The reasons an individual may enroll (or not enroll) in the DPP are multifactorial. First, only 15% of Americans with prediabetes are aware they have this health condition [17]. Secondly, many individuals with prediabetes lack knowledge about appropriate health behavior changes (eg, increasing physical activity and weight loss) needed to prevent T2DM [17,18]. In addition, many individuals with prediabetes have an inaccurate understanding of the risks associated with developing T2DM and its complications [19]. Finally, even individuals who are aware that they have prediabetes and understand the risks may not enroll in the DPP because of practical barriers such as the cost of enrollment, limited time for participation (22 sessions over 12 months), and difficulty with travel to and from DPP sessions [20].

Our intervention addressed or measured (for use as a covariate) each of the factors shown in Figure 1. First, to address low awareness of prediabetes, all participants were called by the clinic’s health coach, were informed that they have prediabetes, and were asked if they would like to participate in the study.

Second, to address low risk perceptions, individuals were randomized to either the VR or 360° version of our videos. The proposed mechanism of action for these videos was based on the tripartite model of risk perception [21], which divides risk perceptions into deliberative risk perceptions (ie, the individual’s estimates of the likelihood of developing a condition), affective risk perceptions (ie, the individual’s level of worry about a particular risk), and experiential risk perceptions (ie, how easy it is to imagine developing a condition). The videos were hypothesized to increase the participants’ affective and experiential risk perceptions, and this would motivate them to enroll in the DPP. Third, to address limited knowledge about prediabetes, T2DM, and the DPP, all participants were sent a link to a website [22], which includes the 3 following components: a self-assessment of risk (using the American Diabetes Association risk score, hypothesized to change deliberative risk perceptions), didactic pages about prediabetes and T2DM, and didactic pages about the DPP and its benefits.

Finally, while this pilot study did not have the resources to intervene on the practical barriers to enrolling the DPP, we measured individuals’ reports of these barriers (eg, cost and time for participation) for use as covariates when estimating the videos’ efficacy.

In this pilot project, all participants watched the same 2 immersive videos. These were developed by our research team after an extensive process of co-development with community members at risk of diabetes. The first video demonstrates how one's vision worsens over years with diabetic eye disease. The second video is a first-person narrative of an individual who progressively develops T2DM, oral health issues, and heart disease. Both videos conclude with a positive message that enrolling in the DPP may aid in preventing these potential negative outcomes. Both videos include a male or female voice-over in either Spanish or English (selected by the participant at the start of the video).

The study protocol was approved by the University of Utah Institutional Review Board prior to study initiation (approval number: 00115941). Informed consent (in English or Spanish) was obtained from all participants, via consent cover letter, prior to data collection.

This project was conducted in partnership with the Midvale Community Building Community Clinic. This clinic serves a primarily low-income Hispanic population in Midvale, Utah. This clinic offers the DPP to its patients at no cost.

Potential participants were identified in the clinic’s electronic health record by the clinic’s health coach (using hemoglobin A_1c values 5.7-6.4). They were then informed via telephone that they have prediabetes and asked if they were willing to participate in the study. During this call, individuals were screened to ensure they owned a smartphone, which was required for study participation.

Individuals who agreed to participate were contacted by the research assistant to meet in person at the clinic. During the in-person meeting, the study was explained, and the participant underwent informed consent. All study materials and the videos were provided in either English or Spanish according to the participant’s preference.

The baseline questionnaire collected information on demographics and a yes or no question on whether the individual had prediabetes (a check on whether they understood the notification from the clinic’s health coach); it also included an 18-item validated measure of risk perception [23]. After completion of the baseline questionnaire, the participants were randomized to receive either the mobile 360° video or the mobile VR experience. Individuals in the VR group were provided with a cardboard headset and headphones and watched the videos privately in the clinic conference room. Individuals in the mobile 360° video group watched the videos on their smartphone, privately in the clinic conference room. Technical issues that the participants experienced with either platform were noted for future refinement of the intervention. Prior to leaving the clinic, the participants were given a flier for the DPP offered by the clinic, which included enrollment instructions. Within 2 days of the baseline meeting, each participant was sent an SMS message with a link to the “doihaveprediabetes” website [22], an educational website that provides information on prediabetes, T2DM, and the DPP.

Two weeks later, the participants were sent a follow-up survey to their mobile phone, which included questions about whether they enrolled in the DPP, a repeat assessment of their risk perceptions, qualitative feedback about the videos (VR or 360) and informational website, as well as a standardized measure of health literacy [24]. The questionnaire concluded with a set of 8 Likert-type questions about the importance of practical barriers or facilitators to enrollment in the DPP, including the following: language the DPP is offered in; availability of childcare at the DPP; accessibility of the DPP in terms of location; time requirements and scheduling [25]; and the desire to participate in the DPP if it were delivered by internet. To compensate for participation in this trial, the participants were emailed a US $75 electronic gift card.

To test whether there were significant differences in the distributions of the participants’ demographics for completers (participants who at least provided data on DPP enrollment in the follow-up survey) vs noncompleters (participants who did not start the follow-up survey) and between completers randomized to VR vs 360° video, we used chi-squared tests for categorical variables and t test for continuous variables.

The primary outcome of interest was self-reported enrollment in the DPP. We used logistic regression to compare the effects of the 2 modalities of video delivery on the likelihood of enrollment in the DPP. This model adjusted for any baseline difference in demographics between groups, a dichotomous variable that indicated whether the individual was aware that they had prediabetes, and the participants’ rating of the importance of practical factors that affected their decision of whether to enroll as covariates.

To test whether the videos caused changes in risk perception, we first used paired t tests to determine whether there were significant within-subject changes in risk perceptions. We then used an unpaired t test to compare changes in risk perception by intervention modality. An exploratory mediation analysis was planned if there had been significant changes in risk perceptions. All analyses were conducted using statistical programming language R (R Foundation for Statistical Computing) [26].

Approximately 240 patients were contacted by the clinic’s health coach and notified that they had prediabetes. A total of 209 consented and participated in baseline assessments, and 116 participants provided complete follow-up data. The majority of the loss to follow-up occurred in the first few months of data collection because we originally compensated participants with a gift card at the end of the baseline assessment rather than after the follow-up assessment. Some data were also lost because some questions were not mandatory in the online follow-up questionnaire (both issues were addressed about half of the way through the pilot).

Table 1 provides the measured demographics of individuals who were randomized to VR vs 360° video; there were no significant differences in demographics between the groups.

A total of 116 participants provided data on DPP enrollment; overall enrollment in the DPP was 44/116 (37.9%). Enrollment among those randomized to VR was 25/56 (44.6%), while enrollment among those randomized to the 360° video was 19/60 (31.6%). To determine if the presentation modality was associated with differential enrollment rates after adjusting for relevant covariates, we created a logistic regression model with enrollment as the outcome and intervention modality, awareness of prediabetes (a check on whether they understood the notification from the clinics health coach), and participants’ ratings of the importance of factors that might affect their decision of whether or not to enroll (“which factors were important in your decision of whether or not to enroll in the DPP?” distance, time cost, etc) as covariates. Table 2 shows the results of that model.

Complete pre-post risk perception data were available for 96 people. Table 3 shows the prevideo and postvideo scores for total risk perception score, and for each component of the risk perception scale. Neither the total risk perception score nor its components changed significantly. T tests comparing the changes by intervention modality found no significant difference in risk perception changes between groups: total risk score (t=-0.6; P=.54), deliberative risk score (t=-0.6; P=.54); affective risk score (t=0.44; P=.65); and experiential risk score (t=-1.6; P=.10)

In the follow-up survey, depending upon the video modality they received, the participants were asked “What did you think of the VR/Mobile 360° video?” Many participants provided extremely short feedback such as “good” or “educational.” Textbox 1 provides a sampling of the more detailed comments that were provided, divided into those that were positive and negative in tone.

In this pilot study, we tested whether the presentation of 2 brief motivational mobile phone videos delivered via mobile 360° vs mobile virtual reality had differential effects on risk perceptions and enrollment in the DPP. We found an absolute difference in the enrollment of 13% between groups that, while not statistically significant in this project, might be practically important. We also found that risk perceptions did not differ by modality and did not change for individuals in either group. We believe these results suggest several avenues for further investigation.

While we are unaware of prior research comparing the efficacy of VR to 360° Video on changing individuals’ health beliefs and behaviors, a few studies have tested the use of VR for health behavior change [27]. For example, Ahn et al [28,29] compared the effects of a pamphlet on the health risks of sugary drink consumption alone with a VR simulation of a virtual person gaining weight as a result of regularly consuming soft drinks, with both interventions combined. They found that the combined intervention was more effective than either alone, leading to lower self-reported consumption of soft drinks. Interestingly, they found that the risk perceptions of participants who experienced the VR increased. By contrast, in this study, we found that our immersive videos did not lead to changes in risk perception. Clearly, further research is needed to understand the mechanism of action of immersive videos intended to change beliefs and behaviors.

There are several strengths to this study. First, this study addressed the pragmatic question of whether the greater immersiveness of VR is needed (vs 360° video) for persuasive videos to affect individuals’ health beliefs and behaviors. Second, the study interventions sought to isolate the effect of the mode of video delivery on enrollment by addressing other factors that might affect enrollment. We notified all individuals of their prediabetes to address low awareness of prediabetes; we also sent all participants a website URL to educate them to address their lack of understanding of prediabetes, T2DM, and the DPP. Finally, we measured contextual factors related to the decision about whether to enroll in the DPP (eg, ratings of importance of travel, distance, and cost) for use as covariates in estimating the effect of the intervention.

This study has several limitations. First, our measure of DPP enrollment was based on self-report, leading to the potential for social desirability bias in our results. In addition, we only measured whether people signed up for the DPP, not whether they engaged with the program. In future work, we plan to collect objective data from the DPP program on both enrollment and engagement. Second, this study was an uncontrolled pilot study; therefore, it is possible that simply informing people that they have prediabetes and educating them about the DPP led to their enrollment, independent of the video interventions. However, prior research on notifying individuals that they have prediabetes and educating them about the DPP has found much lower rates of enrollment than we found in this study. For example, as part of a large trial of community-based DPPs, investigators contacted 7500 community members with a letter notifying them that they have prediabetes and educating them about the DPP; they found that that only 1.7% of those who were contacted enrolled in the DPP [7]. The true magnitude of the efficacy of our videos will need to be tested in a controlled trial.

Based on the results of this pilot, we are planning a trial that will compare the efficacy of notification or education alone vs notification or education + VR videos vs notification or education + 360° on objectively measured enrollment and engagement in the DPP among individuals with prediabetes. Future work will also assess multiple potential mechanisms of action for the videos, assessing whether risk perceptions [23], narrative transportation [30], and immersion [31] are associated with the videos’ efficacy and whether the effects of the intervention are moderated by the individuals’ health literacy, numeracy, or practical barriers to enrolling.

Increasing enrollment in the evidence-based DPP is a national priority. We present a comparison of the presentation of brief motivational mobile phone videos in virtual reality vs 360° video on risk perceptions and enrollment in the DPP. Our results suggest that further work is warranted to determine the replicability of our findings in other populations, to examine the mechanism of action of the videos, and to assess any moderators of their efficacy.