Virtual Reality Analgesia for Burn Joint Flexibility: A Randomized Controlled Trial

Abstract

Objective:

We conducted a randomized controlled study to determine the effects of virtual reality (VR) distraction on pain and range of motion (ROM) in patients hospitalized for burn care during active physical therapy exercises.

Method:

Thirty-nine participants aged 15 to 66 (M = 36) years with significant burn injuries (mean burn size 14% TBSA) participated. Under therapist supervision, using a within-subjects design, participants performed unassisted active ROM exercises both with and without VR distraction in a randomized order. Therapists provided participants with instructions but did not physically assist with stretches. Maximum active ROM was measured using a goniometer. A 0 –100 Graphic Rating Scale (GRS) was used to assess the cognitive, affective, and sensory components of pain. A GRS rating of the amount of “fun” during stretching served as a measure of positive experience.

Results:

Participants reported lower mean GRS ratings during VR, relative to No VR, for worst pain, pain unpleasantness, and time spent thinking about pain. They also reported having a more positive experience during VR than during No VR. However, patients did not show greater ROM during VR.

Conclusion:

Immersive VR reduced pain during ROM exercises that were under the control of the patient.

Introduction

One of the greatest challenges in physical rehabilitation is for patients to participate in therapeutic exercises and then continue exercises on their own outside of the treatment sessions. In this respect, patients with severe burn injuries are a particularly challenging population. When severe burn injuries occur over joints, aggressive physical and/or occupational therapy is necessary. As burn injuries heal—and particularly when they scar—there is the danger that skin over joint areas will contract and lock normally flexible body parts in rigid positions. Appropriate rehabilitation exercises increase range of motion (ROM; Godleski et al., 2013) and improve functional outcomes after the wound has healed (Ehde, Patterson, & Fordyce, 1998). The treatment for such contractures is aggressive ROM; failure to maintain joint flexibility can necessitate surgery in which scarred tissue is replaced with grafts of normal skin. Although ROM therapy can prevent such negative outcomes, it is usually unpleasant and often extremely painful (Carrougher et al., 2009; Gamst-Jensen, Vedel, Lindberg-Larsen, & Egerod, 2014). Pain and anxiety are significant barriers to ROM stretching exercises, particularly in pediatric populations (Hoffman, Patterson, Carrougher, & Sharar, 2001; Schmitt et al., 2011). Unfortunately, standard of care pain medication is not usually sufficient to control the procedural pain associated with ROM stretching exercises (Hoffman et al., 2001). Moreover, high doses of opioid analgesics often have immediate side effects such as nausea, constipation, and sleepiness and can lead to dependence issues over the long term (Berger & Whistler, 2010; Cherny et al., 2001; Shang & Gan, 2003).

Psychological factors, such as patient’s anxiety and fear of pain, can increase how much pain they experience during physical therapy skin stretching exercises. Fortunately, psychological treatments can be used in addition to traditional pain medications, with little or no additional side effects. Although conventional distractions have been shown to be helpful, (Miller, Hickman, & Lemasters, 1992; Patterson, 1995), strong distractions are needed. For example, listening to music during painful procedures has been shown to reduce pain, but in one recent study, the amount of pain reduction achieved by music distraction was only one point on a 10-point scale (Hartling et al., 2013).

Immersive virtual reality (VR) represents a compelling technology which could potentially facilitate physical rehabilitation, needed for burn injuries. By providing converging evidence to several senses (e.g., visual and auditory and motor), immersive VR is well positioned to be an unusually attention-demanding technology. Patients look into a pair of goggles, and have the illusion of going inside a 3D computer-generated world, where they can use a computer input device (e.g., a computer mouse) to interact with objects they see in the virtual environment, with special effects, sound effects, and music. Research has shown that patients report large reductions in pain when they are playing VR, compared with treatment as usual (Hoffman, Patterson, & Carrougher, 2000). In laboratory fMRI brain scan studies with healthy volunteers, participants reported large reductions in pain during VR, and their brains showed large reductions in pain-related brain activity during VR (Hoffman et al., 2004, 2007). One of the early clinical applications of VR to burn care was to act as an analgesic during frequent, repetitive, and painfully invasive dressing changes (Hoff-man, Doctor, Patterson, Carrougher, & Furness, 2000). VR has since been used as a form of cognitive distraction from pain associated with other etiologies as well as burn injuries; to date there have been more than 50 published reports or reviews on the clinical effectiveness of VR distraction analgesia (see reviews by Hoffman et al., 2011; Mahrer & Gold, 2009; Scapin et al., in press; Triberti, Repetto, & Riva, 2014).

An increasing application of VR to burn rehabilitation has been to facilitate ROM exercises. VR has been shown to reduce pain during active-assisted burn therapy exercises (i.e., therapist-facilitated exercises), although it has not yet been shown to increase ROM over control conditions (Carrougher et al., 2009; Schmitt et al., 2011). An important feature of active-assisted ROM is that the amount of joint movement during therapy is controlled by the therapist.

Thus, a question that has not been addressed is what the effect of VR distraction is when the ROM exercises are under the control of the patient. A first step in answering this question can be identified in the literature on exercise adherence in the general population. The Dual Process Model provides a foundation for such research, and holds that there is interplay between cognitive processes and physiological cues from the body during exercise. At low exercise intensity cognitive processes are thought to dominate, and the affective response is typically positive. However, as exercise intensity increases beyond the level of comfort (physiological homeostasis), physiological cues dominate and affective responses become more negative (Ekkekakis, 2003, 2017). A logical step to promote exercise adherence in this situation is to either (a) distract the exerciser from discomfort, or (b) make the experience more enjoyable (Glen, Eston, Loetscher, & Parfitt, 2017). For example, music only works as a distraction if the exerciser chooses the music that is most interesting and pleasant for them. This is relevant to the study at hand because VR has been shown to be more immersive than music alone, and VR is a superior distraction for a number of reasons (Glen et al., 2017).

There is also a link between the acute affective responses to exercise and future exercise participation (American College of Sports Medicine, 2014). Specifically, the more enjoyable the experience, the more likely a person is to exercise in the future (Ekkekakis, 2017). Conversely, exercise that produces negative affect can act as a deterrent to continued exercise participation. Thus, it is important to find modalities of exercise that produce a positive affective response if we want to encourage future exercise adherence. Studies have also shown that when exercise is prescribed versus self-selected, then the affective responses are more negative (Focht et al., 2015). The circumplex model of emotional processing posits a bidirectional relationship between pain and affect (i.e., emotional state). It follows that the analgesic effects of VR distraction may be mediated—at least in part—through central mechanisms involving affect. Because the experience of having “fun” is a potential surrogate label for “positive affect,” providing a VR experience that maximally enhances the patient’s positive affect could increase the analgesic potency of VR distraction (Sharar et al., 2016).

Several factors work against a patient’s motivation to adhere to a burn rehabilitation therapy program. ROM exercise is almost always painful in joints where the skin is contracted. Burn hospitalization in general and the circumstances surrounding the burn injury are negative, therefore frequent therapy exercises can be retraumatizing (Taal & Faber, 1997). Given these factors, the ideal mode of distraction should be highly immersive to offset the already negative affect associated with the exercises. The immersive nature of VR is particularly well-suited to provide the necessary distraction to make the exercises more pleasant. ROM exercises following a burn injury must often be engaged in for months to be effective. Thus, engagement and adherence to treatment is critical to maximizing functional recovery and achieving an optimal quality of life.

Given these considerations, the current study was designed to evaluate the effects of VR distraction on pain and “fun” in patients during joint ROM that was under the patient’s control. We hypothesized that participants in VR would report lower pain as measured by ratings of pain intensity, pain unpleasantness, and time thinking about pain, relative to a No VR condition. Moreover, we hypothesized that ROM would be increased in the VR condition, and that this treatment would improve the participant’s affective experience during therapy, relative to the No VR condition.

Method

Thirty-nine individuals aged 15 to 66 (M = 36) years with significant burn injuries (mean burn size 14% TBSA) participated in this study. Seventy-nine percent of the patients were Caucasian, 13% were Asian, 5% were African American, and 3% were other. Inclusion criteria included patients hospitalized for the care of a burn injury, being at least 15 years old, requiring occupational and physical therapy for ROM exercises, and having the ability to perform those exercises without therapist assistance. The primary joints ranged were upper body, including hands, wrists, forearm, elbow, and shoulder. Patients were ineligible if they had significant facial, ear, or scalp injuries that prevented the use of VR, were under the age of 15 years, or were incapable of rating their pain intensity or completing the other study measures. Exclusion criteria also included any etiology that might interfere with decisional capacity, including a history of traumatic brain injury, significant psychiatric disorder, current delirium, psychosis, or any form of organic brain disorder. Patients were also ineligible if they were receiving prophylaxis for alcohol or drug withdrawal, did not speak or understand English, or had developmental disability, extreme susceptibility to motion sickness, or a seizure history. Because our equipment doesn’t allow for sterilization required for Clostridium difficile bacterial infection, the patients were excluded if they tested positive for this infection. Participants received the same standard analgesic medication in both treatment conditions (VR vs. No VR). Typically this included a long-acting orally administered opioid and/or a preprocedural short-acting opioid, sometimes combined with oral anxiolytic benzodiazepines. Individual medication regimens were determined by the treating physician, and the decisions about these regimens were made independently of the study protocol. Study participation was voluntary and was not compensated.

Patients who met eligibility requirements signed an informed consent form approved by the Institutional Review Board. They were then randomly assigned to one of two treatment orders, VR followed by No VR, or vice versa (see CONSORT diagram, Figure 1). The therapist chose one or two of the most painful or most troublesome joints to be ranged with exercise. Under therapist supervision, using a within-subjects design and during the same treatment session, participants performed active ROM exercises both with and without VR distraction. Using a hand that was not involved in the target ROM exercises, while looking through the VR goggles, the participants moved a computer mouse to control perspective and aim snowballs in the icy canyon virtual environment. The “left click” on the computer mouse was a trigger button to throw snowballs at the virtual objects, such as snowmen and flying fish. The participants saw the sky when they looked up, a canyon wall when they looked to the left or right, and a flowing river when they looked down. They heard background music and sound effects such as a splash when a snowball hit the river. The therapist did not physically assist with stretches. Instead they supervised the sessions, in which the participants did their own exercises with VR and without VR. In both treatment conditions, the participant moved their injured extremity through the identical sequence of exercises. The duration of the treatment was held constant for both treatment conditions. The average treatment duration was approximately 3 min per treatment condition (mean during No VR 2.8 ± 1.6 min vs. mean during VR 2.8 ± 1.7 min).

Figure 1.

CONSORT flow diagram. *We did not record the screening for all time periods. Therefore, we calculated the average number of patients screened per year using known recorded screening data. **For this study we did not collect “No Consent” information. However, according to our current orthopedic-trauma study, 48% of patients who are approached decline to participate in the study. Here, the number who declined, 29 participants, is an approximate number based on 48%. See the online article for the color version of this figure.

The participants were told that they were in a study to determine the impact of immersive VR distraction on pain and ROM during burn injury physical therapy. Before each intervention session, the therapist provided the participants with instruction on how to range their affected joint. During each intervention session the therapist supervised the exercises without any physical interactions with the participants. The participant did their own ranging.

The VR system consisted of an Alienware 17 laptop with NVIDIA GForce GTX 770M video card, Intel Core i7–4700MQ CPU @2.4GHz, 16 GB RAM, and Windows 7 64-bit operating system. During VR, the participants followed a predetermined path, “gliding” through an icy 3D virtual canyon called Snow-World (www.vrpain.com). While looking through the VR goggles, participants moved a computer mouse to control perspective and aim snowballs in the icy canyon virtual environment. They “left click” the computer mouse trigger button to throw snowballs at avatars representing snowmen, penguins, wooly mammoths, and flying fish. The participants saw the sky when they looked up, a canyon wall when they looked to the left or right, and a flowing river when they looked down. They heard background music and sound effects such as a splash when a snowball hit the river. Participants looked into a pair of “military research” quality NVIS MX90 VR goggles (http://www.nvisinc.com/) that blocked their view of the real world. The MX90 VR goggles provided 90° diagonal field of view for each of the rectangular eyepieces, with 100% overlap between the right and left eye images, minimizing eyestrain. The goggles were held in place near the participants’ eyes by a custom-made, articulated arm goggle holding system (Maani et al., 2011), with little or no contact between the patient and the VR goggles. In other words, patients did not have to wear a VR helmet on their head; instead, they could just look into the goggles suspended near their eyes. The MX90 VR goggles were unusually immersive/wide field of view, stimulating considerable peripheral vision in addition to foveal vision. They allowed participation by individuals who had burns on their heads or faces that precluded them from wearing a helmet. Although expensive, wide field of view has been shown to significantly increase analgesic effectiveness (Hoffman et al., 2006). The MX90 VR goggles used in the present research study were $36,500. Fortunately, because of recent mass production/mass marketing, the price of wide field of view VR goggles (110 degrees FOV) has dropped in 2016 to around $500 (Hoffman et al., 2014). Because all participants received stretching exercises with and without VR, study participants and providers administering interventions and assessing outcomes were not blind to treatment condition.

Outcome Measures

The two primary outcome measures were maximum joint ROM and worst pain intensity during the ROM exercise. Maximum ROM achieved by the participants without the therapist’s physical assistance was measured after each treatment condition using a goniometer. Worst pain intensity (the sensory component of pain) was measured using a 0 –100 Graphic Rating Scale (GRS) in response to the query of “worst pain” experienced during therapy. GRSs were also used to measure the two secondary pain outcome measures—”time spent thinking about pain” (cognitive component of pain) and “pain unpleasantness” (emotional component of pain). All measurements took place immediately after each treatment condition. Evidence supports the reliability and validity of GRSs for detecting changes in these pain domains over time (Gracely, McGrath, & Dubner, 1978; Jensen, 2003; Jensen & Karoly, 1992).

In addition to the GRS pain ratings, each participant provided their subjective assessments (using a similar 0 100 GRS tool) of the amount of “fun” experienced during the therapy session (0 “No fun” to 100 “Extremely fun”). This subjective outcome variable has been shown to reflect positive affect and a positive user experience during the use of VR distraction in the setting of laboratory pain (Sharar et al., 2016).

Study Design and Data Analysis

A within-subjects experimental design was used, with each participant serving as his or her own control. All tests for significance were two-tailed, with level of α =.05. To test the primary study hypothesis that patients would experience less pain during VR, we performed a single paired t test, with pain during ROM, as the dependent variable and VR condition as the independent variable. To test our primary hypothesis that patients would show greater ROM during VR, we performed a single paired t test, with maximum ROM, achieved as the dependent variable, and VR condition as the independent variable. To test the secondary study hypotheses that VR would reduce the secondary pain measures of cognitive and emotional pain, and would improve positive affect, we performed several single paired t tests with “time spent thinking about pain,” “pain unpleasantness,” and “fun” (as a surrogate for positive affect) as dependent variables, and VR condition as the independent variable.

Results

No significant practice effect was found for worst pain GRS ratings; F(1, 36) =.50, p >.05 NS, partial eta squared .014, MSE 140.15. Therefore, the GRS data were collapsed across order prior to performing the planned paired t tests. The results of the paired t test analyses indicated that participants reported lower mean worst pain ratings during VR than No VR (No VR M = 64.4 ± 28.3; VR M = 52.9 ± 27.8, t(37) 3.10, p <.005, SD= 23.58; Cohen’s d = 0.51, a medium effect size; see Figure 2). The Cohen’s d= is based on the average SD from two means. The program used to calculate Cohen’s d′ corrects for dependence between means, using Morris and DeShon’s (2002) equation 8; http://cognitiveflexibility.org/effectsize/. Twenty-one of the 39 patients reported worst pain intensity of 70 or higher during ROM with no VR (on a 0 to 100 scale).

Figure 2.

Compared with No virtual reality (VR), participants (n = 39) reported large reductions in pain during immersive VR during “active” range of motion (ROM) exercises. See the online article for the color version of this figure

A significant practice effect was found for maximum joint ROM, F(1, 36) = 24.29, p < .001, MSE = 492.64, partial eta squared = .40. Maximum ROM was significantly greater during the second treatment condition, regardless of treatment order. Specifically, participants who received VR first and No VR second were able to stretch their joint(s) further during No VR, whereas participants who received No VR first and VR second were able to stretch their joint further during VR. No significant effect of VR on maximum ROM was found, however, when collapsed across order (No VR M = 59.0 ± 44.8 degrees; VR M ± 58.9 ± 43.6 degrees), t(37) p = .94 NS.

Pain unpleasantness was also significantly lower during VR than during No VR (No VR M = 52.7 [SD = 28.8]; VR M = 29.3 [SD = 24.7], t(36) = 5.18, p < .001; SD = 27.44, Cohen’s d = 0.86, a medium effect size, see Figure 2). Time spent thinking about pain was significantly lower during VR (No VR M = 59.0 [SD = 37.3]; VR M = 24.9 [SD = 27.6], t(37) = 6.27, p < .001, SD = 33.49, Cohen’s d = 1.05, a large effect size; see Figure 2). The participants also reported having a more positive experience during VR (No VR M = 28.2 [SD = 31.7]; VR M = 63.98 [SD = 23.8]), t(37) = 6.91, p < .001, SD = 31.71; Cohen’s d = 1.15, a large effect size; see Figure 3). No practice effects were found for pain unpleasantness, F(1, 36) = .043, p < .05, partial eta squared = .001, MSE = 156.54. No significant practice effects were found for time spent thinking about pain, F(1, 36) = 3.92, p < .05, NS, partial eta squared = .098, MSE = 2037.42. A significant practice effect was found for ratings of “fun during physical therapy,” F(1, 36) = 5.417, p < .05, partial eta squared = .13, MSE = 2433.13.

Figure 3.

Immersive virtual reality (VR) increased fun scores (positive experience) compared with No VR. See the online article for the color version of this figure.

Discussion

This study was one of the few randomized controlled trials conducted on the impact of VR on ROM exercises after burn injuries. The focus of this particular trial was on unassisted active ROM; that is, ROM that is under the control of the patient, without the physical assistance of a therapist. Consistent with previous studies, patients reported less pain and a more positive experience when they received VR than when they did not. Patients reported significant reductions in pain during VR, for measures of cognitive pain (time spent thinking about pain), producing a large effect size, based on Cohen d′ calculations (Cohen, 1988). Patients reported significant reductions in the emotional component of pain (pain unpleasantness), a medium effect size, based on Cohen d′. calculations. Also patients reported significant reductions in pain intensity (ratings of worst pain), a medium effect size based on Cohen d′ calculations. Patients reported a significant increase in ratings using a surrogate measure of positive experience during ROM, a strong effect size, according to Cohen d′ However, in the current study, there was no difference in ROM during VR as opposed to during standard treatment, a finding that is consistent with studies on active-assisted ROM with VR (Carrougher et al., 2009; Hoffman et al., 2001; Schmitt et al., 2011).

The reduction of pain during active physical therapy exercises (the therapist did not physically assist with stretches) with VR was consistent with a number of previous studies that investigated active-assisted ROM administered by a therapist (Carrougher et al., 2009; Hoffman et al., 2001; Schmitt et al., 2011). In addition to being the first to examine this effect with unassisted active ROM, this was one of the largest between-subjects randomized controlled trials to demonstrate VR analgesia during ROM exercises. Thus, this study bolsters an expanding body of literature on the benefits of immersive distraction on painful wound care or therapies associated with burn care (Faber, Patterson, & Bremer, 2013; Hoffman et al., 2011; Jeffs et al., 2014; McSherry et al., 2018; Scapin et al., in press; Triberti et al., 2014). What was unknown to date was whether patients would show VR analgesic effect when they were doing the ROM exercises themselves, rather than a therapist. The fact that patients found therapy less painful and more pleasant when doing self-controlled exercise has potential implications for independence in care. If elaborate distractions such as VR can make therapies conducted by the patient less painful and more pleasant, then the possibility of increased compliance with home health exercise increases. With inpatient burn stays being shortened and the availability of outpatient therapy becoming increasingly scarce (mainly because of cost), approaches that decrease pain and increase compliance during therapies are highly desirable, particularly if they do not involve analgesic medications which can themselves cause problems.

Consistent with previous studies that examined active-assisted ROM (Carrougher et al., 2009), the active ROM that occurred during VR in this study did not differ from the control group. We anticipated that the reduced pain during exercises would translate into greater ROM. However, although patients were given instructions to range as far as they were capable, it is possible that VR prevented the patients from concentrating on the task at hand and maximizing their ROM. The null effect of VR on ROM might also be an artifact of the study using a within-subjects design and the impact of practice effects. From a clinical standpoint, patients usually report practice effect: Exercises are harder at the beginning of the session or the beginning of the day when joints are more stiff, or not yet warmed up. Greater ROM tends to be achieved as the day or session progresses. The second condition in this study always produced greater ROM as a result of joints already being warmed up in the first part of the session. In the current study, a significant practice effect occurred. The potential positive effect of doing ROM under VR was not strong enough to overcome this practice effect.

This study evaluated the effects and potential benefits of VR distraction analgesia on pain and maximum joint ROM during active ROM exercises (conducted without physical assistance from a therapist). We found that participants reported significant reductions in several aspects of pain (e.g., intensity, unpleasantness, and time spent thinking about the pain) and had a more positive experience while performing exercises during VR, relative to the No VR control condition. Thus, VR made participation in unassisted active ROM stretching less painful and more pleasant. Although not a direct test of the dual process model of exercise adherence, our findings were consistent with what this theory would predict. For exercises that are painful and produce negative affect, such as ROM over a burned joint, more immersive distraction will be more likely to overcome barriers and increase adherence. As hospital lengths of stays are getting shorter for economic reasons, patients are being required to do more therapy on their own once discharged. The notion that a positive experience with ROM therapy will increase their therapy adherence at home is worthy of further investigation.

Study Limitations

There were several limitations to this study that should be considered when evaluating the findings. The results of the pain measures relied on patient reported outcomes, and this study only evaluated the effects of VR on a single session of therapy. Also, no baseline warm-up was provided to the participants to prepare them for the ROM sessions, and this ideally should be provided to each condition in typical studies of VR. It is also possible that the order effects of the within-subjects design limited our ability to evaluate the true effect of VR on ROM; future designs where ROM can be done in VR on separate days or on separate limbs might better answer this question. Previous research has indicated that the VR analgesia can continue over a number of physical therapy sessions; (Hoffman et al., 2001) presumably, ROM improves with greater numbers of the physical therapy sessions. Another limitation was that the VR world used in the current study was not specifically designed to be used by patients during physical therapy ROM exercises. For example, the software used did not reflect a link to a body part doing ROM; it was a distraction that was independent of the exercise at hand, in the world they experienced. Although the software has been used in a number of previous studies exploring VR analgesia during physical therapy ROM exercises (e.g., Carrougher et al., 2009; Sharar et al., 2016), future research will ideally explore the use of new VR and/or augmented reality worlds that include body position tracking (e.g., the participants might see an avatar of their body moving in the virtual world). This can be accomplished through indirect, abstract, and avatar-based movement visualizations (Ferreira dos Santos et al., 2016; House et al., 2016; Kiper et al., 2018; Ortiz-Catalan et al., 2016).

The VR system used in the current study was too expensive for home use, and was a system that has been used for years preceding recent technological advancement in this area. However, with such advances, the price of the VR helmets has recently become less expensive than a new smartphone in price (e.g., now approximately $500 per helmet), and preliminary results suggest that the new generation of lower cost VR systems can also be quite effective for reducing pain of individuals with burns (Ford, Manegold, Randall, Aballay, & Duncan, 2018; Hoffman et al., 2014). VR helmets are getting less expensive, more immersive, and more portable, and this trend is likely to continue.

Conclusion

Despite the study’s limitations, the findings support the potential beneficial effects of VR for making active ROM exercises more comfortable and enjoyable for patients who need this treatment. Even if reduced pain does not translate immediately into greater ROM, if patients find exercises less painful and more enjoyable they are likely to participate more and longer in therapies, ultimately resulting in better long-term outcomes. Research to evaluate this possibility is warranted.

Impact and Implications.

With current advancements in technology, the availability and potential health applications have exploded over the past few years. The current project is one of the early randomized trials done in the area of applying VR to rehabilitation medicine, much less any form of health care. We were able to demonstrate that VR decreased pain ratings when patients were exercising a limb that had been burned in a way that affected range of motion. Although our study did not demonstrate that VR increased limb range of motion during physical therapy, the fact that the patients reported less pain and found therapy more enjoyable has potential positive implications. With greater emphasis on home health care, interventions that result in patients participating more in their own care and being more motivated to exercise are particularly important. Given the popularity and advancements in technology, the melding of it with health care is particularly compelling.