained and pro-vaccine CHW workforce that receives regular training and support, however this is rarely the case for most CHW cadres who often are the primary providers of maternal child health worldwide. Treatment group CHWs indicated a need for tools to share complex concepts and reinforce their credibility with patients, and reported that these digital tools were very helpful in their communication and information delivery regarding vaccines. In the next phase of the study, we will examine health impact of the intervention on patient vaccination rates.