Meet the founders using AI to improve mental healthcare

Posted: 12 October, 2021
Men smiling

1 in 4 people in England will experience a mental health problem every year (Mind). With the impact of COVID-19, lockdowns, and increased awareness of mental health disorders, the number of people seeking treatment is growing exponentially. But globally, meeting this demand clinically has proved challenging.

In response, startups providing mental wellbeing tools are gaining traction – and investors are noticing. Meditation apps Calm and Headspace have both raised significant capital in the last two years and deals are increasing in volume.

Limbic, founded through Entrepreneur First by Dr Ross Harper and Sebastiaan (Bas) de Vries, is different. They’re going beyond the self-care and mindfulness solutions to address one of the sector’s biggest pain points, head-on. 

Patients need evidence-based approaches that are still personal and human for treatment of mental health disorders. But they’ll often face long waiting times, and a lot of hoops to jump through, to get it.

Limbic’s AI-powered therapy assistant supports patients to access the right treatment pathways and provides Cognitive Behavioural Therapy while on the waitlist – allowing clinicians to access more information, and improving patient experience and drop-out rates.

Their software is currently being used by 15% of the NHS’ Improving Access to Psychological Therapies (IAPT) services – the NHS’ talk therapy programme for common mental health disorders in adults -and is seeing a monthly growth in patient usage of 113%. 

Applying AI to as complex an area as mental health might seem counterintuitive. In fact, we might expect our inner lives to be one of the last things that a computer could understand. 

But rather than eliminate the human touchpoint altogether, Limbic serves to optimise patient experience and improve predictions and outcomes. And the results are showing it.

We spoke to the founders about how they harnessed their deeply technical backgrounds to build the company; how their technology works; and how they’re making a difference for clinicians and patients alike.

The Founders

Ross began his career in academia, completing a PhD in Computational Neuroscience.

“I was always doing side projects”, he recalls. “I co-founded a viral marketing project after I finished my undergraduate degree, which paid off my student debt. I definitely had the entrepreneurial bug.”

He made the decision to turn down a postdoctoral offer and pursue his ambition with our platform.

“It seemed the right opportunity to explore this entrepreneurial side of myself.”

CTO Bas also has a deeply technical background – however, he’d developed his skillset outside of academia. Programming since age 16, he’d dropped out of university to build a peer-to-peer payment app in The Netherlands. The app got successful traction but after an acquisition deal went awry, he wanted to start building again – so applied to EF.

The two arrived as individuals to our 9th cohort in London in 2017. “The second I arrived, I was in a room full of like-minded people”, Ross remembers.

“We strongly align in the way we think about things.”

Bas was the first of the pair to reach out, having read Ross’s profile. “He wanted to create something big,” Bas says. 

“That really appealed to me. Then, when I considered the potential of combining his academic experience with my more applied programming experience, I realised that we were going to be a really strong match.”

“What stuck out for me about Bas was our communication”, says Ross.

“We’re both really fast. We get what we both mean and we get it quickly. And while we’re from different backgrounds, we strongly align in the way that we think about things.”

With their skillsets aligned, and their bond formed, it was time to get to the business of founding a company.

The Problem

Bas and Ross’s focus was on building something with huge room for impact.

“We’re both incredibly driven by finding big problems to solve, and quickly landed on mental health as a space to explore,” Ross explains.

“Mental health is the largest cause of disability worldwide. As an area of medicine it’s growing ridiculously fast.

Now, we’re faced by a massive problem where patient need outweighs clinical supply. So there’s a real societal issue for us to tackle.

At the same time, it presents an intellectual challenge for us as founders. It’s a very clinical, academic area of medicine. It’s cognitive, hard to pin down and hard to quantify. We wanted to build something that bucked that trend, and was very software focussed – deploying our probabilistic approaches to data to make a real difference.

We knew we could do it. Bas can package products for consumers like no-one I’ve ever met. This meant that he could take my background in neuroscience, machine learning and modelling and turn it into a real business rather than a research project.”

With a clear problem to solve – a disparity between the need, and the resource, for mental healthcare – they set to work finding the pain point where they could make a difference.

“We spoke to a lot of patients and therapists. We realised that the space for us was not going to be in building a digital treatment – we weren’t going to build something you prescribed”, says Bas.

“We learnt about the large inefficiencies in the assessment process for the NHS. We took a real critical lens to this, observing each care pathway component as if it were a part of a machine, and looked at where AI and clever product design could augment and accelerate the process.”

Limbic Self-Care

The Solution

“We realised that rather than provide a treatment, we could augment the therapist, to improve patient pathways as they go through the process,” Bas explains further.

“We began with a patient app, and a therapist app, that was adopted by 100 private practices in the UK. From this, we were able to learn a lot about the sort of problems that patients were experiencing, and how we could optimise for this.”

This analysis resulted in three product solutions:

  1. Limbic Access, which helps patients refer into services. By speaking with a web-based, conversational, AI-powered chatbot, the platform groups patients and identifies the right treatment pathways – saving clinical time, and helping the therapist gather more information.
  2. Limbic Self-Care, a mobile app for patients on the waiting list for treatment. The app supports the patients during this time with daily check ins, and by sharing CBT tips and techniques. While doing so, it collects information that can support monitoring and triage.
  3. Limbic Care, which supports patients and clinicians during the treatment process. This records negative thought and activity diaries to identify key patterns, and creates personalised recommendations for patients in between sessions

The presentation of Limbic as an AI chatbot is perhaps more commonly seen in other, less personal settings such as FinTech and digital assistants. Many might be reluctant to believe that an area so intrinsically ‘human’ as mental health could be served by software.

“We're not replacing the therapist - we're augmenting the therapist.”

“Crucially, we’re not replacing the therapist. We’re augmenting the therapist – addressing inefficiencies in the patient care pathway, ensuring that they can make the most of their time and providing a better experience.

We wanted to mimic that experience of speaking with your therapist – there is an opportunity for a friendly, interactive interface here. The software was built by Bas with that in mind to engage patients and clinicians.”

There are important advantages of using AI for these services too. Crucially, with each user interaction the software gets smarter.

“You can think of it as this: our first vision was that we would have a trainee psychologist in the pocket of every patient. This trainee would help each user group differently, and work with clinicians to create value”, Ross illustrates.

“With every year that goes by, and with the data we’re collecting through our current positioning in the care pathway, our trainee psychologist becomes a junior psychologist. And then a senior psychologist. With each of those steps, we’re increasing the value we can create.”

From MVP to serving 15% of NHS talk therapy services

With a product in place, the team now set about getting it into the hands of patients and clinicians. But with an industry as conservative to change as medicine, it wasn’t going to be easy.

“Quite rightly, there needs to be a great deal of evidence and general consensus for the medical community to shift their thinking on any topic – and especially one as important as mental healthcare,” Ross notes.

“But the way we solved this, and were able to enter this market, was supported by the fact that we’re not a digital treatment. We don’t say, ‘this is going to solve depression or anxiety, give this untested treatment a go’. You couldn’t convince a community of that – you’d have to go through the normal process of gaining clinical certification.

Instead, our approach was “we know what your workflow looks like, we understand where your major problems are – now here’s a solution.’”

Based on evidence from early adopters, it appears the solution is working.

“You’ve got three stakeholders when you’re talking about Limbic. 

You’ve got the patient, who’s an end user. You’ve got the clinician – also an end user. And you’ve got the services and the healthcare players, who are the buyers of our product [Limbic operates on a B2B model].”

In therapy

Feedback from patients suggest that the software is having a positive impact on their treatment journeys. 91.64% of patients said that the tool helped them access care, and the service has saved over 3900 weeks of wait time.

“Patients love Limbic because it makes it way easier to do every part of their care journey. They can do their negative thought diaries, just through a conversational interface, and they can refer into the service really easily with database lookups. Their wait times are reduced, and it’s a more interactive and engaging journey”, says Ross.

“But the clinicians really like it. It’s supporting their normal workflow, and allowing them to see more patients per unit time. 

This is a critical moment for them too – currently you’re seeing massive amounts of clinician burnout because they can’t meet patient demand. Limbic Access has already saved 359 hours of clinical time, because it does the bulk of the work around referrals.”

And as for the third stakeholder, the service providers? 

“By saving this time, Limbic reduces costs, and frees up clinical time to do more. So they like it too.”

Harnessing digital technology to fight the pandemic

If the need to enhance mental health services wasn’t clear before COVID-19 it certainly is now. The widespread effects of grief, isolation, and increased uncertainty in a world disrupted beyond belief has taken its toll. We asked if Limbic had seen any changes in the ways their services were accessed as a result of the pandemic.

“Initially there was chaos”, says Ross. “It was ‘all hands on deck’ in the NHS, and there was far less openness to taking on projects not directly related to COVID-19 – and for good reason, because we were in a state of emergency.

But it became clear that the mental health effects of the pandemic were profound, and the mismatch between patient need and clinical resources has become even worse. This has made the problem solved by Limbic more and more urgent.”

Another unexpected effect of the pandemic has been health services’ increased openness to new digital solutions.

“The NHS has undergone a forced experiment in digitisation. This is an opportunity for healthtech startups like Limbic. The barriers are briefly down so that we can start creating value and demonstrating benefits in healthcare settings that are suddenly much more open to digital solutions and their impact.”

Looking ahead: Tackling one of the biggest problems of our generation

With Limbic having now processed roughly 16,000  referrals, the company is looking to expand their offering, and their impact. 

“Our goal is to assist on every step of the care journey. We want to get our software smarter, and to improve our solution. For example, while Limbic Access is currently reducing around 50% of patient wait time, we believe we can get that down to about 80-90%”, says Bas.

Speaking to the company’s growth, Ross explains there’s also room to expand the AI components of their software. There are three areas that currently, Limbic doesn’t support:

  1. Diagnosis to a clinical level of accuracy.
  2. Risk monitoring and prediction, particularly of self harm and suicidality
  3. Treatment planning – devising the best treatment plan for the patient based on information they collect

“You currently need a therapist for these,” says Ross.

“Our goal over the next two years – and we believe we can do it in that time – is to use our position and the data we’re collecting to build a diagnostic model. And spoiler alert, we’ve begun to do it and it’s working incredibly well. 

For the next two, risk monitoring and treatment planning, we’ll build an algorithm that can assess people’s digital profiles and establish what’s likely to work for them. For example, based on the information you’ve input, and analysis of those with similar experience, we may say that acceptance based therapy is going to work better than a behavioural approach. So we’ll be building this into the product over the next two years. 

When we do that, we believe we’ll be in a really strong position to reduce costs for healthcare players.”

The team also plan to expand their services into more of, and beyond, the NHS. In the UK, they’re now exploring relationships with private health providers, and are now looking at new geographies, having onboarded clinical leads in the US to join their small but fast-growing team.

As the opportunities for their business grow, we asked if the team still carried the same motivations they joined Entrepreneur First with just a few years ago.

“Mental health is one of the biggest problems of our generation”, reflects Bas. “And it’s really exciting to make an impact in that space.

In addition to this, it’s intellectually challenging. We have taken a new look and approach to solving problems in this space that no one has really done before. It’s really rewarding.”

“Hearing how we’ve helped takes it out of being this intellectual problem we wanted to solve, and makes it incredibly personal.”

While those motivations remain core to the founders and their mission, Ross notes a further driving force now that the product is in the hands of their users.

“Limbic is being used by 12,500 patients at the moment, and I’d defy anyone not to be touched by hearing how we’ve helped. It takes it out of being this intellectual problem we wanted to solve and makes it incredibly personal. 

You see people talking about the depth of their psychological issues and the impact it’s having on their lives and ability to be happy. And then in that same breath, they say that Limbic made life better for them. That created a fire under the team – when the product got launched, it created this extra energy to keep going in the direction we were going.”

Looking ahead, we asked Ross and Bas what their vision for the future, with Limbic’s support, looks like.

“In five years time, if Limbic carries on growing at the rate we’re growing – and we fully intend to – then every individual with symptoms or diagnosable mental health issues, will have an effective high quality digital therapist available to them. 

And they’ll have it 24/7, wherever they are, regardless of the socioeconomic background they come from. That’s where we want the world to be.”