The future of digital mental health in the post-pandemic world: Evidence-based,blended,responsive and implementable

Much has been written about the challenges and opportunities for digital mental health brought about by COVID-19. Indeed, the pandemic has led to one of the biggest surges in uptake of digital health interventions, although not always by choice. Yet, this has also been a time of anxiety and there is an undeniable desire to return to our old lives and old ways of working. Hence, it may be an opportune moment to examine the role of digital health interventions in the near future.

Overall, the dramatic changes in how we have delivered care over the last 2 years has increased comfort with technology for many clinicians and patients, at least with telehealth. However, digital interventions are far more than the use of telehealth and recommendations of online information sources. For many mental health clinicians in Australia and New Zealand, digital health care already includes online psychological therapies, apps for tracking mood, shared decision-making tools and ‘smart’ medical record systems that prompt guideline concordance. The World Health Organization classifies the broad array of Digital Interventions for Health into those for clients, healthcare providers, health system managers and data services (World Health Organization, 2018). Here, we primarily examine digital health services in reference to those used for, or with clients, although the other aspects of digital interventions are equally important in implementing and maintaining client-facing interventions.

Digital intervention tools are already commonplace in the self-management of general mental health, and in the clinical care for people with both chronic physical and mental ill-health. This is not surprising given the ubiquitous permeation of digital technologies in most people’s lives in the developed world today. Policy makers have also recognised the need to leverage digital technologies as highlighted by the 2021 Australian Parliamentary report into Mental Health and Suicide Prevention and the 2019 New Zealand Digital Health Strategic Framework. This is at least partly because of the immense workforce shortages in mental health across Australia and New Zealand. When effective, the interventions can decrease the time required of clinicians to deliver care. Highly skilled clinicians can spend their valuable time delivering care that is reliant on the human interaction, rather than routine tasks which do not do justice to their training and experience. Technology in this instance can be an effective ally rather than a feared foe from a dystopian future.

In the future, digital interventions may play an even greater role as technology evolves along with our understanding of mental disorders. Engaging new mechanisms of brain function, cognitive mechanisms and behavioural change through contemporaneous interactions with technologies hold promise for several mental health conditions. For example, core mechanisms such as ruminations can be engaged in real time with promotion of behavioural strategies to counter them. This could be combined with the data gathering ability of commercial wearable devices to deliver interventions concurrently. More nuanced phenotyping made possible by such devices could also tailor or even personalise interventions by targeting subgroups where interventions are effective. These mechanisms are impossible without digital technologies due to the pragmatic limitations of the current clinician-patient interaction. Advances in artificial intelligence (AI) and improved access to technologies could also lead to greater use of virtual reality, chatbots and other emerging applications.

Against these possibilities, we propose some principles for adopting digital interventions in the immediate future beyond telehealth and COVID-19. We propose that

1. Digital technologies should be evidence-based. An important concern regarding many digital technologies has been the lack of evidence regarding their safety, efficacy and effectiveness. From over 1000 publicly available apps focused on wellness and stress management in 2020, only 2% had any research supporting them. Even when present, the quality of evidence was poor. This has been partly related to factors such as rapid evolution of technologies, the slow time-scale of evidence generation, and finding the appropriate niche for testing digital technologies for treating mental ill-health. However, there are several examples of success. The use of Moderated Online Social Therapy (MOST) at the point of discharge from early psychosis services (Horyzons project) was associated with significant gains in vocational functioning, and in preventing re-hospitalisations in a randomised controlled trial (RCT; Alvarez-Jimenez et al., 2021), although the intervention group did not differ from the comparison group with respect to social functioning. Some smartphone apps have also been associated with improvements in depressive and anxiety symptoms in RCTs, although much of this evidence is not among samples with mental disorders and there were limitations in the quality of evidence. Thus, there is a greater need to evaluate digital interventions in high-quality RCTs that consider expectancy effects, rater blinding and attrition. Adequate funding and opportunity for rigorous research is central to improving the evidence base. New models of simultaneous implementation and research, using paradigms such as A/B testing may also be necessary in the fast-paced world of digital technologies.

2. Digital interventions should be blended with clinician interventions whenever possible. Blended interventions refer to the combination of face-to-face sessions with those delivered through digital technologies, usually online. There is emerging evidence for their effectiveness (Erbe et al., 2017), which has been facilitated by addressing the challenges of stand-alone digital therapies, such as uptake, as well as the challenges of face-to-face therapies, particularly maintenance of intervention effects. Furthermore, digital and in-person interventions can act synergistically to enhance their efficacy. For example, digital tools can help with practice tasks to maintain effects of in-person cognitive behavioural therapy sessions, while therapists can support patients to utilise digital information more effectively. In our experience with pilot blended interventions (e.g. ACTRN12621001123831), clinicians and participants have found the interventions to be highly acceptable when adequate time and resources were able to be dedicated to intervention delivery. Our qualitative data indicate that consumers strongly demand blended therapy models that add to rather than replace options currently available to them.

3. Digital interventions should be responsive to patients’ changing needs. A significant challenge for several previous digital interventions has been their static nature, which meant they were unable to respond to patients’ changing symptoms or other needs. With improvements in technology and greater provider comfort, it has become possible to dynamically respond to participants’ changes in mental state. For example, participant self-rating of their moods combined with clinician feedback loops has been effective at preventing relapses to manic or hypomanic episodes (Faurholt-Jepsen et al., 2020). In these systems, participants rate their mood daily on an app, which when falls outside of an expected range triggers a clinician response. In the future, the need for active participant rating can reduce if pattern recognition from mobile phone-use, or other passively sensed data (e.g. movement patterns), can be reliable proxy markers of self-reported mood. Such interventions can substantially decrease the costs associated with traditional models of care where clinicians spend considerable time trying to contact patients to ensure that they are well. While the value of such human contact is not to be discounted, it may be useful to have such digital systems available when clinicians and patients choose to use them and to expand relapse monitoring to a greater proportion of those who need it.

4. Digital interventions should have feasible and scalable implementation plans. As is probably obvious to most practitioners, provision of digital interventions alone is unlikely to lead to patients or clinicians adopting, using, and sustaining their use. This is particularly important given the competing priorities of most patients’ lives, and the limited time available for most clinicians. Clinician training, reminders, embedding champions in multi-disciplinary teams and addressing barriers specific to the local context are important strategies to ensure successful implementation. At a consumer level, better mechanisms to improve engagement using human centred design principles can improve uptake. In our experience with implementation of the aforementioned MOST intervention in Victoria, the use of peer support, coaching, reminders and clinician training were associated with greater use of the platform. Costs of implementation and evidence of cost effectiveness should be important considerations in translation. This will also enable public or private reimbursement systems to invest resources to utilise such digital interventions. After all, it is difficult to imagine clinicians being able to engage in blended or responsive digital interventions if they do not have time and remuneration to perform tasks relating to the intervention.

Despite the possibility of effective, implementable digital tools, it is unwise to dismiss the potential risks of digital interventions. The concerns regarding privacy and confidentiality regarding digital media also extend to digital health interventions. In a cross-sectional study of 36 top-rated apps for depression and smoking cessation (Huckvale et al., 2019), 80% of the apps transmitted user data to Facebook and Google, but only 40% of those apps disclosed this in a privacy policy. It is even more concerning given many users do not read privacy policies for apps. Such data sharing risks are an important consideration given the intersection of commercial and public health interests in the digital space. Therefore, it is crucial to improve privacy protections for apps or other interventions utilised for mental healthcare delivery. Positively, most evidence-based digital interventions in mental health ensure adequate privacy protections given the ethical and regulatory requirements of research. Furthermore, regulatory bodies have begun to act. For example, the Therapeutic Goods Administration in Australia has introduced recent regulation of Software as Medical Devices (SaMD). In this rapidly changing environment, clinicians may have an important role in being aware of the regulations and considering privacy protections prior to recommending apps or other digital interventions. There is a need for clinicians to have greater access to ready information about safe, effective apps and software. The NZ Ministry of Health supports a ‘Health Navigator’ tool which allows clinicians to easily review information on apps. However, the development of this resource indicates that publicly available privacy policies were reviewed rather than following a requirement to submit a privacy protection plan through a regulatory authority.

Many are also concerned about the harmful impacts of digital technologies, especially among adolescents, and the risks of entrenching their screen use by offering digital modalities of care. This is not surprising given the increasing complexity of artificial intelligence systems driving commercial digital media. Given the emerging data on the risk of mental ill-health from the use of social media, it is vital to monitor risks for each patient and at a population level with further research. However, it should be borne in mind that the use of technologies is here to stay. The key is to develop platforms and guidelines for the use of technology designed to be in the service of, and to support people’s wellbeing rather than to harvest attention.

A related consideration in the use of digital tools is the concern regarding the loss of in-person experiences in assessment, engagement, therapy and monitoring. This has been particularly accentuated by the enforced use of telehealth during the COVID-19 pandemic, when many patients and clinicians yearned for the recommencement of in-person sessions. This clearly highlights the importance of the complex and valuable nature of the human interaction which AI may not replace. But this has also clarified the role of digital interventions not as a replacement for all in-person work, but as useful additions and augmentations to the same.

As we transition into a peri-pandemic or more optimistically, a post-pandemic world, it is imperative to have a balanced view of digital interventions in mental health. Effective digital tools, especially when blended and responsive can radically improve the availability of mental health care in our corner of resource-rich, yet manpower-poor world. The whole is more than the sum of its parts; blended technologies embedded within existing care modalities are likely to be better than stand-alone face to face or digital interventions. High-quality research is key to expanding the evidence base for digital interventions, and improving the research funding environment will support the translational pipeline. Improving safety and privacy protections will also allay concerns for both clinicians and patients. Appropriate resources and targeted implementation strategies may be key to greater uptake of effective digital interventions, alongside in-person clinical work.