Novel treatments targeted at functional outcome in early psychosis are important as functional recovery is less common than symptomatic recovery (1), and the two are not always intrinsically linked (2). Continued functional impairment has significant impact both personally and economically and is often rated as the most important treatment goal in people experiencing their first episode of psychosis (3). One treatment approach targeting functional outcomes broadly is social cognition remediation therapy. Social cognition, including the domains of emotion recognition, theory of mind, and attributional bias, is often poor in both established psychosis (4) and first episode psychosis (5), is strongly related to social functioning (6), and contributes unique variance in predicting functional outcome (7, 8). Although varied in approach, social cognitive remediation programs all appear to improve social cognition measures (9). Global approaches that deliver both social cognition skills training and real-world application of these skills also appear to improve functional outcome at least in established psychosis (9). In a previous study, we adapted one of these established global type social cognition programs, Social Cognition and Interaction Training (10), in an Australian sample of first episode psychosis patients (11). We found in a small pilot study that this group intervention was acceptable to young people, but feasibility was low as consent rate and maintaining attendance in the group was relatively poor (11). For some young people, the thought of attending traditional individual or group therapy provoked considerable anxiety (11). In this sense, we were aware that our intervention was not reaching a wider group who might benefit from social cognition training, especially as poor social cognition is associated with anxiety and negative symptoms (12).

Over the last few decades, technological approaches to delivering healthcare have developed rapidly. This is also the case in mental health problems where the use of apps and social media sites has been used to encourage engagement and widen the reach of interventions to other populations (13). Although emerging, the use of technology to deliver interventions has received relatively little attention in the psychosis field, perhaps due to concerns regarding young people with psychosis not using technology or that technology may be unhelpfully incorporated into their symptoms (14). However, there is no consistent evidence to suggest either of these concerns are the case (13, 15).

Virtual Reality (VR) interventions are showing promise in engaging and helping young people with psychosis often by providing safe exposure to challenging environments. Virtual world platforms are being used to deliver physical and psychological treatments to people with a range of health problems, including obesity, autism, intellectual disabilities, and diabetes (16–19). A virtual world is an online shared community environment where members can interact in a custom-built, simulated world. The community interacts in the simulated world using either text-based, 2-D and/or 3-D graphical models called avatars (20). Although under-researched, virtual world platforms have the potential to reduce communication barriers and improve access to support and treatment in people with mental health disorders (21).

One of most widely used virtual world environments is the online and freely available social world Second Life©. Second Life© has been used as a virtual world platform in healthcare education, delivery, and engagement and has considerable potential in delivering psychological therapies (21, 22). For example, Second Life© has been used to treat social anxiety and has been found to be feasible and acceptable with effects sizes potentially comparable to face-to-face Cognitive Behavior Therapy (CBT) for Seasonal Affective Disorder (23) and in treating social cognition problems in autism (24). However, we know of no other study that has used Second Life© in the treatment of psychosis. We adapted a traditionally face-to-face delivered social cognition and interactional training (SCIT) (25, 26) to a virtual world in Second Life© (SCIT-VR) using a codesign process involving young people, clinicians and web designers. This work is described in detail elsewhere (27). The title of this trial is “virtual reality as a method of delivering social cognitive therapy in early psychosis (VEEP).” Our aim was to test the feasibility and acceptability of the SCIT-VR intervention approach in a group of young people who were recovering from a first episode of psychosis.

We conducted a proof-of-concept nonrandomized study of an established group social cognition intervention (SCIT) adapted for delivery through a virtual world (Second Life©), SCIT-VR. We found that the intervention was generally feasible as the intervention completion rate was reasonable, and the delivery of the intervention by relatively inexperienced clinicians in an NHS setting was possible. Our completion rate is similar to that of the trial conducted by Bartholomeusz et al. (11) where 75% of participants completed the intervention. The intervention was mostly considered acceptable as measured by attendance rate, post-session feedback questionnaires and qualitative interviews. There were no concerns about safety either from adverse events or from individual participant interviews. There was a suggestion that the intervention had face validity as pre-post changes were observed on measures of social cognition and anxiety. There was no change in participants' intention to change or self-efficacy scores. The intervention was not primarily aimed at improving intention to engage in further therapy or social activities but we anticipated this might be a beneficial effect of the approach. Further studies might want to consider the use of virtual worlds as interventions aimed more directly at engagement and self-efficacy as part of a pathway to accessing more structured and face-to-face interventions.

The small scale pilot face-to-face version of this study was conducted by Bartholomeusz et al. (11). The average effect size score in their trial was d = 0.29, compared to our g = 0.22. Similarly this trial also found significant improvements on emotion recognition. This is not completely unexpected as we believe the virtual world is a mechanism to provide the SCIT intervention to those who might find it difficult to attend face-to-face therapy. The findings in the current study across a number of measures indicate that the virtual world intervention we have developed has the potential to improve social cognitive domains in those people with early psychosis. However, this is a small pilot study with a small sample size. Therefore, a large randomized controlled trial is required to determine whether providing SCIT in a virtual world can lead to significant improvements in this population.

This is the first study to our knowledge using a virtual world to deliver a structured group therapy in early psychosis. It was relatively ambitious, as a number of social cognition skills would appear to rely on face-to-face social contact to achieve mastery. However, there have been previous studies that have used Second Life© successfully with the aim of improving social cognition skills in those with high functioning autism (24). We believe that the didactic teaching approach that was delivered in Second Life© was possibly more successful than more group-focused sessions as measured by higher levels of feedback on the individual sessions that were more didactic in nature. The sessions with more specific teaching content were rated more highly by the participants than the sessions that involved group work. No participants highlighted that they enjoyed the group or specifically felt part of the group. Future studies using this approach may want to consider whether the intervention needs to be delivered in a group setting. Others, for example, have successfully used an individual approach when using virtual worlds to treat social anxiety (23).

Participants reported a high degree of presence in the virtual environment, which was surprising as the interaction with Second Life© is purely through a computer monitor and headset and not using a head mounted display or more traditional immersive technology and the interview themes also highlighted that people felt high levels of presence in the virtual world. However, the presence questionnaire was completed by a minority of participants. Previous research has attempted to assess the sense of presence using Second Life©; for example, nursing students reported relatively high degree of presence in Second Life©, which was related to attitude towards using the environment (45). We have considered whether a more immersive environment such as traditional VR or using 360 videos might lead to a greater acquisition of social cognition skills as they more approximate real-life situations.

These findings support the growing number of VR interventions for people who experience psychosis. According to a recent literature review, there are a distinct lack of studies examining the potential impact of VR interventions for early psychosis (46) Therefore, our trial is timely. The key advantage of VR is that real-time behavior can be observed when interacting in the VR world. Furthermore, the environment can be modified and controlled to elicit certain responses (47).

There are a number of limitations of the study. Firstly, there was no control group and therefore we cannot be sure that any improvements are due to the intervention. There were challenges in recruiting patients, which impacted on the feasibility of the study, with only 36% of eligible patients consenting to take part in the intervention. This differs from previous research, which showed that two thirds of patients with psychosis were willing to be contacted about taking part in research (48). However consent rates for those eligible participants diagnosed with psychosis differ between trials. For example, while Wood et al. (49) reported a 67% consent rate (30/45), Kanniston et al. (50) reported a 33% consent rate (1139/3417), and Schrank et al. (51) reported a mean consent rate of 40.0%. One possible reason for our recruitment rate is that some potential participants may not have had access to the internet or may have had limited digital skills, reflecting an ongoing concern around digital exclusion for some people. This issue is a concern more generally in digital mental health studies and requires further consideration in finding ways to reduce digital inequalities (52). More frequent contact between clinicians and researchers has been shown to improve recruitment. Therefore this should be reviewed in future research (53).

Whilst we had a reasonable retention rate, we had problems keeping people engaged in the study. There were some problems with the technology during some of the sessions which required the secondary therapist to rectify; on two occasions this was not possible. This was highlighted in the qualitative interviews. One of the limitations of using Second Life © was that it was not functional on tablets or hand-held devices and it could not be streamed through a gaming console. A number of participants did not own a laptop or desktop computer. We were able to loan a computer to participants who did not have access to one, but for those who do not own or have access to a computer participation in an intervention delivered on a platform such as Second Life© is compromised.

Given the results of this small proof-of-concept study, there is sufficient evidence to consider a full-scale efficacy trial in this population. Challenges in designing such a trial are whether to consider adjunctive experiences that might be more immersive such as combining other VR approaches. Based on the feedback, we would also need to consider whether all parts of the intervention need to be delivered in a group setting and how we might be able to overcome technical problems and access to those without a computer. The delivery of evidence based psychological therapies to this group continues to prove challenging, and the use of technologies such as Second Life© shows promise in widening access and engagement for this hard to reach group.

The datasets generated for this study are available on request to the corresponding author.

This study was carried out in accordance with the recommendations of the 'West Midlands – Solihull Research Ethics Committee' with written informed consent from all subjects. All subjects gave written informed consent in accordance with the Declaration of Helsinki. The protocol was approved by the 'West Midlands – Solihull Research Ethics Committee'.

AT, SB, MB, IV, and DT wrote the grant application for this feasibility and pilot trial. AT, AR, DT, and FE were responsible for coordinating the design of the virtual world. AT, AR, and FE conducted the co-design workshops. AT, FE, FL, and AR undertook the trial. AT drafted the manuscript. All investigators have been involved in revising the report, and all authors have seen and approved the final version.

The VEEP study was sponsored by the University of Warwick and supported by the PsyIMPACT programme of MQ: MQ16PI100011. Transforming mental health (Register charity in England/Wales: 113991 & Scotland: SC046075).

SB is a director of a not-for-profit company aimed at commercialising mental health apps.

The remaining authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.