Why Patients Tolerate VR Exposure Better Than In-Vivo

The Real Clinical Problem: Dropout Rates in Traditional Exposure Therapy

Exposure therapy is one of the psychological interventions with the strongest empirical support for anxiety disorders and related conditions. Meta-analyses have shown large effect sizes in specific phobias, social phobia, and panic disorder with agoraphobia. In real clinical practice, however, some patients who would be candidates for treatment never start it, another group starts but drops out before completing the protocol, and a third group completes the protocol but does not benefit from treatment to the expected degree.

The literature places dropout rates in classic exposure therapy within ranges that vary by disorder, country, and study design. What matters to the clinician is not the exact figure but the pattern: dropout is not marginal. The effective population-level efficacy of a treatment depends not only on its efficacy for those who complete it, but on what proportion of potential patients actually comes to benefit from it.

In this article we review the psychological mechanisms that appear to mediate better tolerance of virtual reality exposure, the documented patient preferences when given a choice, and the practical consequences this has for clinical workflow and dropout prevention.

Three Psychological Mechanisms Behind Better VRET Tolerance

The first mechanism is perceived control. In-vivo exposure, especially when it involves traveling to an external environment (a park with dogs, an airplane, a tall bridge), represents for many patients an irreversible commitment: once there, leaving is costly, visible, and sometimes socially embarrassing. VR exposure, by contrast, offers immediate reversibility: the patient knows they can remove the headset within seconds, and that certainty reduces the catastrophic anticipation of being trapped in the feared situation.

The concept of perceived control is well established in the literature on anxiety. Models such as Barlow's identify it as a central variable in generalized psychological vulnerability. Reducing the sense of uncontrollability during the treatment-approach phase makes it easier for the patient to agree to start exposure. Paradoxically, that same perceived control seems to help the patient stay in the virtual environment longer than they would tolerate in vivo, because the subjective cost of continuing is lower.

The second mechanism is fine-grained, reproducible gradation. In in-vivo exposure, the hierarchy depends on whatever the real environment happens to offer at that moment: a particular dog, a plane taking off that day, an available height in a given building. In VR exposure, the clinician controls the stimulus with precision: the dog's size, distance, behavior, the presence of human figures, ambient sound. The hierarchy can be graduated in smaller, more reproducible steps, avoiding the overly abrupt jumps that in-vivo exposure sometimes triggers, which can push anxiety responses beyond the window of tolerance.

The third mechanism is the privacy of the therapeutic setting. In-vivo exposure, in certain cases (social phobia, agoraphobia), means being exposed in public, which adds a layer of anticipatory shame. VR exposure takes place in the consulting room, in private, which lowers that barrier. For patients with social anxiety this is especially relevant: a first presentation to a simulated audience in the office is far more tolerable than a real presentation, and it allows the habituation foundation to be built before moving on to the public setting.

Documented Preferences: García-Palacios et al. and Patient Choice

A classic study in the field is the work by García-Palacios, Hoffman, See, Tsai, and Botella (2001) on patient preferences for VR exposure versus in-vivo exposure. In a sample of people with a fear of spiders, the researchers presented participants with a hypothetical choice between the two treatments. The majority preferred VR exposure, citing the greater sense of control, privacy, and the ability to leave the environment at any time as their main reasons.

This finding matters because patient preference is not a cosmetic variable. There is abundant literature in general psychotherapy showing that when the treatment offered matches patient preference, adherence and outcomes improve. In exposure therapy, where initial treatment refusal is one of the main access barriers, offering a modality with a lower barrier to entry can translate into more patients agreeing to start treatment.

The Valencia research group (Botella, García-Palacios, Baños) has underscored this point in later publications: VR exposure is not necessarily more effective than in-vivo exposure for those who complete it, but it widens the pool of patients willing to start and finish treatment. This matters in terms of public health and real-world clinical management.

Sufficient Emotional Response Without Overwhelm

An important caveat. Better VRET tolerance does not mean that exposure is somehow "softer" in a therapeutic sense. For exposure to work, the conditioned fear network must be activated sufficiently to allow the learning to be updated. If VR were so mildly activating that the patient did not experience real anxiety, there would be no clinically useful exposure.

The literature on presence (Slater and Wilbur, Diemer et al., Schubert) shows that when a scenario evokes sufficient presence, the autonomic and subjective response reaches levels comparable to real exposure. The difference between VR and in-vivo exposure, then, lies not in the intensity of activation but in the patient's relationship to that activation: in VR, the patient experiences significant anxiety but with a sense of control that allows them to sustain it rather than flee.

This balance is delicate and must be monitored. A poorly indicated VR session can produce insufficient activation (low presence, a cognitively distanced patient), in which case exposure loses its therapeutic value. A well-indicated session produces the optimal range: significant clinical activation, sustained for as long as needed for habituation, without exceeding the window of tolerance.

Practical Consequences for Clinical Workflow

Better VRET tolerance has several operational implications. First, in the initial presentation of treatment: when a patient with dog phobia, fear of heights, or social anxiety hears the proposal of in-vivo exposure, an initial refusal response is common. Offering a first VR phase lowers that barrier and allows the work to begin. The clinical conversation shifts from "I'm going to have to face the situation" to "I'm going to start in a controlled environment and we'll move forward together."

Second, in managing the hierarchy: the fine gradation of VR allows building intermediate steps that would be difficult or impossible in in-vivo exposure. In dog phobia, for example, a "calm medium-sized dog at eight meters" is a situation that is hard to reproduce reliably on the street. In VR it is a controlled parameter that can be used across several consecutive sessions if the patient needs more time to habituate.

Third, in dropout prevention: the initial sessions are critical. If the patient perceives treatment as manageable, they complete more sessions; if they perceive it as overwhelming, they drop out. VR helps build that initial perception without sacrificing the clinical activation needed for treatment to work.

Fourth, in the transition to in-vivo exposure: VR should not necessarily be understood as a complete substitute for in-vivo exposure. In many indications, the hierarchy culminates in real exposure (visiting relatives who have a dog, attending a work meeting, riding a real elevator). VR builds the habituation foundation that makes that final step viable.

Limitations of This Pattern and Equity Considerations

The argument for better VRET tolerance has nuances that the honest clinician must consider. First, not all patients prefer VR. Some prefer in-vivo exposure from the start, whether out of personal preference, distrust of technology, or practical considerations. The choice must be informed and respected.

Second, some patients do not tolerate the headset well for specific reasons: persistent cybersickness, vestibular problems, sensitivity to residual latency. In these cases, the theoretical improvement in tolerance does not materialize, and it is best to return to traditional exposure. The headset is not universal.

Third, the improved tolerance is an average effect. Any clinical intervention has heterogeneity of response. The argument should be read as "on average, a significant proportion of patients tolerate VR better," not as "all patients tolerate it better." This nuances the conversation with the patient and avoids creating absolute expectations.

Fourth, accessibility. VR exposure requires infrastructure (a headset, office space, software) that not every clinician can offer. The argument for improved tolerance is not a criticism of those who do good in-vivo exposure work; it is complementary information for those considering expanding their repertoire.

Fifth, there is a risk of misusing the tolerance argument: using it to justify prolonging VR work when transitioning to real exposure would already be appropriate. If better tolerance becomes a subtle form of clinical avoidance ("the patient is more comfortable in VR, so we keep them there for more sessions"), the net effect can be a suboptimal intervention. Improved tolerance is an advantage for entry and adherence, not a therapeutic destination. The treatment plan must include the transition to the real stimulus when the indication calls for it.

The Link to Real Clinical Outcomes: More Adherence, More Results

The final argument worth articulating is the full causal chain. Greater perceived control and better tolerance translate into less dropout. Less dropout translates into more completed sessions. More completed sessions translate, ceteris paribus, into greater effective exposure. Greater effective exposure, in patients who respond well to the habituation mechanism, translates into better clinical outcomes.

This does not mean that VR is automatically more effective than in-vivo exposure among those who complete both treatments; meta-analyses (Powers and Emmelkamp 2008, Opriş 2012, Carl et al. 2019) suggest equivalence for those who finish. It means that, at the level of effective population health, VR can increase the total number of patients who benefit from an exposure intervention because it widens the pool of those who start and sustain it.

This is an operationally important argument for a psychologist making decisions about their practice: the efficacy that matters in the end is not that of the theoretical patient who completes the perfect protocol, but that of the real patient who walks into the office, decides whether to start treatment, and completes — or does not complete — the necessary sessions.

Conclusion: Tolerance Is a First-Order Clinical Variable

We have reviewed the psychological mechanisms (perceived control, fine gradation, privacy) that appear to explain better VRET tolerance, the documented preferences from the García-Palacios and Botella research line, and the practical consequences for clinical workflow and dropout prevention. Tolerance is not an aesthetic nicety: it is a variable that determines how many patients actually benefit from the intervention.

To go deeper, see our article on comparative efficacy of VRET and in-vivo exposure, our piece on presence and immersion in clinical VR, and our case notes on clinical progress with VRET. To see how all of this translates into concrete scenarios and tools, you can book a guided demo with the clinical team.

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.