Effectiveness of Virtual Reality for Pain Management in Musculoskeletal Disorders Across Anatomical Regions: A Systematic Review and Meta‐Analysis

ABSTRACT

Introduction

The use of virtual reality (VR) in physiotherapy is expanding across various fields; however, while extensively researched in neurology, its application in musculoskeletal (MSK) disorders remains underexplored. This review aims to evaluate the effectiveness of VR in pain management across different anatomical regions.

Materials and Methods

The research was conducted using the MEDLINE (via PubMed), Cochrane Library, Scopus, Web of Science, and Embase databases, including randomized controlled trials that evaluated the effectiveness of VR interventions, encompassing immersive VR, specialised non‐immersive VR, and gaming platforms. The primary outcomes focused on pain reduction. Data were extracted from the included studies, and methodological quality was assessed using the Revised Cochrane risk‐of‐bias tool for Randomized Trials (RoB 2). A meta‐analysis was conducted using both fixed‐ and random‐effects models to synthesise the findings.

Results

From 1265 records, 27 articles met the eligibility criteria, comprising a total of 1191 participants. Of these, 20 studies contributed data to the meta‐analyses. Significant findings were observed for pain relief in the knee region (SMD = − 0.33; 95% CI: − 0.55 to − 0.10; I ² = 13%, p < 0.004), with particularly notable effects within the subgroup utilising specialised non‐immersive VR (SMD = − 0.32; 95% CI: − 0.62 to − 0.03; I ² = 10%, p < 0.003). For other anatomical regions, the heterogeneity was substantial, limiting the strength of recommendations for these areas.

Conclusions

VR shows potential for managing pain in MSK disorders, particularly knee conditions, with significant effectiveness using specialised non‐immersive VR. However, high heterogeneity across other regions limits broader recommendations.

1. Introduction

The International Association for the Study of Pain (IASP) defines pain as ‘an unpleasant sensory and emotional experience associated with actual or potential tissue damage, or described in terms of such damage’ (Raja et al. 2020). Recent shifts view pain not just as a result of tissue damage but also as a perceptual experience influenced by protective mechanisms (Moseley 2007; Wall and McMahon 1986). Pain perception is modulated by multiple factors, including somatosensory signals, and is influenced by nervous system plasticity (Stilwell 2019). This plasticity offers potential for reversing central pain sensitisation, a promising therapeutic target (Woolf 2011).

Pain is the primary reason patients seek healthcare, especially for musculoskeletal (MSK) disorders, which affect muscles, bones, joints, and connective tissues, often leading to chronic pain and functional limitations (Piano et al. 2017). MSK conditions impact a large global population, ranking fifth in disability‐adjusted life years (DALYs) and first in years lost due to disability (Liu et al. 2022). Effective management is crucial, as poor pain control can lead to kinesiophobia, maladaptive beliefs, and hinder rehabilitation adherence. Therapies should be progressive, engaging, and aligned with real‐life activities to sustain motivation and adherence (Chaplin, Karatzios, and Benaim 2023).

Given the complexity of pain in MSK disorders, virtual reality (VR) offers a promising solution. VR creates 3D simulations with visual and auditory elements, allowing users to experience a digital environment (Rutkowski et al. 2020). VR can be non‐immersive, immersive, augmented, or mixed, depending on the interaction level (Gumaa and Rehan Youssef 2019; Rutkowski et al. 2020). Based on the software used, VR could be categorised into ‘specialised’ systems—specifically developed for therapeutic use—and ‘gaming VR systems,’ which are commercial VR gaming consoles adaptable for clinical practice (Rutkowski et al. 2020).

From a mechanistic perspective, the primary factor underlying VR's efficacy in MSK pain management appears to be distraction, which involves both cognitive and affective modulation. For instance, immersive VR environments can alter individuals' responses to painful stimuli by influencing nociceptive neural signals, potentially reducing stress hormones and cortical activity associated with pain perception (Gumaa and Rehan Youssef 2019; Hadjiat and Marchand 2022; Nagpal et al. 2022). Additionally, previous studies have shown that VR can enhance cortical reorganization during rehabilitation in neurological patients (Feitosa et al. 2022). This mechanism may also benefit MSK patients by promoting cortical reorganization to restore motor function and reduce pain, while addressing psychological factors such as distress, kinesiophobia, and central sensitisation (Roy et al. 2017).

Recently, two meta‐analyses have been published on this topic. In 2023, Kantha et al. concluded that interactive VR is more effective than no rehabilitation or conventional rehabilitation in reducing pain intensity (Kantha, Lin, and Hsu 2023). In 2024, Lo et al. found that VR‐assisted active training effectively reduces back and neck pain symptoms (Lo et al. 2024). Both reviews assessed VR effectiveness based on hardware type (immersive vs. non‐immersive VR) (Kantha, Lin, and Hsu 2023; Lo et al. 2024). Earlier, in 2018, Collado et al. evaluated exergaming effects on MSK pain and found insufficient evidence to support its effectiveness, largely due to high heterogeneity (Collado et al. 2018). Other reviews, such as those by Gumaa et al. (Gumaa and Rehan Youssef 2019) and Chaplin et al. (Chaplin, Karatzios, and Benaim 2023), focused on specific body areas and included conditions such as fibromyalgia and rheumatoid arthritis, but they did not isolate pain as a primary outcome, instead assessing clinical efficacy and participant satisfaction.

To our knowledge, no studies have compared VR interventions based on software type (commercial vs. specialised). Thus, this study primarily aimed to evaluate the effectiveness of VR‐based interventions for MSK pain management, taking into account both anatomical region and software type. Additionally, we assessed the influence of VR on QoL.

2. Materials and Methods

2.1. Protocol and Registration

This study was designed as a systematic review and meta‐analysis, conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta‐Analyses (PRISMA) 2020 guidelines (Page 2021). The protocol was registered in advance in the PROSPERO database (registration number: CRD42023466977).

2.2. Literature Search and Study Selection

The search was conducted up to 31 March 2024 across the following databases: MEDLINE (via PubMed), Cochrane Library, Scopus, Web of Science, and Embase. A tailored search strategy with specific syntax was applied for each database (Supporting Information S1). Additionally, the references of relevant articles were examined to gather the maximum amount of useful and valid information.

Based on the study aim, the following PICOs were established. Participants included individuals with MSK disorders requiring pain management and rehabilitation. The intervention under investigation was VR, compared with conventional rehabilitation, defined as treatment as usual (TAU). VR interventions included immersive, non‐immersive and gaming VR delivered in inpatient, outpatient or home settings. Treatments involving telerehabilitation, teleconsultation, robotic interventions, or exoskeletons were excluded. The primary outcome of interest was pain reduction, assessed using self‐administered questionnaires and scales such as the visual analogue scale (VAS) (Price et al. 1983) and numeric rating scale (NRS) (Thong et al. 2018). Secondary outcomes included improvements in QoL, measured through self‐administered questionnaires and scales, including the Short Form Health Survey 36 (SF‐36) (Apolone and Mosconi 1998), Short Form Health Survey 12 (SF‐12) (Ware, Kosinski, and Keller 1996), and European quality of life 5 dimensions (EuroQol‐5D‐5L). We included only randomized controlled trials (RCTs) published in English or Italian, excluding all other study designs. Studies involving children, adolescents, systemic rheumatological conditions (e.g., rheumatoid arthritis, fibromyalgia), or neurological diseases were also excluded.

After completing the search, the ‘Deduplicator’ tool was used to remove duplicate records (Forbes et al. 2024). The titles and abstracts of eligible articles were then screened using Rayyan software (Qatar Computing Research Institute, Qatar). Study selection was carried out by four independent reviewers—two with experience in MSK physiotherapy (X.X. and X.X.) and two with expertise in VR (X.X. and X.X.)—who independently screened records by title and abstract according to a predefined inclusion/exclusion criteria template. A fifth reviewer (X.X.) was designated to resolve any disagreements. Subsequently, full‐text articles were obtained, and the same procedure was applied for full‐text screening and methodological quality assessment.

2.3. Data Extraction

A data extraction form was completed for each study, capturing relevant information such as authors and year of publication, number and characteristics of participants, type of intervention and training, outcome measures, and conclusions drawn by the authors. The studies were further categorised into specialised non‐immersive VR, immersive VR, and gaming VR. Additionally, studies were grouped according to the anatomical region addressed.

2.4. Methodological Quality Assessment

The studies were assessed for methodological quality using the Revised Cochrane Risk of Bias Tool (RoB2) (Sterne 2019). Five bias domains were evaluated: randomisation, deviations from intended interventions, missing outcome data, outcome measurement, and selection of reported results. Each domain was classified as ‘low,’ ‘high,’ or ‘some concerns’ based on bias risk. Finally, potential publication bias was examined through visual inspection of funnel plots.

2.5. Data Analysis

Statistical analyses and meta‐analyses were conducted by two authors (X.X. and X.X.) using RevMan 5.4 (Collaboration, T.C. Review Manager [RevMan], 5.4; The Cochrane Collaboration: London, UK, 2020). In the meta‐analyses, interventions were categorised into two outcome groups: pain (primary outcome) and QoL (secondary outcome), with assessment timepoints ranging from 2 to 12 weeks. Subgroup analyses were planned to evaluate pain outcomes associated with different VR treatments, categorised by anatomical regions targeted: upper limb (UL), neck, lower back (LB) and lower limb (LL). VR treatments were divided into specialised non‐immersive VR, immersive VR, and gaming VR. A meta‐analysis was also performed to assess QoL improvement for each anatomical region analysed, irrespective of VR type.

For studies using the same measurement tools, mean difference (MD) was used to compare treatment effects. For different tools, standardized mean difference (SMD) was applied to standardise and pool results. MD and standard deviation (SD) were estimated from standard error (SE) or interquartile range (Higgins et al. 2019). If data were unavailable, the corresponding author was contacted within 2 ‐weeks.

Statistical heterogeneity was assessed using the I² statistic with a 50% cut‐off. A fixed‐effect model was used in meta‐analyses, assuming a common effect size. If significant heterogeneity was observed, a random‐effects model was applied to account for variability across studies (Higgins et al. 2019).

3. Results

3.1. Article Selection Process

Figure 1 illustrates the PRISMA flow diagram. Our search yielded 1265 results from five electronic databases. After removing 490 duplicates, 775 records remained for screening. Of these, 676 records were excluded from unrelated topics, leaving 96 full‐text articles to be assessed for eligibility. The list of excluded studies is provided in the Supplementary Section (Supporting Information S2). Following full‐text screening, 27 studies met the inclusion criteria for qualitative analysis, and 20 studies were eligible for quantitative analysis. Seven studies were excluded: three due to inconsistencies in outcome data points (Fuchs et al. 2022; Jin et al. 2018; Matheve, Bogaerts, and Timmermans 2020), two that focused on ankle and foot impairments (Punt et al. 2016; Rougereau et al. 2023), and two where the authors did not provide the necessary data (Ebrahimi et al. 2021; Fung et al. 2012). A total of 1191 participants were included in the trials, with 610 patients in VR treatment groups and 581 in TAU groups.

FIGURE 1.

Flow diagram for the study selection process.

3.2. Qualitative Synthesis if the Included Studies

The 27 included studies (Afzal et al. 2022; Cetin, Kose, and Oge 2022; Ebrahimi et al. 2021; Eccleston et al. 2022; Fuchs et al. 2022; Fung et al. 2012; Gianola et al. 2020; Janhunen et al. 2023; Jin et al. 2018; Karakoc et al. 2019; Lin, Lee, and Hsieh 2020; Matheve, Bogaerts, and Timmermans 2020; Menek et al. 2022; Monteiro et al. 2015; Nambi et al. 2021; Naqvi et al. 2022; Nusser et al. 2021; Ozlu et al. 2023; Pekyavas and Ergun 2017; Pournajaf et al. 2022; Punt et al. 2016; Rizzato et al. 2023; Rougereau et al. 2023; Stamm et al. 2022; Tejera et al. 2020; Yelvar et al. 2017) were published between 2012 and 2023 and are summarised in the Tables 1, 2, 3, 4, 5. Eleven studies focused on the use of specialised non‐immersive VR (Ebrahimi et al. 2021; Gianola et al. 2020; Janhunen et al. 2023; Matheve, Bogaerts, and Timmermans 2020; Menek et al. 2022; Nambi et al. 2021; Nusser et al. 2021; Pournajaf et al. 2022; Rizzato et al. 2023; Yelvar et al. 2017; Yu et al. 2023), 10 studies (Cetin, Kose, and Oge 2022; Eccleston et al. 2022; Fuchs et al. 2022; Jin et al. 2018; Naqvi et al. 2022; Nusser et al. 2021; Ozlu et al. 2023; Rougereau et al. 2023; Stamm et al. 2022; Tejera et al. 2020) involved the use of immersive VR and 6 studies (Afzal et al. 2022; Fung et al. 2012; Karakoc et al. 2019; Lin, Lee, and Hsieh 2020; Pekyavas and Ergun 2017; Punt et al. 2016) used gaming VR systems. Among the 27 included studies, 4 targeted UL impairment, 3 focused on neck impairment, 7 were related to LB impairment, 13 concerned LL impairments (11 of them targeted knees impairments and 2 of them focused on the ankle and foot complex).

TABLE 1.

Characteristics of included studies: upper limb.

Author (Year)	Study design	Aim	Group characteristics	Rehabilitation intervention	VR technology	Outcome measures	Timepoints	Conclusions
Menek et al. (2022)	RCT.	To define the effectiveness of video‐based game exercises and structured closed kinetic chain exercises in individuals with rotator cuff rupture.	Individuals with partial rupture of the supraspinatus muscle. 3 groups: CG: n = 15 Age = 52 (± 6.3) CKCEG:n = 15 Age = 50 (± 5.5) VGEG: n = 15 Age = 47 (± 7.1)	2 times/week, for 6 weeks. Each session lasts 40 min, under the supervision of a PT. CG: Conventional treatment. CKCEG: Closed kinetic chain and proprioceptive exercises. VGEG: A video‐based game exercise programme.	Non immersive specialised VR. Fizyosoft (Fizyosoft extremity ROM and Fizyosoft balance system).	VAS. Pain threshold (algometer). ROM. Shoulder approximation force using Fizyosoft balance system. DASH. RCQOL.	T0 = baseline. T1 = 6 weeks.	Improvements in pain threshold, ROM in shoulder FF and ABD, DASH score, and all parameters of the RCQOL questionnaire in VGEG were statistically more significant than CKCEG.
Naqvi et al. (2022)	RCT.	To define the effectiveness of gamification in individuals post‐DRF.	Individuals post‐DRF Age = 18–65 EG: n = 10 CG: n = 10	5 times/week, for 4 weeks. Each session lasts 60 min, under the supervision of a PT. EG: Exercise with immersive‐VR. CG: Conventional rehabilitation	Immersive VR Oculus quest head‐mounted display, (Oculus, USA).	DASH VAS ROM (goniometer) Grip force (Jamar dynamometer)	T0 = baseline. T1 = 2 weeks. T2 = 4 weeks.	Gamification appears to have a significant impact on post‐DRF rehabilitation in terms of pain, ROM, grip strength, and functional independence.
Pekyavas and Ergun (2017)		To compare the short‐term effects of home exercise programme and virtual reality exergaming in patients with SAIS.	Patients with SAIS EG: n = 15 Age = 40.33 (± 13.20) CG: n = 15 Age = 40.60 (± 11.77)	2 times/week, for 6 weeks. Each session lasts 45 min. EG: Supervised VR exergaming programme Exercise training includes bilateral shoulder elevation, boxing, bowling, and tennis games accompanied by avatar. CG: home exercise programme.	Gaming VR. Wii Nintendo.	VAS (rest, activity and night pain). SPADI. Nere and Hawkins tests. SRT. SAT. LSST.	T0 = baseline. T1 = 6 weeks. T2 = 1 month follow‐up.	Intensity of pain was significantly decreased in both groups with the treatment.The EG had significantly better results for all Nere test, SRT and SAT than the CG.
Rizzato et al. (2023)	RCT.	To evaluate the effectiveness and the patients' engagement in using a digital therapy gaming system for in shoulder rehabilitation.	Individuals with shoulder pathologies (impingement syndrome, capsulitis, tendon injuries, degenerative joint or tendon pathologies). EG: n = 11 Age = 59.9 (± 10.2) CG: n = 11 Age = 62.0 (± 10.9)	5 times/week, for 2 weeks. Each session lasts 40 min, under the supervision of a PT. EG: 20 min manual treatment +20 min digital therapy. The digital therapy allowed the subject to complete the rehabilitation programme while playing games and receiving real‐time visual feedback. CG: 20 min manual treatment + 20 min conventional therapy.	Non immersive VR. Device or software name not definite.	ROM and velocity (using software GykoRePower). MVC test (Fmax in push down a ball) maximal isometric force. PENN shoulder score. PACES. Self‐efficacy. Attitude to train at home. Intention to train at home. SUS.	T0 = baseline. T1 = 2 weeks.	The investigated digital therapy resulted as effective as an equivalent non‐digital therapy in shoulder rehabilitation. Promising results in possible patient's exercise engagement at home after the rehabilitation in the medical centre.

Abbreviations: ABD = abduction, ACROM = active cervical range of motion, CG = control group, CKCEG = Closed Kinetic Chain Exercise Group, DASH = Disabilities Of The Arm, Shoulder And Hand Score, DRF = Distal Radius Fracture, EG = experimental group, EQ‐VAS = EuroQol Visual Analogue Scale, ER = external rotation, F = female, FF = forward flexion, Fmax = maximal isometric force, HRT = Head Repositioning Test, HTT = Head To Target Test, LSST = Lateral Scapular Slide Test, M = male, MVC = maximum voluntary contraction, NDI = Neck Disability Index, n = number of participants, NRS = Numerical Rating Scale, OSS = Oxford Shoulder Score, PACES = Italian physical activity enjoyment scale, PT = Physiotherapist, QoL = Quality of Life, RCQOL = Rotator Cuff Quality of Life, RCT = Randomised Controlled Trial, ROM = Range Of Motion, SAIS = Subacromial Impingement Syndrome, SAT = Scapular Assistance Test, SMG = Sensorimotor Group, SPADI = Shoulder Pain and Disability Index, SRT = Scapular Retraction Test, Sub.Imp. = subacromial impingement, SUS = System Usability Scale, Tend. = Tendinopathy, VGEG = Video‐based Game Exercise Group, VR = Virtual Reality.

TABLE 2.

Characteristics of included studies: neck.

Author (Year)	Study design	Aim	Group characteristics	Rehabilitation intervention	VR technology	Outcome measures	Time points	Conclusions
Cetin, Kose, and Oge (2022)	RCT.	To compare the effects of VR and MC exercises in patients CNP.	41 individuals with a minimum of 6 months of neck pain, a baseline NDI score of at least 20% (10 points), and the neck region as the primary pain area. EG: n = 17 Age = 40.0 (± 11.88) CG: n = 17 Age = 41.9 (± 10.76)	Both groups 3 sessions of training per week for 6 weeks, total of 18 sessions. Each session lasted 40 min in both groups. CG: Performed only MC exercises for 40 min (10 repetitions for each exercise). EG: First performed MC exercises for 20 min and then VR for 20 min (5 repetitions for each exercise).	Immersive VR. Oculus Go VR glasses.	Range of motion. JPSE. VAS. SF‐36. Hands. Muscle performance by digital hand dynamometer.	T0 = baseline. T1 = 6 weeks.	VR can be applied for improving proprioception and for decreasing cervical articular pain in CNP patients. In addition, VR may be more effective for decreasing functional limitations in patients.
Nusser et al. (2021)	RCT.	To evaluate the effects of neck‐specific sensorimotor training using a VR device, in comparison with standard rehabilitation programmes, both with and without general sensorimotor training, in patients with non‐traumatic CNP.	Patients with non‐traumatic CNP. EC: n = 17 Age = 51.2 (± 8.8) SMG: n = 16 Age = 53.1 (± 5.7) CG: n = 18 Age = 49.8 (± 8.1)	3 weeks (times for week not specified) EC: Conventional physiotherapy + VR training (120 min). SMG: Conventional physiotherapy + general sensorimotor training (120 min). CG: Conventional physiotherapy (both individual and group).	Non immersive specialised VR. The patients wear an helmet with an integrated monitor (virtual reality head‐mounted display 5DT HMD 800–26 2D, League city, Texas, USA). A 3Space Fastrak system (Polhemus Inc., Colchester, VT, USA) was used for head‐movement tracking.	ACROM. HRT. HTT. NRS. NDI.	T0 = baseline. T1 = 3 weeks.	VR training may increase the effects of a standard rehabilitation programme for patients with non‐traumatic chronic neck pain, especially active cervical range of motion in extension and in relief of headaches.
Tejera et al. (2020)	RCT.	To compare the effects of VR versus exercise treatment on pain intensity, conditioned pain modulation temporal summation and functional and somatosensory outcomes in patients with non‐specific NS‐CNP.	42 individuals with NS‐CNP EG: n = 22 Age = 32.72 (± 11.63) CG: n = 22 Age = 26.68 (± 9.21)	2 times/week, for 4 weeks in both groups. EG: 3 × 10 reps of every exercise, with a 30 rest between exercises. Mobile applications installed, controlled by neck tilting movement; CG: 3 × 10 reps of every exercise, with a 30 rest between exercises. The researcher provided the necessary verbal corrections for the proper execution of the exercises, using the same verbal commands for all participant.	Immersive VR. VR Vox play glasses were used with an HMD clamping system to which a smartphone (LG Q6) was attached.	VAS. CPM. NDI. PCS. Pain‐related fear of movement/(Re)injury. PASS‐20.	T0 = baseline; T1 = immediately after Intervention. T2 = 1 month after intervention. T3 = 3 months after intervention.	VR was not superior to exercise in improving pain intensity, CPM, ROM, neck disability, pain catastrophizing, fear‐avoidance beliefs, PPT or anxiety in patients with NS‐CNP. Kinesiophobia was the only outcome that showed differences between EG and CG at 3 months.

Abbreviations: ACROM = Cervical Range Of Motion, CBP = Chronic Back Pain, CCT = Controlled Clinical Trial, CG = Control Group, CNP = Chronic Neck Pain, CPM = Condition Pain Modulation, EG = Experimental Group, HADS = Hospital Anxiety And Depression Scale, HRT = Head Repositioning Test, HTT = Head to Target Test, JPSE = Joint Position Sense Error, MC = Motor Control, N = Number Of Participants, NDI = Neck Disability Index, NRS = Numeric Rating Scale, NS = Non specific, NS = Non Specific, PASS 20 = Pain Anxiety Symptoms Scale 20‐item version, PCS =  Pain Catastrophizing Scale, PPT = Pain pressure threshold, RCT = Randomised Controlled Trial, SF‐36 = Short Form‐36, SMG = Sensorimotor Group, VAS = Visual Analogue Scale, VR = Virtual, Reality.

TABLE 3.

Characteristics of included studies: low back.

Author (Year)	Study design	Aim	Group characteristics	Rehabilitation intervention	VR technology	Outcome measures	Time points	Conclusions
Afzal et al. (2022)	RTC.	To compare the effects of VR exercises and routine physical therapy on pain and functional disability in patients with CLBP.	84 individuals with chronic non‐radiating low back pain. Age = 25–50 CG: n = 42 EG: n = 42	Both groups received these interventions on alternative days, with 3 sessions per week, totalling 12 sessions. CG: Routine physical Therapy (Hamstring stretching). Back strengthening exercises, including: 10 repetitions of bridging. Prone leg raises. Trunk extension in prone with arms behind the back. Trunk rotation exercises. Knee to chest. Prone position with a diagonal elevation of the arm and the leg. EG: VR exercises + routine physical therapy. VR treatment involved the use of kinetic exergames such as the body ball game and reflex ridge.	Specialised non immersive VR. Kinetic device (model V.2).	VAS. MODS.	T0 = baseline. T1 = 4th session. T2 = 8th session. T3 = 12th session.	The main conclusion of was that VR exercises, when combined with routine physical therapy, had a dominant effect on reducing functional disability and low‐back pain.
Yelvar et al. (2016)	RCT.	To investigate short term effect of the VR on pain, function, and kinesiophobia in patients with subacute and chronic non specific LBP.	46 individuals with diagnosis of subacute and chronic LBP by a physician and patients who have fear of avoidance. CG: n = 22 Age = 52.8 (± 11.5) EG: n = 22 Age = 46.3 (± 3.4)	All participants were treated five times a week for 2 weeks by physiotherapists at the clinics. CG: Physical therapy was performed using physical agent modalities that include a hot pack (15 min), TENS (15 min), deep heat with ultrasound (5 min) and therapeutic exercises. EG: Passively watch the same virtual walking video clip in a sitting position. The patients were asked to imagine as if they were actually walking, while they were watching the video clip.	Specialised non immersive VR. Virtual walking video clip in a sitting position (Vita digital productions, NC, USA) played via an iPod (Apple Inc., CA, USA) with video glasses (Wrap920, Vuzix Coorporation, NY, USA)	VAS. TAMPA. ODI. NHP profile. TUG. 6MWT.	T0 = baseline. T1 = 2 weeks.	Virtual walking integrated physiotherapy reduces pain and kinesiophobia, and improved function in patients with subacute and chronic non‐specific low‐back pain in short term.
Eccleston et al. (2022)	RCT.	To compare the active VR intervention with a sham placebo comparator and a standard care group in adults with LBP.	42 individuals with low back pain for at least 3 months, with average pain intensity of> 4/10 over the past week on a 0 to 10 numeric rating scale, ODI of 26%, and medium (34–41) or high (42–68) T SK score. EG: n = 14 Age = 55.14 (± 10.53) CG: n = 17 Age = 52.76 (± 11.19) STG: n = 11 Age = 57.09 (± 8.34)	Both the intervention group and sham placebo were designed to last 15–60 min for session. There were 5 sessions scheduled for each week, resulting in 30 unique days that were designed to be delivered over 6–8 weeks. EG: Participants always enter the virtual world via the inside space. When in the cabin, they encounter a virtual (disembodied) mentor. The mentor welcomed participants, helped them navigate the space, gave instructions on tasks, presented the psychological content inviting reflection, and set homework in the real (non‐VR) world. The mentor invited the participant to the outside space where they engaged in a gamified ‘fruit‐picking’ activity. The task involved picking or stacking geometrically shaped fruits. Fruits from trees appeared in all 4 quadrants of their peripersonal space. SG: Participants in the sham placebo comparator received the same VR headset as the participants in the DTxP arm. The virtual environment showed a similar seashore environment to the DTxP arm. None of the content provided in the DTxP was presented. Instead, the participants viewed text that instructed them to relax and enjoy the environment as used in the DTxP. CG: No intervention, instructions, or education. During the treatment phase, participants continued with their standard care as usual for 6–8 weeks	Immersive VR. Both the DTxP and sham placebo were built and tested using unity v2019.3.7f1 and delivered on an off‐the‐shelf OCULUS.	NRS. ODI. PROMIS. TSK. EuroQoL‐5D‐5L.	T0 = baseline. T1 = 6–8 weeks. T2 = 9 weeks. T3 = 5 months post randomisation.	The results add to the emerging field of VR interventions for chronic pain management and rehabilitation. Replication is necessary. Future invention could usefully develop personalised, embodied nonhuman agents to guide behaviour change, make more use of the immersive properties of the environment, and further automate that content to allow for scalable use.
Matheve, Bogaerts, and Timmermans (2020)	RCT.	To investigated whether performing these exercises in a non immersive VR environment had an influence on the pain intensity and the time spent thinking of pain during the exercises, when compared to a control group who performed the same exercises without VR distraction.	84 individuals a diagnosis of chronic non‐specific low back pain (> 3 months, ≥ 3 days/week) a baseline pain score between 3 and 8 on a 0 to 10 NPRS and the ability to perform pelvic tilt exercises in a standing position. Age = 18–65 EG: n = 42 Age = 42.1 (± 11.5) CG: n = 42 Age = 44.2 (± 11.9)	A single‐session intervention, which consisted of 2 × 2 min of pelvic tilt exercises in the sagittal plane, with 30 s of rest in between. CG: Performed pelvic tilts in the sagittal plane according to a beep tone. The first time participants heard the tone, they had to their pelvis anteriorly and keep it in an anteriorly tilted position until the next beep, after which participants tilted the pelvis posteriorly, and so on. Participants had to tilt their pelvis 46 times during the first aminutes, and 54 times during the second 2 minutes. EG: Played 2 different games (2 min each), which had to be controlled by pelvic tilts in the sagittal plane.	Specialised non immersive VR Valedo®Pro, Hocoma, Switzerland	NPR. RMDQ. PCS. TSK.	T0 = baseline. T1 = after single exercise session.	Provides evidence for the effectiveness of VR distraction to reduce the pain intensity during exercises for patients with CLBP. Pain‐related fear, pain catastrophizing and baseline pain intensity did not moderate the hypoalgesic effects of VR distraction.
Monteiro et al. (2015)	RCT.	To verify the effect of exercises with Nintendo Wii on CLBP, functional capacity and mood of elderly women.	34 women with non specific CLBP Age = 68 (± 4) EG: n = 16 CG: n = 14	Training lasted 8 weeks and sessions were performed three times week. CG: Did strength exercises and core training. EG: Received the same intervention that CG plus 30 min of virtual physical training (eight exercises) using Nintendo Wii‐motion and Wii balance Board.	Gaming VR. Wii Nintendo.	NRS. Balance (Wii balance board). Sit to stand test. POMS.	T0 = baseline. T1 = 8 weeks.	Physical exercises with Nintendo Wii Fit plus in addition with strength and core training was effective only for sitting capacity (i.e., functional capacity) in elderly women with CLBP. Both types of interventions were effective to decrease pain but did not have effects on balance and mood.
Nambi et al. (2020)	RCT.	To find the short‐term psychological and hormonal effects of VR training on CLBP in American soccer players.	54 individuals with chronic (≥ 3 months) LBP, and 4 to 8 pain intensity on a VAS. Age = 18–25 CG: n = 18 Age = 21.9 (1.8) EG: n = 18 Age = 22.3 (± 1.6) VR + PTG n = 18 Age = 21.4 (± 1.8)	EG: Received VR balance training 30 min in each session for 5 days a week for 4 weeks. CG: Focused on conventional balance training for core muscles The training includes active isotonic and isometric exercise for abdominal muscles (internal oblique, external oblique, transverse abdominus, and rectus abdominus), deep abdominal muscles (psoas major, psoas minor, iliacus, and quadratus lumborum), and back muscles (erector spinae, transverses spinalis, interspinalis, and intertransverse). They performed these exercises 10 to 15 repetitions per day for 5 days per week for 4 weeks. VR + PTG: VR and physical rehabilitation The exercises performed were wall squat, Russian twist, leg lift, plank saw, cobra, and hip raise on the swiss ball. 15 times per set for 3 sets, 5 times per week for 4 weeks	Specialised non immersive VR. Pro‐Kin system PK 252 N (pelvic module balance trunk) MF (TecnoBody, Lanusei, Italy).	VAS. TSK‐17.	T0 = baseline. T1 = 4 weeks. T2 = 6 months.	Training through virtual reality is an effective treatment programme when compared with conventional exercise training programs from a psychological and hormonal analysis perspective in American soccer players with CLBP.
Stamm et al. (2022)	RCT.	To examine the preliminary effectiveness of a VR multimodal therapy for older adults with CBP in a laboratory setting over a period of 4 weeks.	22 individuals with CBP for longer than 6 months. EG: n = 11 Age = 75 (± 5.80) CG: n = 11 Age = 75.5 (± 4.39)	Both groups for 4 weeks, with three sessions per week lasting approximately 30 min EG: Performed a multimodal pain therapy in VR (movement therapy and psychoeducation) in a laboratory setting. CG: Completed a 4‐week conventional multimodal pain therapy (movement therapy as seated exercises and psychoeducation in a group setting)	Immersive VR. VR HMD headset using the ViRST VR application.	NRS. CPGQ. Ffb‐H‐R. TSK‐11. SF‐12. TUI. UEQ.	T0 = baseline. T1 = 4 weeks.	The results showed that a pain intensity reduction can be achieved with the current VR application.

Abbreviations: 6MWT = 6‐Minute Walk Test, CBP = Chronic Back Pain, CCT = Controlled Clinical Trial, CG = Control Group, CLBP = Chronic Low Back Pain, CPAQ‐8 = Chronic Pain Acceptance Questionnaire, CPGQ=Pain Grade Questionnaire, DVPRS = Veterans Pain Rating Scale, EG = Experimental Group, Euroqol‐5D‐5L = European Quality Of Life 5 Dimension‐5 Level Scale, Ffb‐H‐R = Hannover Functional Ability Questionnaire For Measuring Back Pain‐Related Disability, HADS = Hospital Anxiety And Depression Scale, LBP = Low Back Pain, MC = Motor Control, MODS = Modified Oswestry Disability Index, N = Number Of Participants, NHP = Nottingham Health Profile, NPRS = Numeric Pain Rating Scale, NRS = Numeric Rating Scale, ODI = Oswestry Disability Index, PCS = Pain Catastrophizing Scale, PGIC = Patient's Global Impression Of Change, POMS = Profile Of Mood States, PROMIS = Physical Function And Sleep Disturbance, PSEQ‐2 = Pain Self‐Efficacy Questionnaire, RCT = Randomised Controlled Trial, RMDQ = Roland Morris Disability Questionnaire, SD = Standard Deviation, SF‐12 = Short Form 12, SF‐36 = Short Form‐36, SUS = System Usability Scale, TSK‐11 = Tampa Scale Of Kinesiophobia, TUG = Time Up And Go, TUI = Technology Usage Inventory, UEQ = User Experience Questionnaire, VAS = Visual Analogue Scale, VR = Virtual Reality.

TABLE 4.

Characteristics of included studies: knee.

Author (Year)	Study design	Aim	Group characteristics	Rehabilitation intervention	VR technology	Outcome measures	Timepoints	Conclusions
Ebrahimi et al. (2021)	RCT.	To improve the performance of patients with PFP by using VR and to investigate the effectiveness of this method on brain function alterations.	26 individuals with PFP. EG: n = 13 Age = 29.69 (± 5.69) CG: n = 13 Age = 31.76 (± 5.52)	CG: No specific interventions were done. Participants continued their usual and normal lifestyle. They received training education to manage daily programs and activities related to PFP. EG: In addition to the education provided to the control group, the intervention group received 24 sessions of virtual reality training over 8 weeks. According to each patient's ability and progress. Each training session lasted about 40 min, consisting of a 5‐min warm‐up before the game, 30 min of practice with the XBOX Kinect.	Gaming VR. XBOX Kinect™ 360.	mSEBT. VAS. SF‐36. EEG.	T0 = baseline before the intervention started. T1 = after the 8‐week intervention period, following the completion of the 24 VR training sessions. T2 = 1 month after the end of the intervention for all outcome measures, (except for EEG, which was evaluated only at pre‐intervention and post‐intervention stages).	This study demonstrated that long term VR was capable of improving both clinical impairments and brain function in patients with PFP.
Fuchs et al. (2022)	RCT.	To investigate the impact of virtual reality interventions in the immediate post‐operative period following TKA. The study aimed to assess how this intervention affected pain and anxiety levels, as well as long‐term function, compared to conventional physiotherapy.	55 individuals immediate post‐operative period following TKA. EG: n = 30 Age = 70 (± 7) CG: n = 25 Age = 70 (± 7)	EG: Received 15 min of virtual reality (VR) intervention. The intervention was combined with CPM. The VR intervention and CPM were administered three times a day for 10 days. CG: Received 15 min of CPM. This was followed by conventional physiotherapy. The CPM and conventional physiotherapy regimen were similar in terms of duration and frequency to that of the study group.	Immersive VR. Samsung Gear VR.	STAI. VAS. WOMAC.	T0 = baseline assessments were made before the intervention, at the time of admission for the TKA. T1 = day 1 post‐operative. T2 = day 2 post‐operative. T3 = 6 months post‐operatively.	Virtual reality intervention in the immediate post‐operative period following total knee arthroplasty decrease pain and anxiety but did not influence the pain, anxiety, and long‐term function results more than conventional physiotherapy.
Fung et al. (2012)	RCT.	To determine whether Nintendo Wii Fit TM is an acceptable adjunct to physiotherapy treatment in the rehabilitation of balance, lower extremity movement, strength and function in outpatients following total knee replacement.	47 individuals with TKR. EG: n = 27 Age = 67.9 (± 9.5) CG: n = 23 Age = 68.2 (± 12.8)	Both control and study interventions were provided in addition to and following each regularly scheduled 60 min physiotherapy session, in a separate treatment area. CG: Received 15 min of lower extremity exercise EG: Received 15 min of Wii Fit gaming activity.	Gaming VR. Wii Fit games con balance board. Nintendo.	ROM. 2MWT. NPRS. LEFS. ABCS. LOR.	Each participant was assessed on the first study visit and every 2 weeks thereafter until they were discharged from physiotherapy services by their treating clinician.	Wii Fit is potentially acceptable as an adjunct to physiotherapy intervention for outpatients following total knee replacement, provided the games chosen challenge balance and postural control, and use the lower extremities.
Gianola et al. (2020)	RCT.	To compare the effectiveness of early rehabilitation using VR versus traditional rehabilitation in improving functional outcomes after primary TKA.	84 individuals with TKA EG: n = 44 Age = 66.6 years (± 8.7) CG: n = 41 Age = 70.7 (± 8.5)	Both groups underwent rehabilitation sessions with a commitment of 60 min per day. For both groups was until discharge (10 days after the surgical procedure.) Daily sessions until discharge. Both groups performed passive knee motion exercises on a continuous passive motion device for the knee. EG: Received additional rehabilitation sessions based on virtual reality (VR), while the CG: Underwent traditional rehabilitation without the use of VR.	Specialised non immersive VR. Rehabilitation system (VRRS) developed by Khymeia group.	VAS. WOMAC. EQ‐5D. GPE score. FIM. Isometric strength of quadriceps and Hamstrings (measured the strength of these muscle groups using a dynamometer). Knee ROM with goniometry. Proprioception: Evaluated using the stabilometric platform of the VRRS.	T0 = baseline assessment (3–4 days after the TKA surgery). T1 = around 10 days after surgery.	VR‐based rehabilitation is not superior to traditional rehabilitation in terms of pain relief, drugs assumptions and other functional outcomes but seems to improve the global proprioception for patients received TKA.
Yu et al. (2023)	RCT.	To compare the effects of augmented reality based training with conventional therapist‐supervised training on physical performance in the early phase of rehabilitation of patients after TKR surgery.	24 individuals undergone TKR. Age: Average of 68 years old. Group EG: n = 12 Age = 68.39 (± 4.24) Group CG: n = 12 Age = 69.54 (± 3.12)	Both groups performed three sessions per week for a period of 4 weeks. Each session lasted 30 min. EG: Included progressive exercises starting from supine position in week one, moving to sitting and/or prone position in week two, and finally standing exercises during weeks three and four. The exercises included knee extensions and flexions, light weight lifting and movements to improve balance and muscle strength. CG: The programme consisted of exercises to increase the ROM of the knee joint. This included PROM performed by the therapist and 15 min of CPM in each session.	Specialised non immersive VR. Device non specified.	ROM. Muscle strength (digital hand dynamometer). VAS. Balance (dynamic balance). Motion analysis.	T0 = before the surgery. T1 = 3–4 days after TKA. T2 = 4‐week training programme.	The main findings were that a 4‐week period of augmented reality based training and therapist based training both resulted in a greater improvement in ROM, muscle strength, balance, and releasing pain. No significant difference was found between the two groups.
Janhunen et al. (2023)	RCT.	To investigate the effects of a home‐based, exergaming intervention on physical function and pain in older adults after TKR.	52 individuals who had undergone TKR. EG: n = 25 Age: 66.9 (± 3.1) CG: n = 27 Age: 66.4 (± 4.5)	All participants received their usual treatment after TKR. In addition, regardless of the assigned group, all participants received standard protocol home exercise instructions from a physical therapist during the hospital stay. Interventions in the EG and CG were initiated after discharge and lasted for 16 weeks. EG: In this group engaged in unsupervised exergame‐based home exercise. CG: Participants in this group followed an unsupervised home exercise programme based on a standard protocol.	Specialised non immersive VR. Not mention the specific software and hardware utilised.	OKS. TUG. VAS. 10MWT. SPPB. ROM. Isometric knee extension and flexion force tests.	T0 = baseline (within 2 weeks before the day of surgery). T1 = 2 months. T2 = 4 months follow.	The conclusions indicate that in patients undergoing TKR, the use of customised exergames for home‐based training proved to be effective in improving mobility compared to standard exercise. However, there were no significant differences between the two groups in terms of pain reduction.
Jin et al. (2018)	RCT.	To evaluate clinical benefits of VR intervention for postoperative rehabilitation in osteoarthritis (OA) patients undergoing TKA.	66 individuals undergoing unilateral TKA. EG: n = 33 Age = 66.45 (± 3.49) CG: n = 33 Age = 66.30 (± 4.41)	EG: Intervention was applied in the beginning the second day after TKA. Patients were asked to row a boat using knee flexion (interaction of VR) in an immersive virtual environment for 30 min periods, three time a day. CG: Intervention was applied in beginning the second day after TKA. Patients were asked to flex their knees passively using their arms until pain tolerance was reached. They held that position for 20 s followed by relaxing for 40 s. Patients were required to perform three sets of 30 repetitions daily	Immersive VR. Mide Technology Inc. Cangzhou, China.	Womac. HSS. VAS. ROM.	1 day after TKA. 3 days after TKA. 5 days after TKA. 7 days after TKA. Womac and HHS before TKA and 1, 3, 6 months after TKA. ROM 3, 7, and 14 days after TKA.	Clinical application of VR intervention can aid rehabilitation, reduce postoperative pain, and improve functional recovery in OA patients undergoing TKA.
Karakoc Z. et al. (2019)	RCT.	The primary aim was to investigate the effectiveness of Nintendo Wii© balance games added to the accelerated rehabilitation programme after ACL reconstruction.	22 individuals who have undergone arthroscopic ACL reconstruction. EG: n = 14 Age: 31 (± 8.41) CG: n = 8 Age: 24 (± 5.94)	Both groups 3 sessions for week duration: 6 weeks 18 sessions in total. CG: Physical therapy sessions may include muscle strengthening exercises, stretching, joint mobilisation, manual therapy, and knee stability work. EG: Nintendo Wii balance games were added to the rehabilitation programme in the 4th week. These games lasted 40 min after the main rehabilitation programme, and the patients were given a 10 min rest period before starting the Nintendo Wii balance games. In the last week of the virtual rehabilitation programme, the difficulty levels of the games were increased.	Gaming VR. Wii, Nintendo.	VAS. LEFS. Centre of gravity effected side by Wii Balance score by Wii.	T0 = baseline. T1 = 3 weeks. T2 = 6 weeks.	The Nintendo Wii© balance games applied in the clinic under physiotherapist supervision did not change the outcome of the rehabilitation in early period after ACL reconstruction.
Ozlu et al. (2023)	RCT.	To evaluate the effect of a disease‐specific gamified rehabilitation programme using VR glasses on the extent of pain, disability, function and balance in patients with knee OA.	73 individuals with knee osteoarthritis. EG: n = 35 Age = 53.28 (± 10,42) CG: n = 38 Age = 53.71 (± 9.65)	Both groups performed Frequency: 5 days a week for 3 weeks. Both groups received treatment for the same period of time. EG: Additional treatment to conservative therapy with specific game through virtual reality (VR) glasses. CG: Standard conservative treatment. Conservative therapy, including ultrasound therapy and transcutaneous electrical nerve stimulation.	Immersive VR VR glasses of the Oculus quest 128 GB type. Software: A disease‐specific game was developed for use with the VR glasses. However, the article does not provide a specific name for this software.	VAS. LFKS. 6MWT. WOMAC. BBS.	T0 = baseline. T1 = 3 weeks. T2 = 7 weeks.	The implementation of specific VR‐based gamification can improve pain, function and balance in patients with knee osteoarthritis, paving the way for further research on its long‐term effectiveness and its effects on structural healing.
Pournajaf et al. (2022)	RCT.	The primary aim of the study was to evaluate the efficacy of balance training using non‐immersive VR based serious games compared to conventional therapy in patients with TKR.	56 individuals with TKR. EG: n = 29 Age = 68.31 (± 8.12). CG: n = 27 Age = 71.10 (± 5.75)	Both groups received 15 sessions of balance and proprioception training, each session lasting 45 min and conducted five times per week. This training was in addition to their conventional rehabilitation. EG: They performed balance training using non‐immersive VR based serious games with bioFeedback CG: Underwent balance training with conventional therapy, which included exercises to improve autonomy in postural transitions and gait, such as load distribution and gradual load shifting on the operated limb, postural control, and proprioceptive exercises in an upright position.	Specialised non immersive VR. Virtual reality rehabilitation system (VRRS) developed by Khymeia group.	TUG. 10mWT. VAS. MRC. Modified Barthel index.	T0 = baseline. T1 = 3 weeks after.	Balance training using non‐immersive VR based serious games is effective but not superior to conventional therapy after TKR.
Lin, Lee, and Hsieh (2020)	RCT.	To compare the effects of AVGs with those of traditional therapeutic exercise on patients with knee OA.	80 individuals with diagnosis of osteoarthritis. Age: 40–85 years CG: n = 40 Age = 55.9 (± 15.8) EG: n = 40 Age = 58.1 (± 16.9)	Both groups received 3 therapy sessions per week for 4 weeks. Participants in both groups received hot packs to be applied to both knees for 20 min, followed by transcutaneous electrical nerve stimulation (TENS) for 20 min. Then, they received 20 min of active video games or therapeutic exercise, according to their group allocation EG: Video games. Two game sessions, each lasting 10 min. CG: Stabilisation, strength, balance, and posture exercises. Exercise bike for 10 min at a heart rate of 40%–60%. The session lasted 20 min in total.	Specialised non immersive VR. Hot plus system (supreme investment co., Taipei).	WOMAC. Static and Dynamic balance. Physical performance. Quality of life. Psychosocial distress. Fatigue. Work ability. Degree of chronic pain and disability.	T0 = baseline. T1 = 2 weeks. T2 = 4 weeks. T3 = 1 month. T4 = 3 months.	Therapeutic exercises and playing AVGs similarly improved the pain of patients with knee OA; however, playing AVGs improved dynamic balance, physical functional performance, and physical health more than therapeutic exercises.

Abbreviations: 10MWT = 10 m Walk Test, 2MWT = 2‐Minute Walk Test, 6MWT = 6 Minutes Walking Test, ABCS = Activity‐Specific ‐Balance Confidence Scale, ACL = Anterior Cruciate Ligament, Avgs = Active Video‐Game, BBS = Berg Balance Scale, CB&M = Community Balance And Mobility Scale, CCT = Controlled Clinical Trial, CG = Control Group, COG = Centre Of Gravity, CPM = Continuous Passive Movement, EEG = Electroencephalography, EG = Experimental Group, F = Female, FIM = Functional Independent Measure, GPE = Global Perceived Effect, HSS = Hospital For Special Surgery Knee Score, LEFS = Functionality Of The Knee In Patients After, LFKS = Lysholm Functional Knee Score, LOR = Length Of Outpatient Rehabilitation, M = Male, Mbi = Modified Barthel Index, MRC = Medical Research Council, Msebt = Modified Star Excursion Balance Test, N = Number Of Participants, OA = Osteoarthritis, OKS = Oxford Knee Score, PFP = Patellofemoral Pain, PROMS = Passive Range Of Motion, RCT = Randomised Controlled Trial, ROM = Range Of Motion, SF‐36 = Short Form, SPPB = Short Physical Performance Battery, STAI = State‐Trait Anxiety Inventory, TKR = Total Knee Replacement, TUG = The Timed Up And Go Test, VAS = Visual Analogue Scale, VR = Virtual Reality, Womac = Western Ontario And Mcmaster Universities Osteoarthritis.

TABLE 5.

Characteristics of included studies: ankle and foot.

Author (Year)	Study design	Aim	Group characteristics	Rehabilitation intervention	VR technology	Outcome measures	Timepoints	Conclusions
Punt et al. (2016)	RCT.	To compare the effectiveness of exercise training using the Wii Fit™ in ankle sprain patients: With physical therapy.	90 individuals with mild (grade I) or moderate (grade II) lateral ankle sprain.	EG: Exercise training with the Nintendo Wii Fit™	Gaming VR.	FAAM.	T0 = baseline.	In conclusion, the Wii Fit™ could be used as an exercise therapy to treat ankle sprain patients.
				After this initial physical therapy session, patients were instructed to inde‐pendently carry out their rehabilitation programme during 6 weeks, practicing a minimum of two times per week, for 30 min persession, at the difficulty level preferred.		VAS.	T1 = 6 weeks.	However, Wii Fit™ was not more effective than only physical therapy, or no exercise therapy at all. Patients who did not receive treatment showed similar results as people who got any kind of exercise therapy.
			Age: 18–64 years.		Wii Fit
				PTG: All physical therapists who treated patients for this study received treatment guidelines in order to guarantee homogenous treatment modalities, namely joint mobilisation, muscle strengthening, and proprioceptive exercises.	Nintendo.
			EG: n = 30
			Age = 34.7 (± 10.7)
			CG: n = 30	CG: no exercise
			Age = 34.7 (± 11.3)
			FTG: n = 30
			Age = 33.5 (± 9.5)
Rougereau et al. (2023)	RCT.	To evaluate the impact of using virtual reality masks for the management of preoperative anxiety and its impact on postoperative and predischarge anxiety as well as postoperative analgesia during outpatient hallux valgus surgery.	60 individuals with 6 months of pain with appropriate medical management by orthoses and painkillers before surgery old, independent, and living at home, with scheduled outpatient HV surgery.	EG: The session lasted 10 minutes without any external intervention from the staff with the virtual reality headset preoperatively with Oculus Go	Immersive VR.	VAS.	T0 = before the surgery.	The use of a virtual reality hypnosis mask before surgery modestly reduced postoperative and predischarge anxiety as well as early postoperative consumption of higher‐level analgesics in adults with significant preoperative anxiety.
				Patients had different choices with the virtual reality mask (Oculus Go, B07D7HL9KV): Voices (female or male), landscapes (sea, beach, or forest), as well as several musical styles. They were brought to a specific room and explained how to use the virtual reality hypnosis mask.		STAI.	T1 = after the surgery.
			Age = > 18		Oculus Go. (B07D7HL9KV)
			EG: n = 30 paused the virtual reality headset preoperatively and Age = 55 (±13)
				CG: Patients followed the traditional care pathway.
			CG: n = 30 patients followed the traditional care pathway
			Age = 52 (± 15)

Abbreviations: CCT = Controlled Clinical Trial, CG = Control Group, EG = Experimental Group, F = Female, FAAM = Foot And Ankle Ability Measure, M = Male, N = Number Of Participants, PTG = Physical Therapy Group, RCT = Randomised Controlled Trial, STAI = Stay Trait Anxiety Inventory,VAS = Visual Analogue Scale, VR = Virtual Reality.

Four studies targeted UL impairments, with 2 using specialised non‐immersive VR (Menek et al. 2022; Rizzato et al. 2023), 1 using gaming VR (Pekyavas and Ergun 2017), and 1 using immersive VR (Naqvi et al. 2022), totalling 102 patients (51 in VR and 51 in TAU). Menek et al. (Menek et al. 2022) found that non‐immersive VR exercises improved pain and quality of life (QoL) compared with TAU in patients with rotator cuff rupture. Pekyavas et al. (Pekyavas and Ergun 2017) showed that gaming VR exercises reduced pain intensity in patients with subacromial impingement syndrome (SAIS) similar to TAU exercises. Rizzato et al. (Rizzato et al. 2023) found that adding non‐immersive VR to TAU was as effective as TAU alone in shoulder rehabilitation. Naqvi et al. (Naqvi et al. 2022) reported significant pain relief in post‐distal radius fracture (DRF) rehabilitation using immersive VR compared to TAU.

Three studies focused on neck impairments, including 1 with specialised non‐immersive VR (Nusser et al. 2021) and 2 with immersive VR (Cetin, Kose, and Oge 2022; Tejera et al. 2020), with a total of 113 patients (56 in VR and 57 in TAU), all targeting chronic neck pain. Nusser et al. (Nusser et al. 2021) observed improvement in pain with non‐immersive VR sensorimotor training. Cetin et al. (Cetin, Kose, and Oge 2022) found no difference in pain and QoL between motor control exercises (MCE) combined with immersive VR and MCE alone. Tejera et al. (Tejera et al. 2020) reported no significant difference in pain between immersive VR and TAU though kinesiophobia improved in the VR group.

Seven studies targeted chronic low back pain (CLBP), with 2 using immersive VR (Eccleston et al. 2022; Stamm et al. 2022), 3 using specialised non‐immersive VR (Matheve, Bogaerts, and Timmermans 2020; Nambi et al. 2021; Yelvar et al. 2017), and 2 using gaming VR (Afzal et al. 2022; Monteiro et al. 2015), totalling 309 patients (154 in VR and 155 in TAU). Stamm et al. (Stamm et al. 2022) and Ecclestone et al. (Eccleston et al. 2022) compared immersive VR to educational interventions or TAU, finding no significant difference in QoL between the groups. However, all seven studies reported pain reduction in both VR and TAU. Yilmaz Yelvar et al. (Yelvar et al. 2017) demonstrated that virtual walking combined with TAU reduced pain and kinesiophobia compared with TAU alone but did not improve QoL. Nambi et al. (Nambi et al. 2021) showed the effectiveness of non‐immersive VR, both alone and combined with TAU, in pain management. Matheve et al. (Matheve, Bogaerts, and Timmermans 2020) found that VR distraction reduced pain intensity during pelvic tilt exercises in CLBP patients. Monteiro‐Junior et al. (Monteiro et al. 2015) and Afzal et al. (Afzal et al. 2022) both observed that gaming VR combined with TAU reduced pain and functional disability more effectively than TAU alone.

Thirteen studies addressed LL impairments, including 11 on knee impairments (3 using specialised non‐immersive VR), 3 using immersive VR (Gianola et al. 2020; Janhunen et al. 2023; Yu et al. 2023), and 4 using gaming VR (Ebrahimi et al. 2021; Fung et al. 2012; Karakoc et al. 2019; Lin, Lee, and Hsieh 2020), with 547 patients (289 in VR and 258 in TAU). Seven studies focused on total knee arthroplasty (TKA) (Fuchs et al. 2022; Fung et al. 2012; Gianola et al. 2020; Janhunen et al. 2023; Jin et al. 2018; Pournajaf et al. 2022; Yu et al. 2023), two on knee osteoarthritis (OA) (Lin, Lee, and Hsieh 2020; Ozlu et al. 2023), one on patellofemoral pain (PFP) (Ebrahimi et al. 2021), and one on post‐anterior cruciate ligament (ACL) surgery (Karakoc et al. 2019). Five TKA studies found that VR (specialised non‐immersive or exergames) was not superior to TAU for pain and QoL, suggesting VR as a potential alternative or complement to standard treatments. Jin et al. (Jin et al. 2018) and Fuchus et al. (Fuchs et al. 2022) found that immersive VR in the immediate post‐operative period reduced pain and anxiety. In knee OA studies, Yu et al. (Lin, Lee, and Hsieh 2020) concluded that gaming VR improved physical health more than TAU, while Ozlu et al. (Ozlu et al. 2023) showed that immersive VR improved pain, function, and balance. Ebrahimi et al. (Ebrahimi et al. 2021) found that exergaming was more effective than a training education programme in improving pain, QoL, and brain function in women with PFP. Karakoc et al. (Karakoc et al. 2019) reported no difference in outcomes for balance games added to accelerated rehabilitation post‐ACL surgery.

Two studies focused on the ankle‐foot complex. Rougereau et al. (Rougereau et al. 2023) used immersive VR to manage preoperative anxiety in hallux valgus surgery, finding reduced postoperative anxiety and analgesic use. Punt et al. (Punt et al. 2016) demonstrated that exergaming in ankle sprain patients yielded similar improvements in pain, function, return to sport, satisfaction, and effectiveness compared with TAU.

3.3. Methodological Evaluation of Studies

Figure 2 illustrates the risk of bias across the included studies. Of the 27 studies, only one (Janhunen et al. 2023) was classified as having a low risk of bias, while 22 studies (Afzal et al. 2022; Cetin, Kose, and Oge 2022; Ebrahimi et al. 2021; Eccleston et al. 2022; Fuchs et al. 2022; Fung et al. 2012; Jin et al. 2018; Lin, Lee, and Hsieh 2020; Matheve, Bogaerts, and Timmermans 2020; Menek et al. 2022; Monteiro et al. 2015; Nambi et al. 2021; Naqvi et al. 2022; Nusser et al. 2021; Pekyavas and Ergun 2017; Pournajaf et al. 2022; Punt et al. 2016; Rizzato et al. 2023; Stamm et al. 2022; Tejera et al. 2020; Yelvar et al. 2017; Yu et al. 2023) showed moderate risk of bias primarily due to reporting concerns. Four studies had a high risk of bias due to missing outcome data (Gianola et al. 2020; Karakoc et al. 2019; Ozlu et al. 2023; Rougereau et al. 2023) and deviations from intended interventions (Rougereau et al. 2023).

FIGURE 2.

Risk of bias in included studies.

A ‘some concerns’ risk of bias was identified in the randomisation process of five studies (Afzal et al. 2022; Fuchs et al. 2022; Jin et al. 2018; Rougereau et al. 2023; Yelvar et al. 2017).

A high risk of bias due to missing outcome data was observed in four studies (Gianola et al. 2020; Karakoc et al., 2019; Ozlu et al. 2023; Rougereau et al. 2023), each with a dropout rate exceeding 10%.

In the selection of reported results, 25 studies (Afzal et al. 2022; Cetin, Kose, and Oge 2022; Ebrahimi et al. 2021; Eccleston et al. 2022; Fuchs et al. 2022; Fung et al. 2012; Jin et al. 2018; Lin, Lee, and Hsieh 2020; Matheve, Bogaerts, and Timmermans 2020; Menek et al. 2022; Monteiro et al. 2015; Nambi et al. 2021; Naqvi et al. 2022; Nusser et al. 2021; Ozlu et al. 2023; Pekyavas and Ergun 2017; Pournajaf et al. 2022; Punt et al. 2016; Rizzato et al. 2023; Rougereau et al. 2023; Stamm et al. 2022; Tejera et al. 2020; Yelvar et al. 2017; Yu et al. 2023) showed moderate risk of bias. Only two studies (Gianola et al. 2020; Janhunen et al. 2023) demonstrated a low risk, with both providing the study protocol and a detailed analysis plan. Among the remaining studies, 11 (Fuchs et al. 2022; Jin et al. 2018; Karakoc et al. 2019; Nambi et al. 2021; Naqvi et al. 2022; Nusser et al. 2021; Ozlu et al. 2023; Pekyavas and Ergun 2017; Punt et al. 2016; Yelvar et al. 2017; Yu et al. 2023) had no available protocol, while 14 (Afzal et al. 2022; Cetin, Kose, and Oge 2022; Ebrahimi et al. 2021; Eccleston et al. 2022; Fung et al. 2012; Lin, Lee, and Hsieh 2020; Matheve, Bogaerts, and Timmermans 2020; Menek et al. 2022; Monteiro et al. 2015; Pournajaf et al. 2022; Rizzato et al. 2023; Rougereau et al. 2023; Stamm et al. 2022; Tejera et al. 2020) had a protocol but lacked an analysis plan. All studies had a low risk of bias for outcome measurement.

3.4. Effects of Intervention

3.4.1. Effect of VR Treatment on Pain for UL Region

We included four studies with a total of 102 participants. Analysis was conducted using MD with a fixed‐effect model. A subgroup analysis categorised the four studies into gaming VR (Pekyavas and Ergun 2017), specialised immersive VR (Naqvi et al. 2022), and non‐immersive VR (Menek et al. 2022; Pekyavas and Ergun 2017; Rizzato et al. 2023). Three studies (Menek et al. 2022; Naqvi et al. 2022; Pekyavas and Ergun 2017) demonstrated greater effectiveness than the control group, while one study (Rizzato et al. 2023) showed no significant difference between the groups. Due to the high heterogeneity among the results, a pooled analysis could not be performed (Figure 3).

FIGURE 3.

Comparison of VR versus TAU for pain management in UL region. A green block indicates the weight assigned to the study, and the horizontal line depicts the confidence interval.

3.4.2. Effect of VR Treatment on Pain for Neck Region

We included three studies with a total of 103 participants. Given the use of similar assessment tools—VAS (Cetin, Kose, and Oge 2022; Tejera et al. 2020) and NRS (Nusser et al. 2021)—we applied MD with a fixed‐effect model. A subgroup analysis divided the studies into specialised non‐immersive VR (Nusser et al. 2021) and immersive VR (Cetin, Kose, and Oge 2022; Tejera et al. 2020). The results did not show significant differences between the two treatment groups (MD = 0.18; 95% CI: − 0.52 to 0.87; I ² = 49%, p = 0.14) (Figure 4).

FIGURE 4.

Comparison of VR versus TAU for pain management in the neck region. Note: A green block indicates the weight assigned to the study, and the horizontal line depicts the confidence interval. Black rhombi show the overall results.

3.4.3. Effect of VR Treatment on Pain for LB Region

We included six studies with a total of 136 participants. Due to the use of similar assessment tools—VAS (Afzal et al. 2022; Nambi et al. 2021; Yelvar et al. 2017) or NPRS (Eccleston et al. 2022; Monteiro et al. 2015; Stamm et al. 2022)—we applied MD with a fixed‐effect model. Subgroup analysis divided the studies into gaming VR (Afzal et al. 2022; Monteiro et al. 2015), immersive VR (Eccleston et al. 2022; Stamm et al. 2022), and specialised non‐immersive VR (Nambi et al. 2021; Yelvar et al. 2017). The results indicated a significant difference favouring the VR treatment group overall (MD = − 1.38; 95% CI: − 1.59 to − 1.18; I ² = 92%, p < 0.00001), as well as within two subgroups: specialised non‐immersive VR (MD = − 2.01; 95% CI: − 2.28 to − 1.74; I ² = 0%, p < 0.00001) and gaming VR (MD = − 0.48; 95% CI: − 0.83 to − 0.13; I ² = 0%, p = 0.007) (Figure 5). In contrast, the immersive VR subgroup showed no significant difference between treatment groups (MD = 0.58; 95% CI: − 0.53 to 1.69; I ² = 47%, p = 0.31) (Figure 5).

FIGURE 5.

Comparison of VR versus TAU for pain management in the LB region. A green block indicates the weight assigned to the study, and the horizontal line depicts the confidence interval. Black rhombi show the overall results.

3.4.4. Effect of VR Treatment on Pain for LL Region

We included seven studies, all focused on the knee region, with a total of 375 participants. Because of the use of different assessment tools—VAS (Gianola et al. 2020; Janhunen et al. 2023; Karakoc et al. 2019; Ozlu et al. 2023; Pournajaf et al. 2022; Yu et al. 2023) and the pain subscale of the WOMAC questionnaire (Lin, Lee, and Hsieh 2020)—we used SMD with a random‐effects model for analysis. Subgroup analysis categorised the studies into specialised non‐immersive VR (Gianola et al. 2020; Janhunen et al. 2023; Pournajaf et al. 2022; Yu et al. 2023), gaming VR (Karakoc et al. 2019; Lin, Lee, and Hsieh 2020), and immersive VR (Ozlu et al. 2023). Results showed a significant difference favouring the VR treatment group overall (SMD = − 0.33; 95% CI: − 0.55 to − 0.10; I ² = 13%, p < 0.004), as well as within subgroups: specialised non‐immersive VR (SMD = − 0.32; 95% CI: − 0.62 to − 0.03; I ² = 10%, p < 0.003) and immersive VR (SMD = − 0.94; 95% CI: − 1.58 to − 0.30; I ² = not applicable, p < 0.004). The two studies in the gaming VR group (SMD = − 0.09; 95% CI: − 0.49 to 0.30; I ² = 0%, p = 0.64) showed no significant difference between treatment groups (Figure 6).

FIGURE 6.

Comparison of VR versus TAU for pain management in the LL region. A green block indicates the weight assigned to the study, and the horizontal line depicts the confidence interval. Black rhombi show the overall results.

3.4.5. Effect of VR Treatment on QoL Across all Regions

We included seven studies with a total of 309 participants. Due to the use of different assessment tools—SF‐36 (Cetin, Kose, and Oge 2022), SF‐12 (Stamm et al. 2022), EuroQol‐5D‐5L (Eccleston et al. 2022; Gianola et al. 2020), NHP (Yelvar et al. 2017), WHOQOL (Lin, Lee, and Hsieh 2020), and RCQOL (Menek et al. 2022)—analysis was conducted using SMD with a random‐effects model. A subgroup analysis categorised the studies by anatomical region: UL (Menek et al. 2022), neck (Cetin, Kose, and Oge 2022), LB (Eccleston et al. 2022; Stamm et al. 2022; Yelvar et al. 2017), and LL (Gianola et al. 2020; Lin, Lee, and Hsieh 2020). The results showed no significant difference between treatment groups in four studies (Eccleston et al. 2022; Lin, Lee, and Hsieh 2020; Stamm et al. 2022; Yelvar et al. 2017). One study (Cetin, Kose, and Oge 2022) found the VR more effective than the TAU, while two studies (Gianola et al. 2020; Menek et al. 2022) favoured the TAU group over the intervention. Due to high data heterogeneity, pooled analysis was not feasible (Figure 7).

FIGURE 7.

Comparison of VR versus TAU for QoL across all regions. A green block indicates the weight assigned to the study, and the horizontal line depicts the confidence interval.

4. Discussion

This study primarily aimed to evaluate the effectiveness of VR‐based interventions for MSK pain relief, considering both anatomical region and software type. To our knowledge, this review is the first to evaluate VR's effect on pain across all anatomical regions within the MSK domain.

In our meta‐analysis, knee‐related VR treatments showed a statistically significant reduction in pain with low heterogeneity, particularly with non‐immersive VR. This supports VR's role in managing knee conditions such as osteoarthritis (Ozlu et al. 2023), patellofemoral syndrome (Ebrahimi et al. 2021), and post‐operative recovery, including knee arthroplasty (Gianola et al. 2020; Janhunen et al. 2023; Pournajaf et al. 2022; Yu et al. 2023) and ACL surgery (Karakoc et al. 2019). Interestingly, while individual studies, especially those on knee arthroplasty (Gianola et al. 2020; Janhunen et al. 2023; Pournajaf et al. 2022; Yu et al. 2023), did not indicate VR superiority over TAU, the aggregated data in the meta‐analysis reveal a clear advantage, suggesting that VR might offer benefits not apparent in isolated studies. For the neck region, both treatments appear comparable in effectiveness. These results help address the uncertainties raised in the recent study by Byra et al. (Byra and Czernicki 2020) regarding the effectiveness of VR in reducing pain compared with TAU. Our findings suggest that targeted pathology‐specific VR interventions—particularly those designed for knee conditions—could effectively alleviate the mixed pain (i.e., nociceptive‐nociplastic) commonly seen in MSK patients (Kosek et al. 2021). Furthermore, these interventions may reduce reliance on pharmacological treatments and improve the quality of life in postoperative patients (Byra and Czernicki 2020).

Regarding cervical spine disorders, studies focused on chronic neck pain showed no significant differences between VR treatment and TAU in our meta‐analysis, and similarly, no significant improvements were observed for the secondary outcome of QoL across all anatomical regions. Supporting our findings, recent studies by Guo et al. (Guo et al. 2023) and Hao et al. (Hao et al. 2024) provide moderate evidence that VR can be an effective non‐pharmacological approach for reducing pain intensity in patients with neck pain, particularly in the short term. For enhanced effectiveness in chronic neck pain, VR should ideally be integrated into a multimodal intervention (Guo et al. 2023).

For other regions, such as the lower back, shoulder, and foot, high heterogeneity prevents drawing valid conclusions. This variability across most anatomical regions in our review necessitates careful commentary. Potential sources of heterogeneity in our study may stem from the inherently subjective nature of pain perception, which varies widely among individuals due to psychological and cultural influences (Fillingim 2005). The diversity of MSK conditions studied, each with distinct pain mechanisms (e.g., nociceptive vs. nociplastic pain), adds further complexity. These factors, combined with variations in assessment tools and individual patient characteristics, likely contribute to the high heterogeneity observed in our findings. Furthermore, no RCTs were available for regions such as the wrist, hand, elbow, and hip, underscoring a gap in the literature that warrants further exploration through high‐quality research (Byra and Czernicki 2020; Chaplin, Karatzios, and Benaim 2023; Gumaa and Rehan Youssef 2019).

This study aligns with the scoping review by Collado et al. (Collado et al. 2018)which found insufficient evidence for exergames in reducing MSK pain, and with the recent systematic review by Elaraby et al. (Elaraby et al. 2023), which supported VR for improving balance in ankle injuries but found no significant pain reduction. Further high‐quality research is needed to address these limitations.

4.1. Study Limitations

This review has several limitations. High heterogeneity in VR methods, exercise dosages, and targeted impairments complicates generalising VR's effectiveness for pain, as research demands standardized protocols, while clinical practice favours personalised care. The absence of RCTs for certain regions (hip, hand, elbow) limits assessments for these areas. Furthermore, the medium‐to‐low methodological quality of studies and the lack of a GRADE assessment reduce the robustness and clinical applicability of the findings.

4.2. Clinical Implication and Future Study Directions

From a clinical perspective, VR offers diverse options, including immersive, gaming, and non‐immersive experiences, adaptable to individual patient needs and pain mechanisms. While VR cannot replace the expertise of interdisciplinary rehabilitation teams, it can enhance outcomes in MSK rehabilitation (Chaplin, Karatzios, and Benaim 2023), particularly in an era of rapid technological advancement. VR shows promise for pain reduction in MSK disorders (Ahmadpour et al. 2019; Pourmand et al. 2018), improving patient compliance, adherence, motivation (Mouatt et al. 2020) and engaging brain networks that are less accessible through traditional methods (Ahmadpour et al. 2019).

Successful VR integration requires feasibility studies to identify suitable patients and mitigate risks. Collaboration between healthcare professionals and developers is key to creating effective, cost‐efficient, and accessible VR tools. Future research should focus on standardized systems, consistent evaluation, and evidence for various VR modalities, including their effects on nociceptive, neuropathic and nociplastic pain in MSK disorders.

5. Conclusions

This review provides inconclusive recommendations for VR in general MSK pain management. While VR shows effectiveness for knee pain and comparable outcomes to TAU for chronic neck pain, evidence on QoL improvements is limited. Medium‐to‐low study quality and variability in VR interventions highlight the need for larger, rigorous research to clarify the efficacy of VR across MSK conditions.

Author Contributions

Conceptualisation, M.Z. and P.K.; methodology, M.Z., L.C., B.C. and P.K.; software, M.Z., L.C. and B.C.; validation, B.C., L.C. and P.K.; formal analysis, M.Z. and L.C.; investigation, F.M., L.S., S.F. and M.R.; resources, A.R. and G.P.; data curation, M.Z., B.C. and P.K. writing–original draft preparation, M.Z.; writing–review and editing, L.C., B.C. and P.K.; visualisation, G.P.; supervision, P.K.; project administration, M.Z.; All authors have read and agreed to the published version of the manuscript.

Ethics Statement

The authors have nothing to report.

Consent

The authors have nothing to report.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Supporting Information S1

Supporting Information S2

Data Availability Statement

Data are available upon reasonable request to the corresponding author.