VR compared to usual care
The comparison of internet-based VR and usual care showed an effect on Vertigo Symptom Scale (VSS-SF) total score (primary outcome) in favour of VR (
n = 296 participants, − 2.26 points,
p = .02) for patients with dizziness over the last 2 years and still experiencing dizziness triggered by head movements. Secondary outcomes showed mixed effects. Analysis of DHI score (− 5.58 points,
p = .01), and the patient reported improvement (
p < .001) revealed effects in favour of VR. No significant differences were found in the Hospital Anxiety and Depression Scale (HADS) [
35].
No difference in DHI (primary outcome) was reported for patients with dizziness when comparing usual care with a multicomponent program that includes the assessment of fall-risk increasing drugs (FRIDs) stepped mental health care or exercise therapy (
n = 168 participants) in a RCT. No difference of dizziness frequency, presence of anxiety and depressive disorder, QoL and fall frequency was found [
49].
A RCT with 165 participants experiencing balance disorders compared the Otago programme with receiving a fall-prevention booklet and continuing usual activities (optimized usual care). No effect in favour of the intervention could be observed in the primary outcomes mCTSIB, Limits of Stability (LOS), Rhythmic Weight Shift (RWS). Within secondary outcomes, an effect in favour of intervention was shown in the step test (worse leg) (+ 2.10 steps/15 s,
p ≤ .001), in hip abductor muscle strength (+ .02 kg/kg,
p ≤ .001), in the Walk-across Test (WA) (− 2.17 cm,
p ≤ .001), in the Functional Reach Test (FRT) (+ 2.95 cm,
p ≤ .001) and on the Human Activity Profile–Adjusted Activity Score (HAP-AAS) (+ 4.57 points,
p ≤ .001). No effects were reported regarding Sit to Stand Test (STS), the Five Times Sit to Stand Test (5x-STS), muscle strength of quadriceps and dorsiflexors, walking speed, the Step Quick Turn test (SQT), in quality of life as measured by Assessment of Quality of Life (AQoL) and falls measured in the Modified Falls Efficacy Scale (MFES) [
44].
A non-randomized study with 60 participants experiencing balance disorders and a history of falls or having fear of falling investigated additional Cawthorne-Cooksey exercise programme versus conventional physical therapy did not specify a primary outcome. An effect in favour of the intervention (− 0.77 points,
p = .030) as measured by the Visual Analogue Scale of Fear of Falling (VAS-FOF) and in the Dynamic Gait Index (DGI) (+ 1.3 points,
p = .013) was reported. No differences in Berg Balance Scale (BBS) and the likelihood of falls were found [
50].
A RCT with 660 participants with mild to moderate Parkinson’s Disease (Hoehn and Yahr stages 2–3) evaluated the effectiveness of VR versus usual care. The study did not specify a primary outcome
. Mixed results were found: A significant benefit of + 9 points (
p = < .05) on BBS, + 4 points (
p = < .05) in DGI and + 27.5 points for Activities-specific Balance Confidence (ABC) (
p < .05). No significant difference was found in mCTSIB total score, Unified Parkinson’s Disease Rating Scale (UPDRS), Timed-Up and Go test (TUG) and Quality of life measured by the Parkinson’s Disease Questionnaire (PDQ-39) [
30].
When comparing classical physiotherapy (described as “individually tailored and including flexibility, strengthening, posture, breathing balance, walking exercises, and other functional activities”) with additional sensorimotor integration training versus classical physiotherapy (
n = 30 participants with Parkinson’s Disease Hoehn and Yahr stages 2–3, no primary outcome stated), mixed results were found in a RCT. Effects in favour of intervention were found in the 5th position (+ 24.16,
p = .027) and composite (+ 12.8,
p = .042) of Computerized Dynamic Posturography – Sensory Organization Test (CDP-SOT) and in vestibular system score (VEST) in Computerized Dynamic Posturography –Sensory Analysis (CDP-Sensory) (+ 25.43,
p = .048), on BBS (+ 10.34 points,
p = .037) and in TUG (− 4.11 s,
p = .002). No differences were reported for 6th position of CDP-SOT, somatosensory system score (SOM), visual system score (VIS) and visual preference score (PREF) in CDP-Sensory, Unified Parkinson’s Disease Rating Scale (UPDRS) and the Functional Reach Test (FRT) [
47].
Moderate quality of evidence exists, that VR is superior to usual care to improve VDB symptoms, balance and mobility, but not postural control, the impact of VDB on ADL and the presence of anxiety and depression, Parkinson’s disease specific ADL, quality of life, frequency of falls and fear of falling.
VR versus no intervention
Two studies investigated VR versus no intervention.
A RCT (
n = 85 participants with fall-related wrist fractures) showed no differences in primary outcomes (tandem standing with eyes open and closed and walking in a modified figure of eight). In secondary outcomes, no differences were reported when measuring SOLEO, SOLEC, 5x-STS, postural sway, vibration sense, head-shake test, EQ. 5D-VAS and walking variations [
37].
A non-randomized study (
n = 58 participants with multisensory dizziness) stated no primary outcome. Mixed effects were found. An improvement in standing on one leg with eyes closed (SOLEC) (+ 1 s,
p = .038) and in walking heel to toe (− 2 steps,
p = .044). No difference was observed in standing on one leg with eyes open (SOLEO), tandem standing with eyes open and closed, DHI, steps outside during walking in a figure of eight and the risk of falls maintained [
36].
Training computer dynamic posturography exercises compared to no intervention (
n = 139 participants experiencing balance impairment without a vestibular disease, no primary outcome stated) revealed to no differences in SOT, LOS, DHI, TUG and FES-I in a four-arm study, for which other comparison groups are described as follows [
51].
New variations versus established forms of VR
VR in addition to CRM was compared to the CRM alone (
n = 16 participants with BPPV for at least 6 months) in a RCT. Primary outcomes showed mixed effects: A difference in Maximum Excursion (MXE) of LOS (+ 17%,
p < .05) and DGI (+ 4 points,
p = .05) in favour of intervention and no differences in mCTSIB and movement velocity (MVL) of LOS. Secondary outcomes also revealed mixed results: a difference in tandem end sway (1 s in the
p < .05) favouring intervention and no difference in sway in Unilateral Stance Test (US) and VAS [
41].
A RCT with 125 participants (older people referred to a Falls Outpatient Clinic) investigated the Otago exercise programme in groups compared to the Otago exercise programme at home. The primary outcome BBS showed no difference. Secondary outcomes revealed mixed effects. Significant differences in 5x-STS (+ 2.2 s,
p = .005) and TUG (− 2.4 s,
p = .038) were reported. No differences were shown in quality of life measuring the short-form questionnaire SF-36 and on the Fall Efficiency Scale International (FES-I) [
48].
A RCT with 82 participants with dizziness resulting from a vestibular disorder assessed a multimodal version of the Cawthorne-Cooksey programme versus the conventional version and observed no difference in primary outcome DGI. Also secondary outcomes showed no difference measuring STS, Romberg, tandem stand, sensorial, unipedal and handgrip strength, TUG, multidirectional FRT and fall rate [
42].
A four-arm RCT compared VR with computer dynamic posturography exercises to exposure to optokinetic stimuli and exercises at home based on the Cawthorne-Cooksey programme in patients with balance impairment without a vestibular disease. Information about changes in SOT, DHI, TUG and FES-I is missing. No primary outcome was stated [
51].
Moderate quality of evidence exists, that VR in addition to CRM is superior to CRM alone to improve balance. Very low quality of evidence exists, that the Otago exercise programme in groups is superior to the Otago exercise programme at home to improve lower extremity strength and mobility.
CAVR versus other interventions
A three-arm RCT (n = 36 participants) explored virtual reality-based Wii Fit training with subsequent treadmill training in comparison to fall-prevention education with no structured programme for idiopathic Parkinson’s Disease (Hoehn and Yahr stages 2–3). No primary outcome was stated. Mixed results were found. Advantages in gait parameters (+ 12.87 cm/s,
p < .05) in regard to velocity, (+ 15.41 cm,
p < .05) stride length, (+ 16.5 N,,
p < .05) hip flexors, (+ 12.5 N,
p < .05) hip extensors, (+ 14.6 N,
p < .05) knee flexors, (+ 28.1 N,
p < .05) knee extensors, (+ 37.5 N,
p < .05) ankle dorsiflexors and (+ 25.5 N,
p < .05) ankle plantar flexors, as well as (+ 20.5,
p < .05) in vestibular ratio of SOT. Also a significant difference (+ 4.59 points,
p < .05) in the Functional Gait Assessment (FGA) was observed. As the third arm, when the traditional exercise group (CG) was compared with the fall-prevention education group (CoG), all parameters changed significantly in the last follow-up except for the vision component of SOT. No primary outcome was stated. Changes in general were greater when WiiFit was compared with fall-prevention education than when traditional exercises were compared with education [
38].
Home exercises supported by the “Move it to improve it” (Mitii) computer programme versus a printed home programme (
n = 63 participants with vestibular dysfunction) showed no difference in the primary outcome one-leg stand test. No difference in secondary outcomes Motion Sensitivity, VAS, Chair stand test, DHI, DGI, quality of life measured with SF-12 [
43].
A RCT compared in-home virtual reality balance training (TeleWii) to in-clinic sensory integration balance training (
n = 76 participants with Parkinson’s Disease modified Hoehn and Yahr stages 2.5–3). No significant differences in the primary outcome BBS as well as ABC, 10-MW, DGI, Quality of life measuring PDQ-39, and in falls were observed [
34].
Moderate quality of evidence exists, that virtual reality-based Wii Fit training with subsequent treadmill training is superior in comparison to fall-prevention education with no structured programme to improve gait parameters, lower extremity strength, balance and to improve postural control.
Tai chi VR (TCVR) versus no/sham intervention
A RCT (
n = 40 participants experiencing dizziness within the past year, no primary outcome stated) investigated TCVR compared to no intervention and reported significant differences (+ 61 cm,
p = .050) forward, (+ 1.37 cm,
p = .024) backward, (+ 28.57 cm
2,
p = .002) in the maximal sway area of the LOS and (− .23 s,
p = .003) in the 8-ft up-and-go test. No differences were found in right- and leftward of LOS [
39].
TCVR was compared to music percussion as sham intervention (n = 40 participants with visual impairment) in a RCT where no primary outcome was stated. Mixed effects were found. Differences (− 30.1%,
p = .032) of the absolute angle error of the passive knee joint repositioning test, (+ 59.7%,
p = .006) in the visual ratio and (+ 50.3%,
p = .048) in the vestibular ratio of SOT. No difference were reported in concentric isokinetic knee extensor or flexor strength of dominant leg and in somatosensory ratio [
33].
Low quality of evidence exists, that TCVR is superior to no/sham intervention to improve postural control. Very low quality of evidence exists, that TCVR is superior to no/sham intervention to improve mobility.