Can Online Exposure Therapy Work? The Clinical Limits

Can online exposure therapy work?

The honest answer is a qualified yes, and the qualification carries all the weight. The trouble is that the label “online exposure therapy” covers three different practices at once. Each has its own clinical standing, and the results they produce bear little resemblance to one another. They are worth separating before anyone answers, because the confusion is not innocent: whoever sells the third trades on the reputation of the first.

The first version is a session directed by a psychologist at a distance, almost always over video. It is telepsychology applied to one specific procedure: the clinician is present, listens, adjusts the challenge and decides when the trial closes. The second is self-guided material — worksheets, monitoring records and between-session tasks that the patient works through alone and a professional reviews afterwards. The third is the class of apps that sell exposure exercises straight to the public, with no registered practitioner answering for the case.

Their standing is not equivalent. The first has recognised footing and forms part of ordinary practice. The second occupies a bounded place inside a plan that somebody else is directing. The third is not treatment, however it is advertised. Only the first deserves the name online exposure therapy in any clinical sense. Where a practice already works with environments for exposure to feared social situations, the difference is plain enough: either the stimulus is calibrated in real time, or it is left to the judgment of the person who fears it.

So the yield of online exposure therapy is not settled by the channel. It is settled by who holds the clinical decision in each trial. The channel changes the logistics of the appointment. The competence framework changes the outcome. That distinction organises everything that follows.

What changes in the mechanism when the session runs through a screen?

Answering that with any rigour takes a model of the mechanism, and the current model is not habituation. Craske and colleagues framed it as inhibitory learning: extinction does not erase the original fear association, it builds a second association that competes with it. The 2022 update shifts the emphasis onto the retrieval of that new association. What is therapeutic is that it becomes available at the moment it is needed.

From there follows the criterion that organises the whole of practice. What produces learning is expectancy violation — the gap between what the patient predicts and what actually happens. It is not produced by how long they endure, nor by their anxiety falling inside the session. Foa and Kozak's emotional processing model held the opposite reading for decades and still describes a real phenomenon. But a decline in arousal within the session predicts poorly where the patient will stand months later.

With that criterion in hand, online exposure therapy can be judged without rhetoric. For expectancy violation to occur, somebody has to have collected the prediction beforehand, in words and as a number. At a distance that forces the prediction to be said out loud, because the clinician can no longer see the room. And safety signals multiply in the patient's own home with nobody writing them down: the door left ajar, the relative in the kitchen, the phone held in one hand.

The remote format also offers something the consulting room cannot. Craske lists variation of context among the optimisation strategies, and the patient's home is a legitimate context — sometimes the only one that matters. The technical build-out of that idea with virtual reality, with its requirements and its ceilings, is set out in the article on VR exposure delivered remotely. What concerns this article is a different question: who decides.

Is supervision by a psychologist necessary?

Yes, and the reason is technical rather than guild-protective. The guidelines put it soberly. NICE frames exposure as a component of a cognitive behavioural treatment, not as a resource dispensed on its own. Maples-Keller and colleagues describe a field that has grown in the hands of clinicians rather than in self-administration.

It is worth enumerating what the professional actually contributes, because the word support sounds vague and is nothing of the kind:

• Formulates the prediction. Without an explicit expectancy there is nothing to violate, and the trial reduces to endurance.
• Designs the trial that disconfirms it. Chooses the stimulus, the order and the duration so that the result carries information.
• Withdraws the safety signals. They have to be spotted first, which is the hard part.
• Decides the progression. When a step goes up, when it repeats, and when to stop.
• Holds the record. What is written down is what allows the case to be reviewed and answered for.

None of those five operations is performed by a text, and none by a program either. Well-written material informs, but it does not read the prediction of the person reading it. That is why online exposure therapy keeps its value while the clinician stays inside the loop, and loses it the moment the patient is left alone with the instructions. Anyone looking for criteria to choose formal training in VR exposure will find that route described in detail.

There is a point the beginner walks straight past. Professional support does not consist of being connected; it consists of deciding. A video call in which the clinician watches and never intervenes is worth what a recorded video is worth. Presence without decision is not clinical oversight: it is company.

What qualifications authorise a clinician to deliver exposure?

The framework is stricter than the online exposure therapy offering that circulates on the internet suggests. Health-care regulation reserves clinical intervention to practitioners holding a qualifying credential [CITATION TO VERIFY]. In Spain, for instance, there are two routes into it: the health-care psychologist qualification, or the clinical psychology specialty obtained through residency.

To that is added registration with the relevant professional body — in Spain, the regional college of psychologists — with its indemnity insurance and its code of ethics. Registration is not a decorative seal. It is what indicates who is to be held to account when something goes wrong, and what subjects the practice to a disciplinary regime. Once the appointment crosses a regional or national border, that point stops being trivial.

Several profiles therefore fall outside, and they are precisely the ones that turn up attached to exposure material: psychology graduates without the health-care qualification, therapists holding unregulated training, and programmes offered straight to the public. A practitioner working in solo private practice answers exactly as a large clinic does, because the credential does not change with the size of the office.

The framing takes in data handling as well. An appointment held at a distance generates records, sometimes recordings and, where virtual reality is involved, session telemetry. The controller of all of it is the professional, not the tool being used.

None of this is settled by the equipment supplier. A system can require credentials, leave an audit trail and make the report easier to produce, and that helps; but authorisation to practise is not granted by a program. The person who signs the clinical notes is the registered practitioner seeing the patient. That asymmetry explains why a self-administered intervention does not become treatment by refining the software that serves it.

Can a person do exposure therapy on their own?

The question deserves an honest answer rather than a flat no. A person can approach what they fear on their own, and does so daily. What they cannot do is ensure that the approach produces inhibitory learning, because that depends on conditions which are not visible from the inside.

The typical failure of self-administration is not a lack of courage. It is escape at the wrong moment. When distress climbs, withdrawing relieves it at once, and that relief reinforces the withdrawal. The trial ends up confirming the original prediction instead of disconfirming it. An outside observer spots that in seconds; the person living it almost never does.

The second failure is adherence with nothing holding it up. Self-administered programmes report dropout more often than guided ones, and dropout is not distributed at random: the people who needed it most tend to leave earliest [CITATION TO VERIFY]. Self-guided material pays off when it arrives wrapped in a plan, with somebody who asks about it. The mechanisms behind dropout in exposure are described separately and in more detail.

There is a point about the evidence worth stating plainly. The protocols that underwrite exposure were evaluated with professionals delivering them, and human guidance tends to account for much of the difference between one self-administered programme and another [CITATION TO VERIFY]. So the word self-administered describes a route of delivery, not an equivalent grade of evidence.

There is a reasonable use, and it is the one that sustains everyday practice. The patient carries out tasks between appointments that the psychologist has designed, records what happens and brings the record back. That is self-administration with professional support, and it bears no resemblance to online exposure therapy bought in an app store.

How is the format decided in each case?

Format is not chosen by preference but by case. Neither online exposure therapy nor the in-person version is the starting point: the patient is. These are the axes that order the decision in practice:

• Who holds the trial. If nobody collects the prediction or decides the progression, the channel is the least of it.
• Which stimulus is needed. Some themes the patient's home cannot supply, and some only the home supplies.
• How much margin remains if the trial overshoots. At a distance the margin is narrower, and that weighs on the indication.
• What ends up on record. A format that leaves no usable clinical note is no use for answering for the case.
• What has been agreed about the work between appointments. With no written instructions the task gets improvised and stops being comparable from one week to the next.

The limits of the remote format are well known and no amount of bandwidth resolves them. On comparative efficacy, Carl and colleagues do not find virtual reality exposure falling below in vivo exposure. That finding is worth reading carefully: it refers to interventions delivered by clinicians, not to material a patient works through alone. Extending it to self-administration would be a leap the data do not license.

In the consulting room the practical conversation almost always reduces to two questions: which part of the route admits a remote format, and which part does not. Book a demonstration to see how a hierarchy is configured inside the system and what is recorded from each trial.

This article is for informational purposes for psychology professionals. It is not clinical advice for any individual case and does not replace the judgment of the licensed psychologist in charge. VRET is professional clinical-support software, not a CE-marked medical device.