Mental health has been a pioneering area for digital therapeutics and a keen area of focus for digital health in recent years. Digital solutions enable wider access to evidence-based mental health assessment and treatment.

The rising prevalence of mental health conditions globally has fuelled the growth of digital biomarkers in the field of mental health.

As per the latest industry reports, the global market for digital biomarkers is growing, with its value estimated to reach $16,174 million by 2028. Pharma, biotech companies, and investors alike are recognising the potential benefits of continuous, passive data collection in disease detection and management.

‍

What are Digital Biomarkers?

Digital biomarkers are objective, quantifiable data that is collected using digital devices and analysed to provide insights into an individual's mental health status. They can capture near-continuous and unobtrusive measurements of behaviours associated with mental health symptoms outside of a clinical setting. This data offers a promising solution for preventive, personalised care and remote monitoring in mental health care. 

‍

Which digital biomarkers have the most potential in mental health?

Attendees at our recent HealthXL meeting on Digital Biomarkers for Mental Health Evaluation and Management believed speech and text analysis have the most potential. Others exhibiting potential in this space include sleep, heart rate variability, social media activity, and physical activity. 

‍

Speech and Text Analysis

This biomarker measures changes in digital speech patterns, such as speech rate, intonation, and pronunciation. It can provide unique insight into early signs of cognitive impairment that would be inaudible to the naked ear.

Companies like Kintsugi, EVOCAL Health, Ellipsis Health, and Sonde Health have used AI-powered voice analysis to develop digital biomarkers to detect signs of neuropsychological disorders such as depression and anxiety.

The Ellipsis Health App has shown feasibility in using voice recordings to screen for depression and anxiety among various age groups.

‍

Sleep Quality

Regarded as one of the most mature digital biomarkers, changes in sleep patterns can provide early warning signs of mental health conditions. Sleep is a universally measurable endpoint and is captured by wearables and smartphone apps, which makes it scalable. Wrist-worn accelerometers have been shown to detect sleep patterns associated with psychiatric conditions.

Though a gap between medical and consumer-grade products remains, the data that consumer sleep technology can collect is continuously increasing. Research has shown promising results for the prediction of mood state and mood episodes when AI is applied to consumer-grade sleep data. The evolving multisensory capability of consumer sleep technologies holds great potential in disease detection and monitoring of mental health.

‍

HR Variability

Heart rate variability (HRV) is a promising biomarker for mental health resilience. It provides an objective and quantifiable measure associated with an individual's ability to adapt to stressors.

HRV can provide early warning signs of mental health conditions, and also act as a marker of recovery and progress during treatment. Biofeedback devices incorporating HRV act as an educational technique to help patients manage symptoms of anxiety. Similar to sleep, consumer devices measure heart rate variability. Research has shown promising results for using consumer-grade wearables for heart rate variability. For example, the Apple Watch accurately measures HRV for stress monitoring.

‍

Social Media and Online Activity

This biomarker measures changes in social media activity, such as the frequency and timing of posts, number of followers, and engagement with content. These factors can be indicative of behaviours associated with mental health symptoms.

For example, an individual posting on social media platforms throughout the night may be suffering from insomnia, which is indicative of depression. An analysis of Twitter posts between March and May 2020, could detect pandemic-related increases in coping mechanisms. Similar methods have been employed to assess the effects of psychiatric medications. Tracking online activity may provide greater transparency into psychiatric symptoms than self-report surveys.

‍

Physical Activity

Physical activity has been shown to serve as a digital biomarker in the passive monitoring of depression and anxiety. It is measured using digital technologies like wearables, actigraphy, and GPS technology.

Research shows that indicators such as activity time, level, intensity, movement speed, and step counts are significantly correlated with depression. GPS data has recently offered insights into the relationship between reduced mobility and poorer mental health during the COVID-19 pandemic. In patients with schizophrenia, less GPS mobility is associated with greater negative symptom severity and motivational deficits.

It is clear that digital biomarkers will play a vital part in the assessment and treatment of mental health issues. Yet, key challenges remain for their adoption and scaling.

‍

Challenges in the Adoption of Digital Biomarkers in Mental Health

Low clinical validation

Clinical validation of digital biomarkers in mental health remains remarkably low despite its potential benefits. Validating digital biomarkers presents several challenges such as providing accuracy, consistency, and relevance to clinical outcomes in mental health.

Developing digital biomarkers requires the collection of large diverse data sets at scale, and over long periods of time, to provide consistency and equity in model predictions. Models need to be exposed to different data types, to account for differences in sub-populations' behaviours. This process would be challenging for a single longitudinal study or clinic to collect. 

Researchers in the mental health space have called for an open data source to be created. The source must pool de-identified data to enhance the ability to create and validate digital biomarkers in mental health. Furthermore, identifying appropriate control arms for validation studies of digital mental health solutions remains a challenge.

Reportedly, just 2% of apps on the marketplace leverage smartphones' full capabilities in digital phenotyping. Further research is needed to understand the way in which these capabilities can be deployed clinically in mental health, to validate clinically relevant measures, and to prove acceptance of such solutions among people with mental health conditions.

‍

Vague regulation for digital devices in mental health

Current regulatory pathways were developed for traditional medical devices. But with the rapid and constant evolution of digital devices, regulatory approaches need to evolve alongside them or risk becoming a barrier to improvements in care.

The FDA approves digital therapeutics that meet the definition of Software as a Medical Device (SaMD). The De Novo or 510k pathways regulate these. Solutions that pose only a low risk to patients, may qualify for 'enforcement discretion' and can be made available without active FDA oversight. 

With so few predicates existing in the mental health space, the suitability of the 510k clearance for solutions of this nature is uncertain. Further, the quality of trial design and results for a digital mental health device approved in the De Novo Pathway has been questioned previously. Solutions with similar functions can qualify for different regulatory pathways, based on their intended use cases, indicating room for interpretation in where categorical boundaries lie. A well-defined framework for regulating digital biomarkers may help to accelerate validation and adoption in mental health.

‍

Data privacy and ethical concerns

Digital biomarkers have the capability to capture a large amount of personal information, and in mental health, this poses several legal and ethical concerns. Psychiatric patients may be particularly cautious regarding sharing personal data about their mental health condition, due to concerns stemming from issues of stigma and discrimination. The idea of being ‘monitored’ may not sit well with certain patient populations. Regulations for digital biomarkers for mental health that safeguard privacy and data protection principles help establish trust among all stakeholders.

‍

What is the future of digital biomarkers in mental health?

Opportunity to expand our understanding of mental health conditions

The evolving multimodal nature of passive data presents an invaluable opportunity to expand our holistic understanding of the experience of people with mental health conditions. It has the potential to widely impact patient outcomes through earlier identification of mental health conditions at scale.

‍

Personalisation of mental health care with evolving AI, ML technologies

Advancements in AI and ML models will become a crucial factor in accelerating the development of clinically relevant digital biomarkers in the future.

An emerging application of digital biomarkers in mental health is the provision of closed-loop mental health interventions, using 'just-in-time-adaptive-intervention'. So a device may infer low mood in the context of social isolation and offer a relevant intervention. But it may infer low mood in the context of poor sleep and recommend another intervention.

Although in its infancy, the personalisation of mental health care might be the future of digital biomarkers in mental health.

‍

Opportunity for Big Tech to drive growth?

The global wearable and smartphone market is set to reach a combined $706 billion by 2030. The amount of passive patient data will also increase exponentially with this growth. So what do tech companies like Apple, Google plan to do with this data and the technology?

Their evolved data collection and analysis methods could make them big players in digital biomarkers for mental health. But clinical validation and regulatory concerns still limit their use in medical research and practice, and to date, digital biomarkers have not been a major focus area for these companies.  

Will the popularity of consumer health companies force medical technology companies to develop more convenient and affordable products for passive data collection? Or will the result be a strategic partnership between them to create and validate consumer-friendly products suitable for medical purposes? It will be interesting to see if and how they play a role in the advancement of mental health care going forward.

‍

For a wider look at Big Tech in Digital Health, download our free report.