The International Association for the Study of Pain defines pain as an “unpleasant sensory and emotional experience associated with actual or potential tissue damage [1].” As a disease, pain carries a huge global burden with a prevalence of 20% and incidence of 10%. It negatively affects one’s psychological and social functioning, thereby impinging on quality of life too. Its tangible costs cannot be understated as well, with research demonstrating a loss in productivity owing to absenteeism and diminished job performance as a result [2-7].

Acute pain is a commonly encountered clinical entity in up to 84% of patients presenting to medical services [8]. Acute pain is sudden in its onset and is typically expected to last for a short time (≤6 weeks). Usually, it can be attributed to a specific event or illness, but at times, it may be iatrogenic [9].

Either pharmacologic (analgesics) and nonpharmacologic (interventions) can be used for addressing acute pain. Within analgesics, opioids are often prescribed [10]. This strategy, although clinically effective, has several disadvantages. Opioids are notorious for their deleterious side effects, including tolerance, dependence, and hyperalgesia [11]. Additionally, some analgesics also require invasive procedures to be administered, such as with intrathecal infusions, which carry their own set of clinical risks and side effects. Similarly, concerns have been raised toward nonpharmacologic approaches (ie, transcutaneous electrical nerve stimulation, hot or cold compress), with regard to their efficacy and appropriateness in the setting of acute pain [10,11].

In light of these findings, as well as the recent recommendations from the American Pain Society and the American Society of Anaesthesiologists, there remains an urgent need to characterize and identify alternative modalities for acute pain relief. In particular, there is a need for these therapies to be clinical efficacious, be minimally invasive, and potentiate low levels or negligible side effects [12]

Virtual reality is a burgeoning technology which is in its infancy of uptake for clinical utilization. As a technology, virtual reality allows for users to be immersed in a virtual environment through multisensorial stimulation [13,14].

Over the last decade, increasing attention and research has been directed toward assessing the utility of virtual reality in managing acute pain. While there is presently no clear explanation of virtual reality’s mechanism of pain relief available, several theories which span the realms of biology and psychology exist to elucidate its efficacy [15,16]. To date, it is also worthy of mention that virtual reality therapy has been successfully used as an analgesic in several acute clinical contexts, ranging from pediatric phlebotomy to dressing changes for burns and postcardiac surgery [17-19].

The main motivation for our group in undertaking this systematic review was to provide a comprehensive literature review to inform the clinical utilization and testing of virtual reality therapy.

Primarily, this encompassed understanding the applicability of virtual reality in facilitating analgesia during acute pain for inpatient populations. This also included an understanding of virtual reality therapy headsets, content being used for the indication, and the effects of virtual reality on anxiolysis, since it has been implicated in facilitating this effect, which in turn, modulates the patient’s perception of pain [20,21]. Furthermore, we aimed to delineate the effect of virtual reality on physiological parameters (autonomic arousal); literature has suggested that these are biomarkers of pain, and thus, are theoretically correlated with pain responses [22]. Last, given the impact that it would have on patient safety and clinical uptake, we wanted to understand virtual reality therapy’s side effect profile.

Prior to undertaking this review, there was a gap in the literature on virtual reality therapy, in the setting of facilitating inpatient acute analgesia, which was of practical relevance to the clinician. This, in our opinion, appeared to be a barrier of clinical uptake which we aimed to address through this initiative by providing a holistic overview.

The following review was conducted in line with PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) guidelines. The search was undertaken on PubMed, Ovid Medline, EMBASE, and the Cochrane Database of Systematic Reviews up to and including May 2018 by VS, JS, AN, and SK and repeated up to and including January 2019 by VS, RRW, OP, and JS. The search was carried out without any limit of the years, and articles were restricted to those in English. The databases were searched independently by the aforementioned authors. Once shortlisted, full texts were ordered and read. The bibliographies of articles selected for the review were also screened for suitable additional articles to be included in this review. Inclusion in the review was selected by consensus between the screening authors.

Inclusion criteria were primary studies utilizing virtual reality in the management of acute pain in a clinical setting. Interventions were considered to be virtual reality only if they employed an audio or visual multimedia environment with which the patients could view or interact (ie, games and videos). Acute pain was defined as pain that was less than 6 weeks in duration and associated with an acute condition or medical procedure. The context was selected to make the findings relevant to the inpatient treatment of patients for clinicians. Studies with both adult and pediatric populations were suitable for inclusion in the review.

Articles that were reviews, case series, or case reports were excluded from this review. Experimental studies of a nonclinical nature (eg, pain induced via cold-pressor test) were excluded in a bid to focus on clinically relevant pain reduction which could be easily extrapolated to clinical practice.

Search terms used across all databases were (virtual reality OR vr) AND pain. Studies were then filtered manually as per the inclusion criteria for acute pain associated with acute conditions or procedures.

Data were extracted manually for analysis by VS, SK, and JA in tabular form. Due to the heterogeneity of the studied populations, variations in technologies utilized, and heterogeneity in the endpoints of the studies; pooling of data for meta-analysis was not considered appropriate. In addition, meta-analysis of data from rapidly evolving medical technologies of various generations was deemed inappropriate [23]. This is due to the lack of similarity between technologies, the impact of incremental innovation between generations of the same technology and the presence of operator dependence (from a clinician and patient perspective) on its performance [23]. As such, a narrative approach was followed for this review.

Data items of interest for the studies included: year of study, study design, sample size, clinical setting, population, nature of the intervention, control or comparison, virtual reality content type, main outcomes measures, outcome measurement tools, and technical specifications of virtual reality devices employed. Technical specifications of virtual reality devices included type of head mounted display, display utilized, weight of device, field of view, computer, video card, controller, virtual reality content type, and virtual reality content used.

The risk of bias was assessed by VS and RRW using the modified Downs and Black List [24] and scored on a scale of 1-10 as illustrated in Table 1.

The primary measures of interest were degree of analgesia afforded by virtual reality therapy, degree of anxiolysis afforded by virtual reality therapy, effect of virtual reality on physiological parameters, side effects precipitated by virtual reality, measures of pain assessment, virtual reality content type, and types of equipment utilized.

No patients were involved in the design, recruitment, or conduct of the study. There was no intention a priori that the results of this review would be disseminated to patients included in the trials of the review.

The findings of this review illustrated that there was a significant reduction of pain related to virtual reality therapy utilization in 67% of the studies (12/18). Although acknowledging limitations in interpreting these findings (see limitations below), this evidence is corroborated by the findings of other high-quality studies [40-42], supporting the use of virtual reality therapy as a nonpharmacologic adjunct in facilitating analgesia within a clinical context. An effort was also made to critically appraise the studies which failed to demonstrate any significant differences in pain (Multimedia Appendix 3). The intention here was to examine the studies for factors which may have contributed toward null findings and were considered to be limitations of the study by the authors themselves.

To the practicing clinician, these findings are of relevance as they suggest that virtual reality therapy can be considered as an inpatient adjunct for acute pain, particularly in the context of facilitating procedural analgesia (12/18). However, it is likely that this performance will vary by indication and the patient population to which it is being applied. This variability in performance, therefore, should prompt consideration toward pilot testing, as an initial step, for any specific clinical use in order to establish its appropriateness as a therapeutic modality.

To further elaborate, although the exact mechanisms behind how virtual reality facilitates analgesia are still unknown; there are several plausible theories which may explain its therapeutic effect. One school of thought suggests that virtual reality therapy enacts changes on a neurobiological level, and thereby, facilitates analgesia in a manner similar to a drug. Functional magnetic resonance imaging has been utilized to demonstrate this in experimental models [43]. During episodes of pain stimulus, areas of the neuroanatomic pain matrix (insula, anterior cingulate cortex, thalamus, primary and secondary somatosensory cortices) demonstrated increased levels of activity. When virtual reality therapy is administered to patients during these episodes, a reduction greater than 50% is observed in the activity of the pain matrix, which corresponds with a fall in patient-reported pain ratings [43]. Similarly, experimental models have also demonstrated that this analgesic effect of virtual reality therapy can be controlled in a dose-dependent fashion [44-46].

Alternatively, it has also been theorized that virtual reality enacts its functions on a psychological level through the distraction it provides. The Gate Control Theory [16] proposes that the amount of attention given to a painful stimulus affects the person’s interpretation of it. In line with this, the Multiple Resource Theory [15] also suggests that humans have a finite capacity to provide attention toward and process pain. As such, it is plausible that by rerouting or drawing these mental faculties away from the noxious stimulus, through a mechanism such as virtual reality therapy, that this would successfully attenuate the perception of pain [42,47,48].

Within the brain, the limbic system and amygdala are implicated in mediating anxiety, and this is often experienced by patients prior to a medical procedure [49,50]. While the ability to be anxious is essential for survival, increased levels of anxiety in a clinical environment can lead to worsening perceptions of pain, decreased thresholds for pain, and less cooperative patients [20,21]. As such, a rationale does exist for controlling anxiety in the context of facilitating analgesia for patients.

The findings of the review with regard to anxiolysis were equivocal, with 50% (9/18) of the studies demonstrating a significant anxiolytic effect. A recent systematic review [51] demonstrated significantly reduced anxiety scores in individuals undergoing virtual reality therapy for treatment of anxiety disorders in comparison to those of controls. Similarly, there is also some suggestion that virtual reality therapy generates positive emotions and improvements in mood which dampen preprocedural patient anxiety. Also, it is supposed that similar to its analgesic effects, anxiolytic properties occur as a result of the abovementioned psychological alterations [52-54].

There appears to be merit in further evaluating virtual reality therapy for its anxiolytic effect. Particularly, as the potential benefits of anxiolysis extend beyond the mitigation of procedural pain to include an improved patient experience [55].

Several studies have attempted to utilize changes in physiological markers or autonomic arousal as surrogate marker of analgesic effect [56]. From a biological perspective, this is not surprising since acute pain activates the sympathoadrenal fight or flight response, which in turn produces autonomic arousal effects (ie, increased respiratory rate, heart rate, blood pressure, skin sweating—galvanic skin response) [22,56-58].

In this review, 39% of studies (7/18) explored the relationship between virtual reality therapy–facilitated analgesia and its effect on a variety of physiological parameters. Although our findings somewhat suggested that heart rate correlated with pain scores [30,39], it was not possible to consistently ascribe utility toward using autonomic arousal as a surrogate marker for analgesic effect. There were several reasons for this.

First, parameters utilized and investigated across studies appeared to be heterogeneous and inconsistent, making it difficult to draw firm conclusions. Next, there was also evidence to suggest that not all physiological markers respond similarly to pain stimulus and subsequent analgesia [17,19,30,34,37-39]. Additionally, it is also known that arousal induced by pain is not static as the participant may be able to influence it either consciously or subconsciously by utilizing their own coping strategies (ie, heightened respiratory rate can be consciously altered by slowing down one’s breathing) [56,58,59]. This, therefore, will arguably impact the ability of physiological markers to be precise and consistent markers of pain. However, it is worth mentioning that the literature seems to suggest that both respiratory rate and galvanic skin response appear to be consistent markers of pain response, whereas cardiovascular changes appear to be less useful [56,58].

Considering these findings, it is safe to say that there is no firm evidence to suggest that virtual reality therapy can either affect autonomic arousal or demonstrate its analgesic properties through modulation of these parameters. Testing these parameters in a uniform and consistent manner, at least within a research context, is merited.

In this review, studies reported a low incidence (0%-8%) of adverse effects in participants utilizing virtual reality therapy; however, it should be reiterated that only 6 studies screened for side effects.

Some of the main side-effects associated with virtual reality were nausea, vomiting, eye strain, and dizziness; cumulatively referred to as cybersickness [60]. The most widely accepted theory explaining cybersickness relates to the Sensory Conflict Theory. This refers to the discrepancy which occurs between the ocular and vestibular systems when the senses do not receive the usual sensory feedback that would be expected in such a scenario. This lack of synchronization is believed to cause cybersickness [61,62].

This is of relevance to the clinician using the technology for several reasons. For one, this alludes to a vulnerable population of patients who are susceptible to these side effects and who should be excluded from its use, such as patients with vestibular abnormalities, with seizure disorders, and who experience migraines or headaches [63]. Additionally, this alludes to a number of methods that can be considered to reduce the incidence of these effects during use of the technology. Although a discussion of these is beyond the scope of this review, the following articles provide adequate reference material [64-68].

Most virtual reality interactive hardware consists of a combination of a head mounted display, built-in biaural headphones for sound, and a trackpad or joystick for manipulation or navigation of the virtual environment, to provide the user with an immersive experience [69].

In our review, all studies utilized a head mounted display for the administration of virtual reality to participants. These ranged from portable hardware, such as a helmet or piece of cardboard, to more sophisticated hardware systems, where participants were connected to an external processing unit.

A head mounted display displays content via 2 screens placed in front of the user’s eyes which are stereoscopic in nature. The images displayed are angled to provide a variation in depth perception, which is interpreted by the brain as having 3D characteristics and features. In addition, the head mounted displays track user interaction in real time, which updates the virtual content that is reflected to the user simultaneously [70]. This can be either through tracking head orientation or position of the user’s physical movements, as well as walking and jumping [69,71]. None of the studies in this review used systems capable of positional tracking. This is understandable, considering that clinical procedures, including those examined, typically necessitate controlled patient positioning.

In this study, 50% of the content was an active form of virtual reality, which entailed an element of interaction with the environment by the participant. In contrast, the remainder administered a passive form of virtual reality, where participants could only observe the content. This is worth mentioning as the available evidence suggests that the analgesia afforded by active virtual reality is significantly more than that offered by the passive form [72-74]. No study in our review, however, explicitly investigated this difference.

The findings of this review should be interpreted considering the following limitations.

First, it is important to note that the results of this review could be influenced by publication bias. Particularly when considering that Fanelli et al [75] demonstrated that approximately 90% of literature in the fields of psychiatry, psychology, and clinical medicine report positive findings. As a result, the performance of virtual reality as an analgesic could be overestimated. Second, this limitation is further exacerbated by the use of a narrative approach employing descriptive statistics, as was the case for this review. Similarly, the search strategy was restricted to health databases given the interest in clinically relevant findings. However, in doing so, we neglected technical databases such as IEEE and ACM, which might have had further data of relevance. Unfortunately, this is an issue that plagues research in the field of medical technology as it attempts to incorporate the two very separate domains of health and technology. To bridge this, we have provided information in the discussion to supplement areas where more technical knowledge may have been required. Third, as the sample sizes of the studies included in this review were generally small and based on very specific inpatient populations, the generalizability of the findings may be limited. In addition to this, a large variety of measurement tools were implemented to quantify the outcome of pain. This unfortunately precludes meta-analysis of the data, which would have otherwise been useful to quantify accurate treatment effects.

The systematic review highlighted the need for further large-scale prospective studies to be conducted in order to investigate the efficacy of virtual reality therapy in facilitating analgesia and anxiolysis. Additionally, this review also highlighted the need for investigators to screen patients for cybersickness-related side effects as part of their study design. Finally, it is also suggested that future clinical studies explore the differences between active and passive forms of virtual reality in facilitating analgesia.

This review sheds light on the efficacy of utilizing virtual reality therapy for the reduction of acute pain and procedural anxiety within an inpatient setting to hopefully offer a novel and practical perspective on the matter. Furthermore, it demonstrated a low incidence of adverse side effects in the populations being sampled. For clinical use, there appeared to be a preference for head mounted display to display virtual reality content. Although no differences between active and passive virtual reality were identified in this review, the literature appeared to suggest that active virtual reality would facilitate a higher level of analgesia in comparison to that facilitated by passive virtual reality [72-74]. Finally, although there was no evidence found to suggest an effect of virtual reality therapy on physiological parameters (autonomic arousal) as a surrogate biomarker of pain, this review also suggested merit in continuing this line of investigation in a rigorous and reproducible manner. It is hoped that this study serves to inform future trials to assess the efficacy of virtual reality in the treatment of acute pain.