We convene the rural CAP using the “broad community” model to capture participants’ opinions about vaccination [20]. We work directly with Cooperative Extension Educators at Oklahoma State University (OSU) to identify and partner with (1) key stakeholders (eg, community leaders) and (2) members of the community for panel development. Cooperative Extension is a key branch of the land-grant system and includes community outreach and health-related engagement with populations across the state, including rural adults [21]. We use purposive sampling to identify, with educators’ input, and recruit key partners within rural counties with the lowest vaccination rates at the time of recruitment. We recruit a minimum of two individuals in the bottom 10% of vaccinated counties across Oklahoma to participate in the CAP. Our target total sample size for phase 1 is 15-20 CAP participants in order to achieve this benchmark, which is consistent with existing recommendations in qualitative research [22].

Inclusion criteria include being ≥18 years of age, residing in a target county and being identified as a key stakeholder, and willingness to provide informed consent and complete a semistructured interview. There are no exclusion criteria.

We conduct one-on-one semistructured interviews with each participant (see Multimedia Appendix 1). Interviews include open-ended questions to gauge barriers and facilitators to vaccination. To promote the diversity of opinions, research personnel are trained in using open-ended questions that are developed with empathy-driven, user-centered principles [23] and in using MI-consistent prompts (eg, “Please tell us about the impressions of vaccines in your community” or “Please share your experience with vaccinations”). For individuals who express vaccine resistance, questions focus on their perspectives (eg, “Tell me what you don’t like about vaccines”) and hypotheticals (eg, “Imagine you were to get the vaccine, what would those circumstances look like?”). For those who have been vaccinated, questions similarly focus on community perceptions and participants’ own experiences (eg, “Tell me about your decision to get the vaccine”). Lastly, questions also center on mHealth (eg, “Imagine you were to use an app to help you make health care decisions – what would that look like to you?”). All interviews are about 30-45 minutes via phone, video call, or in-person if preferred. Interviews are recorded and transcribed verbatim for analysis.

Interviews conducted with the CAP are transcribed verbatim and triple-coded by 3 separate members of the research team to inductively identify themes classified broadly as barriers to and facilitators of receiving the COVID-19 vaccine as well as the use of mHealth using content analysis. In addition, all transcripts are reviewed for identification of SCT constructs and opportunities for the use of MI relational skills, technical skills, and existing empirical evidence [24,25]. This approach uses a reflexive strategy to integrate three different researchers’ perspectives into the data set (eg, experience in SCT versus MI) [26]. Coding matrices are used to match MI- and SCT-specific elements to barriers to and facilitators of vaccination to design potential mHealth-specific messages. Discrepancies in coding are discussed to achieve intercoder agreement, and inductive thematic saturation is achieved at the time of analysis by iteratively reviewing the emergence of new themes during triple coding [27]. If saturation is not achieved in that new themes emerge at the time of analysis that are not well integrated into the codebook, additional community members will be recruited to complete interviews (maximum total n=20) [28]. The final data set is entered into the NVivo software (Lumivero, LLC) package [29]. Codes are organized into barriers and facilitators; other categories can be created as they emerge during the coding process.

We leverage multiple recruitment strategies in an intentional effort to reach historically underrepresented areas across Oklahoma. For phase 2, we partner with OSU Cooperative Extension and TrialFacts to target primarily rural counties to recruit a total of 60 participants. Around 60 participants are selected as consistent with recommendations for pilot studies, taking into consideration the focus on evaluating feasibility, as well as practical restraints [30]. TrialFacts is a participant recruitment service company that provides targeted recruitment of specific populations of interest.

Eligible adults are (1) aged ≥18 years, (2) living in a rural area (based on Rural-Urban Commuting Area codes; rural, codes 4-10) [31,32], (3) self-report of never having received any dose of a SARS-CoV-2 vaccine, (4) able to speak English, and (5) able to provide written informed consent. Exclusion criteria include (1) a cognitive or other disability that inhibits smartphone use, (2) inability to participate because of medical or psychiatric conditions diagnosed by a physician or clinician, or (3) enrollment in other COVID-19 research.

RCT participants (N=60) are randomized to one of the 3 treatment groups using simple randomization in the REDCap (Research Electronic Data Capture; Vanderbilt University) program [33,34]. These groups include standard MI (SMI; n=20), intensive MI (IMI; n=20), or mHealth-based MOTIVACC (n=20). SMI consists of one MI session at least 20 minutes in length [35-38] delivered by a trained counselor at the baseline and provision of printed self-help materials. IMI consists of the SMI component plus 3 additional MI sessions delivered by a trained counselor via telephone or video calls at weeks 2, 4, and 6 post enrollment. MOTIVACC consists of an empathy-driven, fully automated, smartphone-based MI program that involves proactive, interactive, and individually tailored messages, plus some embedded images or videos, delivered by our Insight app [39]. MOTIVACC will begin immediately after enrollment and data collection and continue for an 8-week period (about 2-5 messages or questions per day, depending on participants’ interaction and responses). The intervention content in all groups is adapted from the team’s previous work and existing empirical evidence for in-person MI [25], is informed by Phase 1 outcomes, and is designed to tap SCT mechanisms. Insight for Android is used and any participants enrolled in MOTIVACC who do not have an Android phone will be loaned one for the duration of the study. All participants are followed up at week 8 post enrollment via individualized REDCap survey links sent via email [33,34].

Primary assessments (self-reported data) are conducted at baseline (after completing informed consent) and 8 weeks (after completion of the intervention if applicable). Additional data regarding the completion of the intervention contacts is recorded at the time of delivery in REDCap by study personnel (ie, for MI) or consistently and automatically in Insight (ie, for MOTIVACC). Table 1 displays the assessments and measures. The primary RCT outcomes are positive changes in the vaccine hesitancy scale [40,41] and intention to obtain the vaccine [42,43]. The secondary outcome is actual vaccination.

All clinical trial participants receive one US $30 gift card as compensation for completion of baseline and 8-week follow-up data collection.