Does VR therapy actually work? The evidence behind therapeutic VR for pain and anxiety

Does VR therapy actually work? For some things, genuinely yes — with real evidence behind it. For others, the marketing is well ahead of the science. We build RVer, a therapeutic VR app certified as a Class I medical device, used in real clinics for pain, anxiety and rehabilitation. That certification means we can't wave our hands about efficacy — a regulator made us substantiate the claims. So here's the honest version of what the evidence supports, what it doesn't, and how a medical device has to prove it rather than just assert it.

What the evidence actually supports

Therapeutic VR is not one thing. It's several different mechanisms, each with its own evidence base, and they are not equally strong.

• Distraction analgesia — the best-evidenced use. VR reliably reduces acute and procedural pain by occupying attention the brain would otherwise spend on the pain signal. The classic work is burn-wound care, and the effect replicates across dressing changes, venepuncture, dental work and paediatric procedures. Multiple randomised trials and meta-analyses show a moderate, real reduction in reported pain. This is the strongest ground therapeutic VR stands on.
• Acute and procedural anxiety. Closely related to distraction — VR lowers anxiety before and during procedures (dental, pre-operative, paediatric). The signal is consistent, if smaller and noisier than the pain data.
• Exposure therapy for phobias and PTSD. A separate, mature field. VR exposure is well-studied and works because it does what in-vivo exposure does, with control over the stimulus. This is among the most clinically validated uses of VR anywhere.
• Rehabilitation engagement. For stroke, balance and motor rehab, VR tends to match conventional therapy on outcomes while beating it on adherence — patients do more repetitions because it's less tedious. The honest reading: it's rarely a magic improvement, but engagement is a real clinical lever, and dose matters in rehab.

Where the evidence is thin or oversold

Being the people who build this, we'd rather say it than have a clinician say it for us:

• Chronic pain is much harder than acute. Distraction is a short-window mechanism. Claims that VR durably treats chronic pain rest on far weaker, smaller, shorter studies than the acute-pain literature. Treat them with caution.
• The novelty effect is real. Some of the early benefit is "this is new and absorbing." Studies with short follow-up can't tell durable therapeutic effect from a wow that fades. Long-term data is the thing the field is still short on.
• Heterogeneity makes meta-analysis messy. "VR" in one trial is a phone in a cardboard holder; in another it's a tethered headset with hand tracking and a bespoke clinical scene. Pooling them hides as much as it reveals. The content and the dose matter as much as the modality.
• Blinding is genuinely hard. You can't give someone a placebo headset they don't notice. A lot of the noise in the literature is this, not fraud — but it means effect sizes deserve a skeptical read.

How a medical device has to prove it — not just assert it

This is the part most VR-wellness marketing skips. Under MDR 2017/745, a medical device can't claim a therapeutic benefit on vibes. It needs a clinical evaluation: a documented, ongoing argument that the device does what its intended-use statement says, built from the published evidence for the mechanism plus your own data, and kept alive through post-market clinical follow-up.

Practically, that forces a discipline on you that a consumer app never faces. Your claims have to be narrow enough to defend. "Reduces procedural anxiety as a non-pharmacological adjunct" is a claim you can stand behind with literature and measurement. "Cures pain" is not — and a regulator will not let you write it. The certification is, in effect, a filter that strips the overclaiming the rest of the category indulges in. We wrote about that whole road in what it takes to ship a Class I medical VR app.

Measuring outcomes that actually count

If you want to know whether it works for your patients, you measure — with the same instruments clinicians already trust, not bespoke in-app scores designed to look good:

• Pain: a validated scale (VAS / NRS), before, during and after — not a vibe.
• Anxiety: a standard instrument (e.g. STAI) rather than "patients seemed calmer."
• Rehab: function and adherence — repetitions completed, range of motion, drop-off rate — because engagement is the mechanism, so it's the thing to track.
• Comfort: simulator-sickness scores, because a therapy that makes 1 in 5 patients nauseous is not a therapy. We covered that design problem in reducing VR motion sickness in patients.

Why we built RVer to the device bar

We could have shipped therapeutic VR as a wellness app and skipped the dossier entirely — most of the market does. We didn't, because the evidence question is exactly where that market loses clinicians. A nurse deciding whether to use VR for a dressing change is not impressed by a slick demo; they want to know what it claims, what backs the claim, and who is accountable if it goes wrong. A Class I certification answers all three in a language a hospital recognises. The evidence base and the regulatory file are not overhead — they're the difference between a gadget and something a clinician will actually put on a patient.

The honest bottom line

VR therapy works best where it has the strongest evidence and the clearest mechanism: distraction from acute and procedural pain, anxiety in the same settings, exposure therapy, and engagement in rehab. It is weakest where the claims get broadest and the follow-up gets shortest. The useful question is never "does VR therapy work?" in the abstract — it's "what, specifically, is this product claiming, and what backs it?" If a vendor can't answer that crisply, that's your answer. For more on the clinician's side of this, see what clinicians actually need (and what vendors oversell).