In this example study, medical student participants were randomly assigned to interact with a virtual patient who was portrayed as being either obese or nonobese. By using VR, it was possible for participants to interact with patients who differed solely on the characteristic under study (i.e., weight). All other patient characteristics, communications, and behaviors were held constant. By doing so, the example study demonstrated that the patient's body size alone was sufficient to elicit biased reactions from medical students in isolation from other factors often confounded with body size (e.g., comorbid health conditions, communication styles33,34
). This is a clear example of the first advantage of using a VR platform for experimental research. That is, VR allows presentation and control of any object or characteristic, even those that are difficult or impossible to change or control in real life.3,4
In future obesity prevention and treatment research, a near endless array of clinical or interpersonal factors could be manipulated in a VR research environment. It is particularly powerful that VR enables complete control over virtual humans. For example, in addition to visible weight, it may be profitable to study the impact of a virtual patient's or clinician's gender, race, attractiveness, and so on.28,35–39
Work of this nature could illuminate social factors (such as racial or gender match between patient and provider) that influence weight-related medical encounters. In future research, it may also be useful to control nonverbal behavior of virtual humans. In the example study, nonverbal behavior was held constant between the obese and nonobese patient. However, nonverbal behaviors could instead be manipulated to study their effects.40
In doing so, we could learn more about subtle behaviors that influence the outcomes of weight-related communication between patients and providers. Beyond personal behaviors and characteristics, VR allows the creation of any desired scenario, whether or not that scenario could or would occur in reality. This capability might be especially promising for assessing the impact of potential future health care innovations on weight-related medical encounters. For example, simulating a weight-focused medical encounter that integrates genomic information can help us gauge the impact of these technologies before they become widely available.41
Also advantageous is that, within VR, hypothetical research scenarios can be situated in realistic, tangible environments. In the example study, participants interacted with an embodied virtual patient within an immersive, visually realistic clinical environment. Unlike traditional methodologies (e.g., written vignettes), study participants can interact with and react to actual experimental stimuli in environments that contain realistic cues. For this reason, VR research environments are posited to heighten the external validity of research findings.4
A related advantage of using VR is that, by the very nature of the technology, every element in a research scenario is tightly controlled. Therefore, researchers are able to create situations that isolate variables of interest and eliminate extraneous variables and confounds. Moreover, due to this tight control, every study participant can have the exact same experience. This allows near perfect replication between participants, studies, laboratories, or time points. This level of experimental control was evident in the example study in that the virtual patient behaved identically in every interaction with every study participant.
The ability of VR research platforms to provide both high levels of control and high levels of realism is one of its greatest assets.3,4
In traditional research settings,the more realistic an experimental situation is, the less control researchers typically have over the environment.42
For example, the more an environment approaches a real clinic, the less control researchers tend to have over features like event timing, interruptions, and background noise. Virtual reality simulations can overcome this dilemma.
Finally, VR environments can not only be designed to create and experimentally vary study elements, but also be constructed to serve as behavioral measures. Behavioral measures play a crucial role in obesity prevention and management research but are often difficult to design, administer, and quantify. In VR environments, it becomes a relatively simple matter to unobtrusively record and analyze nonverbal behaviors like visual gaze and interpersonal distance that occur in clinical or other interpersonal scenarios.30
In the example study, the VR system automatically collected data that allowed for objective assessment of participants' visual contact with the virtual patient (a proxy for eye contact). The study demonstrated that medical students' level of visual contact with the patient differed depending on the patient's weight. This finding is important because nonverbal behaviors like eye contact relate to patient satisfaction in medical visits.43
Furthermore, previous work has determined that gaze behavior in some VR situations can be indicative of psychological constructs like inter-personal attention and bias.38,40
In addition to embedding nonverbal behavioral assessment in interpersonal VR scenarios, researchers can also develop virtual environments whose primary purpose is to measure a particular behavior of interest. This approach can facilitate assessment of real, quantifiable behaviors in a realistic-looking environment while participants are actually situated in a controlled laboratory. Like other VR environments, those designed for use as behavioral measures can be created to reflect any desired scenario, most notably those related to dietary intake or physical activity. At the same time, because they are virtual, researchers have control over all elements of the measurement scenario and can assess behavior in precise and fine-grained ways. For example, in our group, we are currently testing the effects of a dietary intervention using a virtual environment wherein users select food from a buffet table. Researchers can thereby directly measure several aspects of participant behavior, including not only the types and amounts of foods they select, but also patterns, order, and timing of these choices. Assessment of participant behaviors in a laboratory environment can require fewer resources than behavioral assessment in analogous “real” environments. In addition, virtual scenarios, unlike real environments, are stable over time and therefore may allow for better reliability in longitudinal assessments and for more exact replication between participants and studies. Although there is currently no published work employing virtual simulations to measure obesity-related lifestyle behaviors, such measures have been successfully developed to assess other related health behaviors.44,45