digital numbers in a form of a red heart

Nigel Cannings, CEO and founder of Intelligent Voice, explores technology’s role in mental health diagnostics. 

How can technology play a part in mental health diagnostics? According to the mental health charity MIND, 1 in 4 people will experience a mental health problem of some kind each year in England, and 1 in 6 people report experiencing a common mental health problem (like anxiety and depression) in any given week in England.

However, approximately only 1 in 8 adults with a mental health problem are currently getting any kind of treatment, and the most common treatment offered is psychiatric medication, so demand for services vastly outstrips supply.

Experts believe digitalised mental healthcare could bring lasting benefits to many people who might otherwise go untreated. Whereas enormous strides have been taken in technological solutions for physical health over the last 50 years, the same cannot be said of mental health where standardised analysis techniques are nowhere near as developed as they might be in an area such as blood chemistry.

NLP ‘more accurate’

This is where the analysis of language can come in. Natural language processing is the part of AI that examines the interactions between computers and human language, and how you programme computers to process large data sets of natural language. ‘Deep learning’ in the past few years has accelerated this process, resulting in models that have a much more accurate understanding of human speech patterns.

Two years ago, the charity Relate hit the headlines when its chief executive Aidan Jones said he was looking at the potential of using artificial intelligence (AI) to help his 1,500 online counsellors cope with the demand for services. The charity received money from the National Lottery’s Digital Fund to create targeted, innovative ways to help the communities they support.

The technology in the form of chatbots for use in live chat counselling services would provide “relative anonymity” for clients, Aidan Jones said at the time, which some people preferred as it made them feel more comfortable opening up about their problems. In addition, the AI learned as it interacted with different clients.

Telehealth insight

Automatic speech recognition (ASR) solutions have been with us for many years. Take telehealth, for example, where ASR can be used to gain insight into what is said between a patient and doctor during virtual telehealth consultations, and using the information gained to aid patient-doctor matching initiatives.

After the consultation, ASR can be used to analyse the effectiveness of doctors with the technology used to note where there is a lack of understanding from the patient and/or the doctor side, if the doctor talks too quickly or there frustrated speech from either the doctor or patient.

Conversational analytics can extract behaviourally significant features from voice. Each feature will have a specific value by use case basis which can be established through tuning of the decision engine, and this can help with mental health diagnostics.

In telehealth, Intelligent Voice has created software that uses medical named entity recognition and linking to highlight key conversational topics such as drug names and symptoms and linking them to the International Classification of Diseases 10th Revision (ICD-10) and the unified medical language system.

Robust ASR that is paired with a medical language model that can recognise specific terms can be used to develop the capability for ASR dedicated to mental health diagnostics. Conversational speech is very different from the written word, but robust systems can be trained to recognise words, pitch and tone, and mark areas that raise a red flag.

Pronoun mapping, for example, tracks the speaker’s positioning of themselves within a narrative and correlates with certain mental health conditions, while emotion mapping can detect fear, anger, surprise, happiness, sadness and disgust.

In general, AI-enabled mental health applications can be used as decision-support tools for mental health practitioners; to customise the digital patient interface with human therapist intervention for one-to-one counselling; and to steer patient interaction remotely.

In addition, employer responsibility for mental health monitoring is likely to become more of a focus, accelerated perhaps by the pandemic where so many people’s working lives were turned upside down. The duty of care to employees must include taking care of their mental wellbeing, and many of those who experience mental health issues are either undiagnosed or go without treatment, partly out of concern that being identified as suffering from mental health problems stigmatises them or that accessing services means taking time off work.

It is well recognised that employees who have good mental health are more productive, so it is within every employer’s interest to ensure services are available – perhaps through the use of AI services that can monitor wellbeing.

One of the challenges will be ‘selling’ technology to patients, assuring them there is no infringement of privacy, that the system used to develop diagnostics is robust—algorithms got a bad rep during the lockdown when tens of thousands of students missed out on the exam grades they expected—and that human interaction is there when they want it.

But when you consider the advancement of technology in the physical sphere and the benefits it has brought to health, it is to be hoped that digital mental health diagnostics will experience that same advancement, and lead to more equitable treatment for all.

The original version of this article was first published on Med-Tech Innovation, by Nigel Cannings.

Click here to download our free Good Mental Health Hygiene booklet.