Conversational Agent Vaccine Promotion RCT

Overview

The objective of this study is to assess the use of and satisfaction with the ECA intervention over a 12-month period, its ability to increase SARS-CoV-2 and influenza vaccination in the BMATP community, and the comparative effectiveness of proactive engagement strategies and cultural tailoring on these factors. The Investigators will assess app use and satisfaction among these participants, along with self-reported vaccination attitudes and behavior, at 6 months and 12 months. During this time the investigators will conduct a 2x2 factorial RCT (with the individual the unit of random assignment and measurement) to assess the impact of two app design features on engagement and outcomes: (1) the investigators will manipulate engagement mechanisms (ENGAGEMENT), including reminder notifications and trust-building dialogue by the ECA and, (2) independently manipulate cultural tailoring of vaccination promotion counseling language used by the agent (TAILORING) to either adaptive religiosity (tailored) or secular (non-tailored). The investigators' primary hypotheses are that participants with have significantly greater vaccination completion rates in the high engagement and tailored conditions at 6 months (H1) and 12 months (H2) compared to other conditions. The investigators' secondary hypotheses are that participants will use the app more, be more satisfied with the ECA, be more advanced in their stage of change, and have greater knowledge, self-efficacy, and decisional-balance from baseline to 6 and 12 months, in the high engagement and tailored conditions. In addition to the RCT participants, all members of the 12 participating BMATP churches will be invited to use the app (via snowball recruitment), to increase the utility of the social networking functions of the app by having as many users as possible.

4.1.a.1. Randomization. Study participants will be randomized into one of the four study conditions (ENGAGEMENT x TAILORING) using blocked randomization with randomly permuted block sizes of 2 and 4, by software on the central server referenced when participants first activate the app on their phone. Participants will continue using the app at home, and will be called at 6 months and 12 months to assess outcome measures during a videoconference call with a research assistant. 4.1.a.2. Qualitative Interviews. The investigators will conduct interviews with a sample of participants completing the study (12), as well as with Church and Health Ministry leadership (12) to understand overall acceptance of ECA-ACE and barriers to deployment and use. 4.1.a.3. Qualitative Methods for all Project Focus Groups and Interviews. Videoconference-based focus groups and interviews will be facilitated using a semi-structured interview guide, audio-recorded, and transcribed verbatim by a professional transcription service. The investigators will conduct a thematic analysis of these transcripts guided by the investigators' research questions with the aim of identifying design insights. The investigators will use grounded theory analysis, including constant comparisons, axial and selective coding, and memo writing. Using NVivo 12.5.0 software, two researchers will inductively code transcripts separately, labeling emergent phenomena in the data to arrive at a codebook. Two researchers will then independently apply the codebook to participant transcripts, meeting regularly to discuss discrepancies in the applications of the codes, updates to the codebook, and contradictory data. Sample sizes specified should be adequate to reach thematic saturation relative to the investigators' research questions.