Fully automated VR therapy delivered by a virtual coach significantly reduced fear of heights
How this was rated
Single-blind randomized controlled trial with adequate sample (n=100), published in The Lancet Psychiatry. Direct evidence for fear of heights specifically; extension to other conditions and to communication-specific populations is plausible but requires its own evidence.
Ratings use a simplified four-tier scheme (High, Moderate, Low, Very Low) informed by the GRADE working group. Learn more about how studies are rated.
In a 100-person single-blind RCT, a fully automated VR program with a virtual therapist coach produced large reductions in fear of heights - achieving outcomes comparable to therapist-delivered care without requiring a clinician in the room.
A high-profile RCT in a leading journal demonstrating that automated VR therapy can produce substantial clinical effects in a circumscribed condition (fear of heights); the broader claim - that VR therapy can be delivered without continuous clinician presence - has implementation implications across conditions.
Key findings
- Automated VR therapy produced large reductions in fear of heights compared with the control condition (HIQ Cohen's d=2.0)
- Effect size (d=2.0) greatly exceeded therapist-delivered exposure benchmarks (d=1.1 from prior meta-analysis) - not just comparable but substantially larger
- 49/49 (100%) VR participants showed reduction; 34/49 (69%) dropped below entry criterion; NNT=1.3 to halve fear; 100% follow-up at all timepoints; no adverse events
- Treatment delivered by virtual coach (graduate psychologist in room for headset setup and safety only; therapy content ran automatically)
- Trial registration ISRCTN11898283; hardware HTC Vive; 'Now I Can Do Heights' software (CE-marked Class I medical device)
Background
Most evidence on VR-based therapy comes from trials in which a clinician is present during the VR session - operating the system, guiding exposure, and providing in-the-moment support. That delivery model is appropriate for clinical research but raises a practical question for scaling: if the meaningful active ingredients of VR exposure are graded controlled stimuli and the participant’s engagement with them, can a structured automated program deliver those ingredients without continuous clinician presence? Freeman and colleagues set out to answer that question with a methodologically rigorous RCT.
What the researchers did
The researchers enrolled 100 adults with fear of heights into a single-blind randomized controlled trial. Participants were assigned to either a fully automated VR therapy program or a control condition. The automated program featured a virtual coach (an embodied conversational agent) who guided participants through six sessions of structured exposure to virtual heights, with adaptive difficulty and skill-building components. No human clinician was present during the VR sessions. Outcomes - measured by both self-report and behavioral approach to heights - were assessed at baseline, post-treatment, and follow-up by assessors blinded to condition.
What they found
The automated VR program produced large reductions in fear of heights compared to the control condition. The effect size (HIQ Cohen’s d=2.0) was not merely comparable to therapist-delivered exposure benchmarks - it substantially exceeded the d=1.1 benchmark from prior meta-analyzes of therapist-delivered exposure. Every single VR participant (49/49, 100%) showed some reduction in fear. 34 of 49 (69%) dropped below the study’s entry criterion - meaning they no longer qualified as having clinically significant fear of heights. The number needed to treat to halve fear was 1.3. Follow-up was 100% at all timepoints, and no adverse events were reported.
Critically, a graduate psychologist was present only to help with headset setup and to monitor safety - the therapeutic content ran automatically through the virtual coach system. The clinician was not delivering the therapy. This is a genuinely autonomous automated delivery model.
Why this matters
This trial is one of the first rigorous demonstrations that VR therapy can be delivered effectively without a clinician in the room. For service models where specialist clinicians are scarce or where delivering enough sessions to produce meaningful change is logistically difficult, the implementation implications are substantial. The findings do not directly address communication-specific applications - the trial was about fear of heights, which is structurally different from speaking anxiety, voice work, or social communication practice. But the implementation principle - that structured VR delivery with a virtual coach can produce meaningful effects without continuous clinician presence - is one that subsequent research can extend to other domains.
Limitations
The trial addressed fear of heights specifically; whether the automated delivery model generalizes to more complex conditions - including communication-specific contexts where a clinician’s responsive judgment may be more central - has not been directly demonstrated. Single-condition trials cannot establish a general principle of automation. The follow-up was relatively short, and longer-term durability of automated-delivery effects compared to clinician-delivered effects remains an open question.
Implications for practice
Clinician-light or clinician-free VR delivery is feasible for circumscribed conditions and produces meaningful clinical effects. For service models where access to specialist clinicians is limited, the implementation pattern offers a template - though direct evidence for any specific use case (including communication-specific applications) requires its own trials.
Implications for research
Direct trials of automated or semi-automated VR delivery for communication-specific contexts (graded speaking practice, social rehearsal) would extend Freeman et al.'s implementation model. Health-economic analyzes of clinician-light VR delivery are warranted.
Where this connects to Therapy withVR
The study above is independent research and does not endorse any product. The notes below are commentary from withVR on how the themes in this research relate to features of Therapy withVR. The research findings are not claims about Therapy withVR.
Avatar Speech System
Freeman's virtual coach delivered structured therapy guidance through avatar speech - Therapy withVR's avatar speech system supports the same kind of structured delivery in clinical contexts.
Save Setup
Reproducible session configurations support the structured, session-by-session progression that automated VR delivery requires.
Cite this study
If you reference this study in your work, the canonical citation formats are:
@article{freeman2018,
author = {Freeman, D. and Haselton, P. and Freeman, J. and Spanlang, B. and Kishore, S. and Albery, E. and Denne, M. and Brown, P. and Slater, M. and Nickless, A.},
title = {Automated psychological therapy using immersive virtual reality for treatment of fear of heights: a single-blind, parallel-group, randomised controlled trial},
journal = {The Lancet Psychiatry},
year = {2018},
doi = {10.1016/S2215-0366(18)30226-8},
url = {https://withvr.app/evidence/studies/freeman-2018}
}TY - JOUR
AU - Freeman, D.
AU - Haselton, P.
AU - Freeman, J.
AU - Spanlang, B.
AU - Kishore, S.
AU - Albery, E.
AU - Denne, M.
AU - Brown, P.
AU - Slater, M.
AU - Nickless, A.
TI - Automated psychological therapy using immersive virtual reality for treatment of fear of heights: a single-blind, parallel-group, randomised controlled trial
JO - The Lancet Psychiatry
PY - 2018
DO - 10.1016/S2215-0366(18)30226-8
UR - https://withvr.app/evidence/studies/freeman-2018
ER - Know of research that should be in this hub? If a relevant peer-reviewed study is not listed here, send the reference to hello@withvr.app. The hub is kept up to date as the literature grows.
Funding & independence
Funded by Oxford VR and NIHR Oxford Health Biomedical Research Center. Disclosure: 9 of 10 authors are co-founders or employees of Oxford VR (the company that owns the 'Now I Can Do Heights' software tested in this trial) or its related company Virtual Bodyworks - specifically DF, JF, MS, and BS are co-founders of Oxford VR; DF, JF, PH, EA, and MD are employed by Oxford VR; MS and BS are co-founders of Virtual Bodyworks; SK is employed by Virtual Bodyworks; PB and AN declare no competing interests. This COI is disclosed in the paper's declaration statement. No withVR BV involvement in funding, study design, or authorship. Summary prepared independently by withVR using the published paper.