The Future of Mental Healthcare: Digital Biomarkers, Chatbots, and Blurry Lines
There is a lot of potential to lighten the load of mental health professionals
Global studies have painted a dark picture of the future: Depression and anxiety rates are skyrocketing for Generation Z, the demographic cohort of young people up to the age of 23.
While mental illness is being reported more frequently amongst all demographics, it is daunting to think that 55% of Gen Zers in the US report less than very good mental health. Similarly, depression and self-harm rates are on the rise in the UK amongst younger generations.
The largest hit to our mental health is yet to come: the expected recession after the pandemic. No wonder the mental health tech sector is receiving more investment than ever. But what exactly are these technologies that are likely to help us meet this unprecedented demand?
The divide between self-care and clinical care
To understand future trends, let’s take a glance at the present. The mental health tech ecosystem is already sizeable. Most of the established mental health tech can be classified as telehealth solutions such as Babylon Health, peer-to-peer support networks such as Big White Wall and a host of self-care courses and apps, ranging from meditation exercises, over sleep help to mood tracking¹.
While the telehealth sector is heavily regulated, peer-to-peer and self-care tech solutions pop up at a much higher rate, have a quick iteration cycle, and are rarely scientifically tested.
Naturally, there is a pretty clean separation of prevention/self-help tools from clinical help tools.
Only a few hybrid solutions such as Meru Health are attempting to bridge this gap.
Is this a problem?
In a society where mental illness is rare, a division of self-care and clinical care may be fine. However, we may be facing a future in which every other person or more has mental struggles.
We need a system that begins with prevention from a young age and is able to transition smoothly from one tier of help to another, in case escalation is needed.
We will need a full-stack solution that integrates prevention, self-help, peer-to-peer help, as well as clinical help.
What is the role of technology in such an integrated system? In 2018, the NHS conducted the TOPOL review, an assessment of the digital future of healthcare, including mental healthcare.
Their final report predicts13 overarching trends. Many of these trends suggest a fusion between self-care and clinical care of mental health in the near future. Let’s look at some examples:
Digital biomarkers of mental states
The increasing popularity of smartphones, wearables and sensors opens a new door for mental illness prevention and condition management: the use of digital biomarkers.
“In mental health, digital biomarkers are indicators of mental state that can be derived through the patient’s use of a digital technology.
Commonly cited biomarkers cover physiology (eg heart rate), cognition (eg screen use), behavioural (eg global positioning system) and social (eg call frequency). […] These can be gathered from smartphones and other sensors, as well as social media interactions and usage of various services.” (Topol Report)
Digital biomarkers blur the boundary between prevention and clinical care. These biomarkers are extracted from non-clinical contexts, such social media interactions, device interactions and app interactions and are providing clinical insights.
Given a substantial part of healthcare data is in text form, recent breakthroughs in Natural Language Processing (NLP) are likely to revolutionise the healthcare sector, including mental health.
In NLP, we build Artificial Intelligence applications that work with human language data. Specialised Conversational Agents (chatbots) have become increasingly good at taking some of the workload off humans, for example in the context of support chats.
No question, replacing a human clinical professional with a chatbot is not happening anytime soon. However, chatbots have become pretty good at information extraction, i.e. driving the conversation with a specific objective to learn something about the user.
They have the potential of outsourcing the administrative or trivial tasks of a mental health professional before, during or after consultation, such as identity checking, reporting the history of the present illness, guiding therapeutic exercises etc.
On the other hand, chatbots are already being used as self-help tools for self-awareness, for example applications being Wysa and Woebot. Such applications need to provide a quick route to escalation in case they are dealing with serious mental illness.
Thus, In the future therapists may have to work hand in hand with chatbots that perform administrative or trivial tasks.
Electronic Health Records
Healthcare professionals often spend half of their time documenting patient consultations as electronic health records (EHRs). Why are EHRs so useful? They facilitate research, they prevent errors and they allow interoperability with other clinical as well as non-clinical systems.
In addition, EHRs usually need to be coded with an international standard such as the International Classification of Diseases (ICD). This takes a lot of time and is an error-prone process, often resulting in inconsistent records. NLP can optimise the process of creating EHRs by facilitating:
- digital dictation
- automatic clinical coding
- automatic transcription and summarisation of consultations
A lot of these technologies are still maturing, but progress is fast and they can be expected to make a significant difference in the next ten to twenty years. The benefits are countless:
- patient accessibility of health records (and the capacity to correct them)
- integrations with sensor devices, smartphones, communication platforms
For example, imagine your travel information is automatically added to your EHR and there is no risk of forgetting to mention it. Again, the boundaries between clinical to non-clinical applications are beginning to blur.
The big questions
So what happens if self-care and prevention stop being clearly separable from clinical care? While the benefits of all these trends are obvious, so are the risks.
- How to clinically validate mental health tech, given tech is such a rapidly developing field and clinical validation is such a long process?
- What IS mental health tech in the first place? If any app or platform can collect information reflecting mental states, how to classify mental health tech?
- How to ensure privacy and that information safeguarding standards are met?
- If AI is at play, how to account for bias and cultural sensitivity?
- And how to deal with the issue that financial incentives of mental health tech often do not align with patient wellbeing? In many business models, profit is correlating with patient illness rather than patient wellbeing.
These questions take years to answer. Technology is nowhere near mature enough to replace clinical psychologists, but there is a lot of potential to lighten the load of mental health professionals. As prevention, self-care, and clinical care are slowly merging driven by tech, we have to start thinking about these difficult questions and there is no time like the present.
This article was originally published on medium.