Past, Present, and Future: Editorial on Virtual Reality Applications to Human Services

Abstract

“Virtual Reality” interventions in human services may include 360° video, augmented reality, mixed reality, and fully immersive 3-dimensional virtual reality simulations. A variety of applications have been evaluated in various fields of study, including medicine, social work, psychology, and human performance training. Currently, the state of research of VR interventions in human services has primarily focused on efficacy and effectiveness research, with few studies evaluating “scaling up” or implementation of VR interventions in larger populations. Unfortunately, the state of efficacy and effectiveness studies of VR interventions still remains weak with some applications due to smaller sample sizes, lack of randomized control trials, and a gap in reporting key intervention qualities, dosage, and outcomes. With new developments in combining artificial intelligence with VR, realism and the potential for human interaction with computer generated simulations may boost presence and immersion within these applications. This editorial provides an overview of the state of virtual reality applications in human service provision, potential gaps to be addressed by research in the future, and the development of AI based interactive sequences that may boost use presence.

Keywords:

• Virtual reality
• human services
• editorial

Introduction

Virtual Reality (VR) originated as early as the 1950s and experienced some degree of popular enthusiasm in the late 1980s and 1990s (Mandall, 2013, p. 304). Originally conceptualized as a digital space that individuals could access using computer equipment (Lanier, 1992; Rheingold, 1991), the term "VR" is often used for several technological mediums including augmented reality, mixed reality, and 360° immersive video. Defining VR as an “advanced form of human-computer interface that allows the user to interact with and become immersed in a computer-generated environment in a naturalistic fashion” (Eichenberg, 2012, p. 3), current literary efforts have expanded upon this definition to help define differences between these technological mediums. VR systems include the use of computers for rendering an environment using real-time graphics to create a computer-generated simulated world, head mounted displays for viewing the environment, and interface devices and body tracking sensors for tracking synchronicity between actual movement and movement in the virtual space (Rizzo, Schultheis, & Rothbaum, 2003; Fox, Arena, & Bailenson, 2009). VR provides a user with full immersion and presence within the environment, a platform for exploring thoughts and feelings and "intuitively transform their intentions into actions" (Eichenberg, 2012, p. 4). Presence is established upon level of immersion of experience, with VR users completely distanced from cues of the physical world, lost in a new reality established by cues within the digital environment (Fox et al., 2009). This psychological state occurs when the “the individual perceives himself or herself to be enveloped by, included in, and interacting with an environment that provides a continuous stream of stimuli” (Blascovich et al., 2002, p. 105). With full immersion in an alternative reality, social scientists have utilized VR environments to evaluate the implications of experiencing selected stimuli through computer-generated graphics to evaluate implications of exposure and to train for actual real-world experiences.

Applications for VR have been under evaluation for more than 20 years. Early applications focused on simple phobias (agoraphobia, claustrophobia, fear of flying, etc.) and have progressed to more sophisticated exposure therapies such as VR exposure therapy for Post-Traumatic Stress Disorder (PTSD; Rizzo et al., 2014) in war veterans to analgesic VR for painful procedures (Hoffman et al., 2008). VR has been used to address social anxiety and phobias through exposure to anxiety stimulating scenarios for managing and reducing anxiety symptoms (Opris et al., 2012) and to produce computer-generated addiction craving cues to develop coping for addiction with a variety of substances (Hone-Blanchet, Wensing, & Fecteau, 2014). Additional psychological applications include VR biofeedback applications as well as interactive therapy and diagnostic experiences, such as virtual counselors, virtual diagnostics, and shared group virtual spaces with other human clients. Most interactive characters have been VR based to date as they are just beginning to emerge onto VR platforms in 2019. Many of the described VR applications have been constructed and studied, often with positive results.

VR can be more efficacious for these applications because it provides a sense of ecological validity; the brain automatically interprets the immersion as being real even though it is consciously understood to be virtual. Thus, VR can be helpful for immersive training in a variety of fields. Medicine has been using VR for years for training with applications in anatomical visualization, simulated surgery, and complex medical simulations for critical care and emergency medicine (Alaker, Wynn, & Arulampalam, 2016; Pfandler, Lazarovici, Stefan, Wucherer, & Weigl, 2017). The realistic environments are good at simulating a complex setting and work well with appearing to simulate procedural and medical interventions.

Other forms of immersive technology, such as 360° video, offer an individual immersion without presence; the user cannot fully become part of the rendered space including manipulation such as engagement with movement. Thus, the level of presence and human autonomy within this technology may not provide user choices that facilitate training for behavioral modification. 3-Dimensional VR, on the other hand, involves full immersion with an active component; specifically, the ability to engage in autonomous interaction including manipulation and movement within the environment. Because 3-dimensional VR offers complete immersion within a landscape of choices and opportunities for a user to interface with their environment, it can be a tool for training human behavior. This does not mean that 360° video or other methods of experiencing the digitized world are not useful mechanisms for social science or the human service profession. For instance, one article from this special edition (see Metsis et al., 2019) found that exposure to a 360° immersive video compared to a VR environment yielded similar outcomes when addressing social anxiety in participants and could be used as a means of prototyping. However, few studies through direct comparison have identified strengths and weaknesses of the variety of immersive technologies. Given that research comparing immersive mediums is in its infancy, it is helpful to define the technology, including strengths and challenges, and assist practical users with the best technology for achieving their goal.

Social Work

Social work as a discipline of study has established a foundation of research pertaining to VR interventions within a variety of applications. The mission of the profession of social work has been defined as "to enhance human well-being and help meet the basic human needs of all people, with particular attention to the needs and empowerment of people who are vulnerable, oppressed, and living in poverty" (National Association of Social Workers, 2017). Within this definition, social work has a strong presence in the field of practice, policy, and advocacy of social services. Social workers are often designing, implementing, and evaluating the effect of social services on their intended goals, and thus remain key stakeholders in the design and implementation of social service interventions. There are multiple VR research laboratories in U.S. social work schools currently designing and studying VR social and behavioral interventions (Trahan, Smith, & Benton, 2017), indicating that social work education is aware of the usefulness of the technology. Social work is well positioned with a focus on cultural, political, and sociological factors to assist the field of VR research to move from a focus on efficacy to effectiveness and even implementation. As social work is educationally familiar with factors on a population level that impact delivery of social service intervention, the field may continue to have a large role in assisting with evaluating VR solutions to social and behavioral problems.

Social work interventions will differ because physical interactions and kinetic procedures are far less likely in typical social work interactions. VR provides specific environments which are useful for social work applications, but VR will need to provide much more in the way of rich interaction to find robust and useful applications for social work clients. Unfortunately, a barrier to social work's broader adoption of an identity around research of implementation of technology is a lack of educational curriculum and certification standards. Few social work schools have incorporated educational curriculum for students about how technology can enhance practice, policy, and advocacy. To our knowledge, no social work schools provide curriculum in virtual or augmented reality. Without a foundation of understanding of the use and role of this technology in expanded social service practice, social workers remain unequipped to face the massive proliferation of this technology that is reaching the commercial market. Many social workers do not understand the technology and have not been educated about the technology, perhaps a barrier to greater implementation in social service delivery.

To boost competence and credibility of delivering VR interventions, there remains a need for training and certification procedures for delivering services both ethically and effectively. This effort requires oversight of a defined skill set for certifying competence. As there are ethical issues related to use (see Trahan et al., 2017), only those clinicians that can identify risks, recognize limitations, and utilize technology effectively should use VR for human service interventions.

Efficacy, Effectiveness, and Implementation

The push to move from efficacy to intervention implementation research is well documented. Three stages of research, including efficacy, effectiveness, and implementation or "scaling up" have been described and delimited in literature in prevention science, for the purpose of encouraging growth toward research that disseminates and evaluates interventions within real life circumstances with attention to barriers to utilization (for more information, see Gottfredson et al., 2015). Efficacy research is designed to evaluate an intervention within an optimal treatment setting, with homogenous samples in lab environments, under controlled conditions while effectiveness research is intended to evaluate the intervention in real-life situations delivered by common intervention providers using varied samples (Flay 1986; Flay et al. 2005). Unfortunately, some VR research still suffers from problems with non-rigorous study design and reporting gaps resulting in a confusing picture of the efficacy and effectiveness of some interventions.

Deficiencies in VR research may include small sample sizes, lack of randomized control trials, and problems with evaluating the fidelity of the intervention. Unfortunately, many studies are often incomplete in adherence to reporting standards. These deficiences reduce potential for replication and meta-analysis, dilute potentiality of results, and diminish the credibility of the intervention. In a recent systematic review of interventions for alcohol and nicotine addiction (Trahan, Maynard, Smith, Farina, & Khoo, 2019), it was discovered that small sample sizes, single group and quasi-experimental study design, and problems with external and internal threats to validity rendered the systematic review of interventions with high potential for bias. The length of time required to evaluate large samples due to equipment limitations may result in slow publication turnaround time and positive findings may be more publishable, resulting in potential dissemination bias (Hopewell, Clarke, Stewart, & Tierney, 2007). We have also noticed that publications on VR studies do not often report key components of the intervention that provide us a clear understanding of the actual intervention. Lack of intervention specifications may be due to journal length requirements, and we recommend that editors consider relaxing these requirements due to the nature of reporting all dimensions of the intervention.

To evaluate these interventions with rigor, the Cochrane risk of bias tool (Higgins et al., 2011) needs to be understood and utilized to ensure high quality evaluations. Selection bias, performance bias, attrition bias, detection bias, and reporting bias must be addressed by studies in order to boost scientific objectivity. Studies should apply best practices in design, utilizing randomized control trials or randomized effect trials to reduce selection bias. Studies should address performance and detection bias by blinding participants and researchers. Attrition should be reported to determine the level of potential bias impacting outcomes. Outcomes should be construct measured and operationalized using standardized measurement with good reliability and validity. In addition to statistically significant differences, effect sizes need to be calculated and reported so that meta-analysis can evaluate the pooled results of multiple studies. Power analyses should be reported, despite potential for small samples due to equipment limitations.

Reporting results should follow American Psychological Association's Journal Article Reporting Standards (Kazak, 2018) and Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA; Liberati et al., 2009) guidelines. Furthermore, reporting standards for conflicts of interest can be especially important in developing technological interventions and should be transparent. Reporting should include the role of the researcher in delivering the intervention and attempts to maintain objectivity in this process. We, along with others, have noticed a gap in studies of dosage on clinical effect (Glegg & Levac, 2017; Trahan et al., 2019). Few studies on dosage for interventions have left the field questioning what quantity and quality is required for positive outcomes. Dosage of treatment includes length of treatment, length of exposure sessions, length of time between exposures, and total number of sessions. Qualifications for the intervention provider should be reported, as studies evaluating interventions effects may suffer from underqualified providers. The process of tailoring the intervention to an individual should also be reported. If a generalized environment is tailored to the individual, how was this decided? Manualized protocols should be created, assessed for fidelity, and readers should have opportunities to view and evaluate these protocols. It is also helpful for researchers to report the software and hardware platforms of intervention, including software delivery systems and equipment utilized during exposures such as types of immersive headsets, joysticks, and other movement operating systems. When exposures are paired with a secondary intervention, these interventions should also be clearly outlined regarding dosage, treatment process, and delivery team member(s). All measures that were collected should be reported in the outcome of the study. While these recommendations are generally standard to PRISMA guidelines for reviews (Liberati et al., 2009), we have witnessed many VR studies neglecting this information in reporting their clinical trials. Unfortunately, lack of this information provides a gap in the foundation of efficacy and effectiveness studies required to move the field to implementation studies. To reach a goal of implementation science around the dissemination of VR interventions in the broader community, we recommend that the field of VR interventions focus on first developing typical checks that are necessary to ensure high quality scientifically valid efficacy and effectiveness evaluations.

Generally, once interventions are deemed efficacious and effective, research can begin to focus on the implementation questions, including feasibility of scaling and barriers to adaptation in the larger culture. These studies are the next generation of intervention evolution, to understand how the intervention can impact the broader population and evaluate potential pitfalls of this scaling up process. Recognized as a stage of research that is still in a conceptual infancy (Spoth et al., 2013), this type of research may require more organization and community data about the level of adoption, implementation, and sustainability of the intervention (Gottfredson et al., 2015). In prevention science, one element of this potential direction of research is community willingness to adapt to the new intervention (Gottfredson et al., 2015; Spoth et al., 2013). With a lack of specificity about how to measure "community readiness," authors admit that the lack of ability to measure this construct appears to be a barrier to overcoming our understanding of the scaling process (Gottfredson et al., 2015).

Broad implementation of VR interventions appears on the horizon. VR equipment has become more commercially available in the last 5 years with over a 100 million headsets shipped in 2016 (Forbes Magazine, 2017). This renewed interest in VR is paired with the VR market projected to reach 44 billion dollars by 2024 (PRNewswire, 2019). This market surge has resulted in a considerable reduction in costs associated with equipment. With considerable cost reduction in technology, specifically VR, the proliferation of implementation studies may be just on the horizon. Generally, studies in both efficacy and effectiveness are common to evaluating VR interventions in various domains, including rehabilitation, health, mental illness, substance abuse, and training/education. However, the leap to implementation research remains in the infancy of study. While few VR interventions have successfully been implementation on a wide scale, VR exposure therapy for PTSD has probably the widest exposure, being utilized at more than 65 Veteran’s hospitals in the United States.

Barriers to implementation may reside in the history of cost for VR systems. Until 2015, due to cost, hardware, and software for using VR technology remained out of the grasp of a broader U.S. population. The need for VR display devices is certainly a barrier to adoption. State of the art systems cost upwards of $3,000, though there are alternatives such as inexpensive lens chambers for mobile phones and low-cost mixed reality headsets priced under $400. These experiences are on a variety of independent and incompatible platforms, with standardized compatibility some years off. Even with VR displays in place, there is no recognized standard for how users should interact in VR; using voice, gaze, gestures, or game controller to operate a VR scenario. Thus, using a new and expensive VR technology for social work clients, typically those without great means, results in increased accessibility in social work offices rather than client’s homes.

This trend may change with the commercialization of VR mobile phone technology. VR applications on mobile phones may result in increased delivery to larger subpopulations, resulting in greater opportunities for developing evaluation of the scaling up process. As the price of this technology continues to become more affordable and research begins identifying useful applications for VR, the profession must make efforts to better understand how this technology can be used and what barriers exist, so we do not deepen the digital divide (Dolan, 2016). This divide refers to the difference between populations with access and those that do not have the means to stay up to speed with technological advancements (Dolan, 2016). For instance, if platforms require Internet or large download capability, many individuals may not have the capability or the current PC specs to be able to run VR efforts. With nearly 90% of U.S. adults owning a cell phone (Pew Research Center, Internet Science & Technology, 2013) the proliferation of VR interventions through mobile access has the potential to bridge the gap of the digital divide, increasing accessibility to low income groups, and expanding the reach of VR intervention.

While access remains a crucial barrier to implementation, the user experience needs more attention as well. Glegg and Levac (2017) suggest the adoption of scales that assess a provider’s willingness to utilize technology as an intervention, specific evaluations of the components of environments that stimulate response, such as esthetics, types of feedback, or user history. Usability studies for heterogenous samples may provide some insight into whether adoption of the technology is feasible, while sampling providers about space for equipment and cost benefit analysis of adoption may be helpful. An understanding of the transfer of a VR skill set to everyday behavioral experience is necessary to indicate that VR interventions have real world effects, not just lab outcomes (Glegg & Levac, 2017). Furthermore, there are few studies on contraindications, resulting in a dearth of knowledge about factors that predict poor feasibility and utility. It is still not clear if using VR in some interventions is still more effective than traditional, non-VR strategies, so we must not jump into this movement hastily. Research should be conducted understanding user related issues around access, implementation, and barriers to fill the number of gaps related to the utilization of VR in social work practice. Technology does permit us the opportunity to reach people who may not attempt to target traditional methods.

With limited funding sources for VR research and expensive costs associated with developing content, obtaining funding for such expensive endeavors can be difficult. Due to the expense, VR is best utilized when immersion and presence is necessary to the experience. For example, a non-VR interactive simulation of a conversation about sexual abuse may be suitably executed on a smartphone or webpage with greater potential distribution. Placing that scenario in VR is indicated when the VR element adds something important to the experience. Thus, those considering adoption of VR must consider the compelling reason to conduct an activity in such an environment.

Artificial Intelligence and VR

Prior decades of VR research and clinical applications have focused on technologically low hanging fruit; mainly exploration of a plausible 3-D environment with severe constraints on interactive elements beyond exploratory movements and environmental changes. Agoraphobia, fear of flying, claustrophobia, and other fears can be managed by gradually progressive exposure to the fear inducing environment within VR, hence the need for interactivity is low. Other applications, such as VR exposure therapies for PTSD, sexual trauma scenarios, and alcohol/drug focused environments may rely on prescripted characters and narrative elements. Thus, to date, there has been very little artificial intelligence (AI) in VR applications.

Bringing AI technologies into VR applications opens a world of greater possibilities. With AI, VR users will have the opportunity to interact with characters within VR and have dynamic exchanges with them. This functionality requires AI for natural language understanding, computer-guided behaviors and generation of nonverbal behaviors to express emotionality. Examples where this may be useful could include rehearsal of counseling a virtual person with similar problems, learning to confront bullies, rehearsing a job interview, or the cathartic expression of trauma toward an avatar version of an abusive family member (Rizzo et al., 2014; Talbot et al., 2012). Such technology is available now for a limited set of applications yet there have been few examples of it being incorporated directly into VR, though there are no significant technical limitations to doing so today.

Computer vision is enabling AI systems to recognize objects, people, and their actions in real-world environments in a manner that has never been possible before. Similar technologies can map out 3-D spaces, allowing VR to inhabit real world spaces while maintaining consistent rules and positioning of virtual elements. Coupling this with transparent (see through) displays allows for seamless mixed reality, where we can experience virtual environments or have virtual characters sit across from us in a real-world consultation room. More sophisticated computer visions systems include multimodal sensing AI; virtual human characters will be able to perceive the emotional state and expressions of the human user and adapt their behavior accordingly. Experiments with emotional interactive virtual humans have shown promise, interacting reciprocally, matching nonverbal expression and posture, responding to vocal tone and facial expression. Additional machine learning AI used with such interactive characters have been successful in autonomously diagnosing major depression with a high degree (∼90%) of sensitivity (Stratou & Morency 2017) after a 15- to 20-min interview. Whether these AI technologies become accepted by clients really depends on their ease of use, quality of interaction and their perceived utility. As these AI systems evolve, it will be necessary to study the human factors involved around such interactions as well as learn the best practices for AI systems to interact with people. Standards regarding ethical use of AI are necessary to preserve client privacy, sense of autonomy, and safety.

With profound questions about potential of danger of AI to human relationships and the ethics of using AI (Sarangi & Sharma, 2018), social work may want to explore the proliferation of these technologies to the social world. Social work is uniquely positioned to advocate for those that are the most vulnerable in this transition to a tech driven economy. Because of the mission of social work to enhance social relationships and respect the dignity of all people, we are charged with developing ways to foster appropriate and helpful applications of these new technologies to assist in the evolution toward positive social change. We are amid a technological revolution and as a profession we must not sit by and speculate about the utility or barriers this movement may cause; we must become leaders in this endeavor ensuring our clients are not further disenfranchized and marginalized. When we avoid engaging in the research or practice with these technologies, we avoid a basic social trend, that people are growing more dependent upon technology for their social relationships. By developing positive applications for new technologies, we can shape the future generations of research and practice applications, promoting the positive applications of these technologies to assist in the empowerment of individuals and the delivery of social services.