UK-based digital therapy company Iso on Tuesday announced a $53 million Series B round. The round is the funding it needs to take the company in a new direction: creating more intuitive autonomous text therapy.
In other words, the AI trained thousands of hours of real-life therapy can deliver personalized sessions over chat.
ieso has been around for ten years, and is running a text-only medical service (with a human doctor at one end) through the UK’s National Health Service. So far, the company has provided text-based therapy for about 80,000 patients — though 6,000 are actively receiving therapy, Iso CEO Nigel Pitchford told Nerdshala. This is a total of 460,000 hours of therapy so far.
“We’ll be providing some 400 hours of therapy this evening through our network,” Pitchford said.
This most recent round of funding was led by Morningside, with participation from the Sony Innovation Fund. It also included existing investors IP Group, Molten Ventures and Anand Impact Ventures.
Eventually, ISO wants to move itself from a human-based physician-based system to a largely autonomous system. The idea of AI-based chat therapy isn’t exactly unique in this space (we’ve covered other companies pursuing it) but data behind Ieso’s approach is what the company sees as its secret sauce.
The “undue advantage” of iso is ten years of real-life text-based interactions between patients and physicians, which Pitchford calls a “transcript of care.” that dataset is paired With real-time data on patient clinical outcomes, which the company is collecting along with those transcripts.
“ISO has created one of the most impressive data assets in the space with its text therapy dataset,” said Stephen Brusso, an investment partner at Morningside. The dataset is one aspect of the ieso that it finds most attractive as an investor, and has been called “phenomenal”.
This dataset has been used to track how certain therapeutic interactions or techniques are linked (or not linked) to patient improvement. And, there is some evidence Iso has managed to mine that data for insight. For example, the company published a paper in 2019 Jama Psychiatry Analysis of 90,000 hours of therapy. The paper found that aspects of therapy such as “planning for the future” or certain cognitive behavioral therapy techniques were associated with better patient outcomes.
Overall, the data suggested that about 28 minutes of every hour of therapy, there is a conversation or practice that “directly affects” patient outcome, said Andy Blackford, the company’s group chief science and strategy officer.
Perhaps on the upside, the paper also found that therapeutic empathy was negatively associated with patient outcomes – although other research has suggested clients get better results When they feel their therapists understand them. Blackford interpreted the discovery of empathy as evidence that empathy should be employed alongside other therapeutic techniques.
Ultimately, Pitchford looks at this dataset, and does the sort of analysis done in it. jama Paper as a roadmap for how AI based practitioners will be trained and personalized.
“So essentially, we’re studying what the very best therapists are doing at a very high level, and then reconstructing it so that it reaches people who are unable to access human psychotherapy delivery, who There is a big problem all over the world,” he said.
Even with this dataset, it seems to be operating in an increasingly crowded space. According to the Silicon Valley Bank Trends report, funding for mental health startups is projected to crack $3 billion in 2021, per Q3. This means that a lot of minds are focused on the problems associated with traditional medicine right now.
Bruso sees Iso as one of the few mental health companies that can at least pinpoint real-world health outcomes using their own datasets.
“We believe there is a unique synergy between ISO’s digital products built on real world data and their ability to test these products across their existing user base to generate results data from day one,” he added. “At the end of the day, the products that will stay in space will be which can demonstrate a measurable impact on both individual health and social outcomes.”
Blackford is also aware of how full this space is – in fact, he sees it as a problem for consumers. These apps, i.e. leadership figures, are often designed for patients with self-help, mindfulness, or non-serious mental health diagnoses.
Although ISO can treat people with non-advanced mental health struggles, the company’s focus is also on moderate to severe diagnoses. It is not, to use their words, a “wellness solution” – and can be used by groups that may need more intensive care.
This focus means that safeguards against self-harm need to be particularly strengthened. Blackford says the company has risk escalation protocols in place for its human-based therapy models that have been hailed over the course of ten years as one of the largest mental health providers in the UK. The company plans to incorporate those programs into autonomous medical products in the future.
Right now, Blackford doesn’t anticipate a difficult regulatory path as it looks to tackle high-stakes mental health diagnoses.
“The good thing is that there are examples and predictions that we can use when it comes to market. But the important thing will be, you know, clearly safe and effective products,” he said.
Blackford argued that the high volume of data the company has already collected will allow insights into efficacy and safety to be gathered “many times faster” than you would expect under routine clinical trial conditions.
Going forward, ISO plans to use this round to build its AI-based medical arm and expand its presence in the US. The team is expected to grow to around 200 people by next year, which will give a push to go to market in the next two years.