The increasing pace of development in digital innovations aims to put patients in charge of their health and care and offer new ways to address the challenges of an ageing population, but how are entrepreneurs changing the world of healthcare with the help of technology? We took a look at two start-ups in the AI space that want offer clinicians the tools they need to improve outcomes and give people more control over their care  – and we wanted to find out how they’re doing it.  

"The next tier of AI coming through to hospitals"

Let’s start with Ultromics. 

An Oxford University spin-out, Ultromics claims to have developed “the world’s most accurate echocardiography software” to improve diagnosis of coronary artery disease, one of the leading causes of death in the UK, by more than 90% - compared to the current 80% accuracy of echo tests. 

“If you’ve ever seen an ultrasound image of the heart, it’s very difficult to interpret, so even getting it 80% right is good,” Ross Upton, CEO and Co-Founder of Ultromics, told MobiHealthNews. 

“But then we also found out that it’s only 80% in the very best hospitals, with the most experienced clinicians, and [in] some of the hospitals where clinicians are less experienced it’s only around 65% to 70% accurate. 

“That is quite low, especially when diagnosing the most prevalent disease in the world.”
 

"I think we can be the exemplar company on how we should innovate and then give back to the NHS"

Ross Upton, Ultromics

Starting his PhD at Oxford University, Upton set out to develop a methodology to address this issue, working with Paul Leeson, Professor of Cardiovascular Medicine, now the start-up’s Chief Medical Officer (and Co-Founder), to put together a database of patient images from clinical trials ran at Oxford University and collect longitudinal patient data, following up on the individuals involved for up to one year after having the tests.  

“It was important to do that because we know the test itself is only 80% accurate, so we had to follow the patients up to see what had happened after, to see if the test was right or not," Upton said. 

Leveraging access to this database, they developed the technology now known as Ultromics, extracting “thousands of different features” from patient images, using machine learning to relate those to predict what had happened to patients in a year’s time. The software can now analyse more than 80,000 data points for every echocardiogram scan to spot signs of disease.

Oxford University, aerial view

Ultromics trial to be expanded to 20 NHS hospitals, followed by 10 other sites

Working with Oxford University, Leeson and Upton set up a trial across six cardiology units, which is now set to be extended to 20 NHS hospitals in the UK, to be followed by 10 other sites. A study outlining the results of the clinical trial will be published later this year, but Upton said initial findings suggested the technology had been generating “better results than the 80% accuracy currently performed”.

Once Ultromics receives regulatory clearance – Upton said regulatory bodies would have to develop a new classification around the Ultromics model, as their product was different to the two current approval streams for the application of AI in healthcare -  these NHS sites will be able to use the software under the initial clinical trial agreement "at no cost", while Ultromics will continue to collect trial data to improve their technology.

Upton described it as a “win-win” for both sides: “I think we can be the exemplar company on how we should innovate and then give back to the NHS.” 

Meanwhile, they are looking to bring their technology to market in the US, starting with trialling it at hospitals involved in their scientific advisory board. Once they receive FDA clearance, Upton said Oregon Health and Science would become the first organisation in the US to adopt their technology.  

“Ultromics is actually taking it to the next level where we’re not trying to notify doctors and we’re not trying to localise disease, we’re actually trying to take it to the next step and improve patient outcomes. So, we’ll be the next tier of AI coming through to hospitals.”

The start-up was only launched in 2017 and in May this year it announced that it had raised £10m in a Series A round led by Oxford Sciences Innovation. Asked what his advice would be for other innovators that are looking to translate their idea into a tangible solution, Upton said a university affiliation had a big impact and advised other companies to adopt “the same approach” if they want to work with the NHS or train complex algorithms using its data. 

“…charging the NHS an extortionate amount for a piece of technology when you use the NHS data to generate that technology doesn’t feel quite right. I think it would help companies if they adopted a similar model to the one that we’re adopting.” 

The start-up aiming to reach 100 million people by 2020 with its personal health guide

Launched on the market in late 2016, after six years of research and development, Ada Health’s AI-powered personal health guide was one of world's fastest growing health apps (CE Marked) in 2017. It aims to help users identity the appropriate next steps for their care, asking a series of questions based on medical reasoning and AI to assess symptoms, and its makers say it learns from every interaction, making the experience “increasingly more personalised” - along with the chatbot functionality, it stores data  from previous conversations and general information about users' health in the cloud.  

Founded in 2011, the Berlin-based start-up, – now with offices in London and Munich as well – is on a mission to reach 100 million people by the end of 2020, with its app now being used by more than four million individuals around the world and figures showing that every three seconds a health assessment is completed through the platform.

It has now been launched in English, German, Spanish and Portuguese, according to the markets with the highest uptake, and a French version is due to go live shortly.

Initially, Dr Claire Novorol, the start-up’s Chief Medical Officer, set out to create, along with the other two co-founders, Daniel Nathrath, now CEO, and Dr Martin Hirsch, now Chief Scientific Officer, a clinical decision support tool for clinicians. From trials undertaken in Germany and the UK, however, they soon found that doctors wanted them to create a tool to help collect data from patients ahead of their consultations.

“So we started building this pre-assessment tool and then we realised, why not just make it as user-friendly as possible, put it in an app, put it out there and just let people around the world use it,” Dr Novorol told MobiHealthNews.

"Our core activity started with UK and Germany, but we’re now focusing more and more on the US, and we have a particular interest in the developing world with the project in Africa."

Dr Claire Novorol, Ada Health

Ada Health signs partnership to improve access to services in the developing world

While the company still works with a number of specialist clinics through research partnerships around this tool, the CMO explained it was “no longer the core focus of the company”. 

Instead, they will be launching soon a web version of their personal health guide app, which is currently available on iOS and Android, as Dr Novorol argued the future of healthcare would be "much more patient centric". The start-up has forged a number of partnerships with health providers and payers to offer its personal health guide, and is expected to shortly make an announcement to improve access to services in Africa. 

“It’s very much a global perspective, I guess our core activity started with UK and Germany, but we’re now focusing more and more on the US, and we have a particular interest in the developing world with the project in Africa.”

In October last year, Ada Health announced that it had secured £35m in a funding round led by global investment group Access Industries to strengthen and expand its operations and continue to develop its offerings, and Dr Novorol said:

“I really see a future where there’s a real ecosystem of solutions where the individual is much better equipped to manage their health and take control than they have been in the past."

Twitter: @1Leontina
Contact the author: lpostelnicu@himss.org