HIV is a serious medical illness with significant morbidity and mortality. Advances in treatment have continued to progress with improved regimens both in terms of efficacy and tolerability and in terms of convenience in the form of combination medications in a once- or twice-daily pill known as highly active antiretroviral therapy (HAART). Adherence to HAART has been shown to be correlated with survival [1,2] and quality of life [3]. Primarily because of HAART, HIV is now widely regarded as a chronic and manageable infection [4]. Treatment efficacy, defined as a decrease in viral load (VL), is highly correlated with an almost perfect adherence rate to HAART of at least 90 to 95% [5,6]. However, about 40% of HIV-positive patients have adherence rates to HAART below 90% [7,8].

Several factors have been identified that correlate to poor medication adherence including homelessness [9-11], comorbid mental illness [12,13], and active substance use [14,15]. In addition, health literacy has been shown to be an independent predictor for HAART adherence [16,17]. A survey of Medicare patients revealed that 34% of English-speaking and 54% of Spanish-speaking respondents had inadequate or marginal health literacy [18]. Compounding to this challenge, a study found that although physicians frequently believe that they are using nontechnical terms when communicating with their patients, they actually used nontechnical language in only 12% of encounters [19]. Moreover, a study of audiotaped encounters between patients and physicians revealed that there was a discussion about the patient’s degree of understanding in only 2% of the encounters [20]. These data suggest that changes to how patients are educated about their illness may be needed to optimize medication adherence rates.

Virtual reality (VR) has recently emerged as an effective tool in several branches of medicine [21-26]. Recent advances in VR head-mounted displays (HMDs) have made their use increasingly comfortable, affordable, and immersive [27]. VR radically differs from most other media because it can create a level of immersion that produces a feeling of presence—a “sense of being there” [28]. Due to this distinct experiential quality, VR can augment learning with an immersive experience. In addition, VR has been shown to improve learning outcome gains [29-31], and, compared with educational videos, VR has been shown to increase engagement, positive emotions, and remembering of the presented information [32,33]. In fact, an increasing number of classrooms are utilizing VR technology to provide unique and effective educational experiences for their students [34,35].

In light of this trend, the aim of this study was to investigate the effects of an educational VR experience on HAART adherence in people with HIV. Specifically, people with HIV were enrolled at a county clinic to undergo an interactive VR experience that educates them about the importance of HAART adherence. The participants were asked to self-report how often they take their HAART medications on a Likert-scale. To verify the self-reported compliance, pre- and post- HIV VL and CD4⁺ cell counts were obtained from the participants’ digital chart. In addition, second pre- VL and CD4+ were obtained to detect any trends before the VR experience. Participants also completed a post experience questionnaire (PEQ; Multimedia Appendix 1) that assessed their immediate reaction to the VR experience with regard to its novelty, comfort, educational value, and perceived effectiveness.

From February to September 2018, 107 HIV-infected participants were recruited from Olive View-University of California, Los Angeles (UCLA) Medical Center’s HIV clinic in Los Angeles, California. To enroll, participants had to be diagnosed with HIV, be on HAART, be at least 18 years of age, be fluent in English or Spanish, and have no major uncorrectable problems with vision or hearing. There were no selection criteria for initial VL, CD4⁺, or HAART regimen. Participants were continuously recruited from the clinic until the study’s predetermined time for access to the clinic elapsed. The institutional review board at Olive View-UCLA Medical Center approved the study. Each participant signed a written informed consent. There were no financial incentives for enrolling in the study.

A 7-minute interactive 3-dimensional educational VR experience was created that illustrates a simplified version of the mechanisms of immune cells, HIV, and antiretroviral medications. We used a Dell Windows Mixed Reality HMD. Research assistants helped place the HMD on the participants’ head and adjusted it for optimal clarity and comfort. Participants required no training before the VR experience. The English and Spanish versions of the VR experience had different narrators but otherwise all participants had the same VR experience.

The narrator begins by introducing a virtual person with HIV named Dave and explaining that Dave is made up of cells. A magnifying glass moves to show Dave’s skin cells (Multimedia Appendix 2). The experience then virtually transports the participant into one of Dave’s arteries. Red blood cells can be seen traveling through the artery when a single rod-shaped bacterium swims nearby (Multimedia Appendix 3). For dramatic visual effect, the bacterium is seen releasing green particles representing poison into the artery. Intense music ensues when a white blood cell (WBC) chases the bacterium, which it eventually engulfs (Multimedia Appendix 4). A single HIV then enters the scene, gets inserted into the WBC, and makes many copies of itself until the WBC bursts (Multimedia Appendix 5). With no WBCs around, many bacteria swim freely and release the green poisonous particles. The artery gradually turns green, signifying that Dave is sick (Multimedia Appendix 6). The scene resets and the narrator says, “Let’s see what happens when Dave remembers to take his medications.” Hopeful music plays and the narrator encourages the participant to press a button to release “a dose of life-saving medication.” When the participant presses the button, medication in the form of gold particles enters the artery and creates a shield around the WBC (Multimedia Appendix 7). HIV again tries to attack the WBC but gets repulsed by the shield. The participant is asked to press the button several more times to give more medication as more HIV attempt to attack the WBC (Multimedia Appendix 8). In conclusion, the narrator explains that although Dave feels fine, he is required to continue taking his medications to keep the shield active.

Participants were asked to self-report how often they take their medications on a Likert-scale. This SRQ was asked just before the VR experience and again at least 2 weeks post experience. The initial SRQ was asked in a face-to-face interview, while the follow-up was asked either by phone or with another face-to-face interview. Immediately following the experience, a PEQ was administered.

VL and CD4⁺ cell counts were drawn as per standard clinic protocol. Both the treating physician who ordered the labs and the laboratory were blinded as to which patients were enrolled in the study. Post-, pre-, and second pre-VL/CD4⁺ were recorded from the participants’ digital chart. Given that blood concentrations of HIV ribonucleic acid are expected to decrease rapidly in the first 2 weeks after HAART initiation [36,37], post-VL/CD4⁺ were defined as the first labs that were obtained at least 14 days post experience, whereas pre-VL/CD4⁺ were defined as the most recent labs that were obtained before or on the same day as the experience. Finally, second pre-VL/CD4⁺ were defined as the most recent labs that were obtained before the pre-VL/CD4⁺. All identifying information was removed to protect the privacy of the participants. Demographic data obtained from each participant were age, gender, and primary spoken language.

All analyses were performed by use of IBM SPSS software, version 22. The raw VL were converted to a logarithmic scale because of the large variance in the data. Spearman correlation was performed on the initial SRQ scores and pre-VL. A 2-tailed paired t test analysis was performed to compare the initial and follow-up SRQ scores. A 2-tailed paired t test was also performed to compare pre- to post-VL, and pre- to post-CD4⁺. To assess for the possibility that the VL were decreasing even before the VR experience, a 2-tailed paired t test was performed on the second pre-VL and pre-VL. For the CD4⁺ analysis, only participants who had below normal pre-CD4⁺ (defined as 500 cells/mm³) were included. Participants who did not have pre-VL, second pre-VL, or post-VL were excluded from this analysis. Finally, the PEQ was correlated with age using Spearman correlation.

When taken together, several points of evidence strongly suggest that participants indeed were more compliant with their HAART medications because of the VR experience: (1) the SRQ scores were correlated with VL, (2) the SRQ scores improved from pre to post experience, (3) the VL decreased from pre to post experience but not from serial VL obtained before the experience, (4) the CD4⁺ cell counts improved from pre to post experience but not from serial counts obtains before the experience, and (5) almost all participants (94%) answered on the PEQ that they are more likely to take their medications because of the VR experience.

Although these preliminary results are promising, a number of limitations need to be addressed. First, the study did not randomize participants and no controls were used. Future studies should ideally use a placebo-controlled design. Second, the study was conducted only at a county clinic and therefore may not necessarily be generalizable to other settings or populations. Third, the VL and CD4⁺ counts were obtained as per the clinic’s standard of care. A superior study would control the timing of the blood draws to more thoroughly account for extraneous variables in the data. Fourth, the duration for which increased HAART adherence is maintained is unclear. A longitudinal study with more participants would allow for more robust statistical analyses. It would also be worth testing whether repeating the experience every few months can maintain improved compliance over time. Finally, collecting further demographic data on, for example, health literacy, homelessness, substance use, and comorbid mental illness could yield interesting correlations that may help advise which subpopulations would be most likely to benefit from the VR experience.

This study suggests that patient education using VR is effective for increasing HAART adherence. If the results are confirmed, VR’s effectiveness and relative low cost offers a great opportunity for clinics to implement a simple solution that may improve both the morbidity and mortality of their patients. Future studies should attempt to generalize the results to other settings, populations, and illnesses—especially those in which patients feel healthy and may not feel the need for taking their medications as scheduled. Compared with other traditional media platforms such as videos, VR is more immersive and thus has a greater positive emotional impact, which can improve engagement and learning. VR is rapidly becoming more accessible, affordable, and immersive, and more studies are needed to further explore opportunities for using this maturing technology for improving peoples’ lives.