First posted on our sister title, Digital Health Age.
Data
Within the healthtech sector, start-ups offer perhaps the most exciting range of products and solutions designed to improve the health and wellbeing of populations.
Launched in 2017 as a competitive grant, the Pfizer Healthcare Hub: London is designed to help start-ups with innovative healthcare technologies scale-up their product and advance their business. The competition gives three start-ups the chance to win a share of £50,000 alongside support from Pfizer to help the companies grow and reach more patients and healthcare providers faster.
Last year, digital health companies Cera, Give Vision and Echo won the competition and have since went onto expand their businesses through initiatives such as funding rounds, marketing and growth support and specialist consulting.
Now, the 2018 edition of the Hub has just concluded its pitch event during which 10 shortlisted candidates presented to the Pfizer team about the benefits of their products or services to healthcare in the UK.
Digital Health Age had the chance to sit down with Dr Hamish Graham, Pfizer Healthcare Hub: London Manager, as well as two of the finalists, to talk about the event as well as innovation within the NHS.
In the last of a series of articles surrounding the Pfizer Healthcare Hub: London, founder and CEO of Illumr, Jason Lee, talks to Digital Health Age reporter Reece Armstrong about the use of AI and data within healthcare.
Q: How did the event go yesterday?
Really well. I was very impressed with how organised it was and by the great panel of judges. I actually find it really enjoyable, it can quite nerve wracking standing in front of these different groups of people, but we got some really good feedback and are hopeful that we might get through.
Q: Could you tell me a little bit about Illumr?
What we do is we help organisations better understand and predict patterns of behaviour. To do that we’re using some remarkable technologies and have got over a decade’s worth of academic research. What currently makes our data analytics software great is that we can take data and self-organise it in a 3D environment to reveal previously hidden patterns. What makes us unique globally in AI is that we have a new methodology and a proven ability to reveal insights that all tools might miss. So, I’ve been going to global organisations like Pfizer to say that we can find things that your data scientists are blind to right now and we’ve proven time and time again that we can do that.
Q: Do you think AI has been overhyped within healthcare and where do you see it being the useful?
Actually I don’t think it has been overhyped because we are just getting to the point now where the media are sending a message out to the world that it’s very important. It’s not being applied to the right verticals and healthcare and life sciences are two of those areas where it has only just begun to make an impact. So, I don’t think it’s been overhyped but I think there’s an expectation that it can solve everything and that’s where the hype probably exists because AI cannot solve everything, it’s not going to solve all diseases, it’s not going to solve all the problems and make everyone healthy again. What it can do is obviously uncover things that may never have been seen before and help identify where patients can get better outcomes. I think the value it can provide is certainly a game changer but it’s not going to solve all problems, I think we’re only just beginning to see the benefits of AI in healthcare.
Q: The NHS is sitting on a large amount of data. Do you think the NHS has utilised it in the most beneficial way and what can Illumr for that data in the NHS?
I don’t think it has utilised it in the best way and that’s because all large organisations really have information silos, they have a lot of different systems where the information sits in and often their data sets haven’t been joined or one part of the organisation doesn’t know another part has certain data. We’re getting to a point now where people are realising that these datasets are important, they’re very valuable and within those datasets are intrinsic insights that may have never been seen before. And once again I think we’re just on the starting point of knowledge discovery within these datasets. At Illumr, we have a new methodology and a proven ability to find insights that all other methodologies and tools may miss. That’s important for several reasons. The first thing is a lot of the media are talking about deep learning right now and that’s fine if you know what you’re looking for because you can train that computer to find the patterns, sort of like supervised learning. The other side of AI that not many people talk about is unsupervised learning and that’s when you can’t give the computer an example. So let’s take drug discovery for instance, the only way you’re going to find signals in data is to get in a trained data scientist to read them and come up with a hypothesis as to where that emerging signal lies. That can take weeks to prove if that signal even exists and because a human is doing it means you’ve got human bias involved. You may end up with a solution that doesn’t match the reality, or you may lose the signal completely because you don’t know what to look for. What makes us unique is that we can go through these datasets to reveal patterns in a hypothesis free environment. So sometimes the patterns we reveal are things that have never been seen or even hypothesised before. And if you’re looking at new drug discovery this is really important because we might identify a small outlying factor that’s indicative of a new multibillion drug discovery. Or, if you look at clinical trials data, we may be able to identify a sub-population in a trial where those patients are going to get a better cancer outcome. Ultimately, we can make a company a lot of money through drug discovery or we can save lives by identifying a better cancer outcome for certain subsectors of patients. That’s really what’s exciting about working with Pfizer or the NHS, we’re hopefully getting involved in a lot of projects and are helping make really good outcomes.
Q: And where do you see the biggest advances in diseases or new discoveries coming from in regards to data analytics?
What I said yesterday is that we are a very horizontal solution, so we go across a number of different sectors. Within healthcare and life sciences what we like to solve are pain points. We are still about six months away from having a product that anyone can use themselves and that we can sell. But we’re getting use cases under our belts to demonstrate we can solve big problems with big customers. In healthcare it is stuff that I mentioned earlier, finding and identifying the subpopulation in clinical trials for better cancer outcomes. That’s very much a patient side for an outcome. For the providers, it’s identifying maybe a new method of fraud in insurance claims that they haven’t seen before. Or maybe we can identify medical waste in order to stop waste happening. If we can do that for providers and patients and identify that type of outcome, that can really make a dramatic impact. We’re very much data agnostic, we can take in dirty data sets with missing values and get these actionable insights very quickly.
Q: Lastly, as a start-up what’s the most difficult thing about gaining traction within the UK healthcare market and how important are initiatives like the Pfizer Healthcare Hub: London?
Well getting traction within any vertical is difficult but because we’re B2B all of the organisations we target are very large. For a start-up it is a very long journey, it can take over 18 months to get your first sale. What you spend most of your time doing is finding the right person in that organisation to talk to. That can be really, really difficult. And for us, we’re just starting out on the journey in healthcare and life sciences and the Pfizer Healthcare Hub: London is critical. It’s basically one of the only ways we can get into this market. What the accelerator does for this is enable us to talk to the key decision makers, those people that have pain points and are working on clinical trials; who want to find that better cancer outcome. For us we need that guidance, we have something that is very powerful and very disruptive but it needs to be guided by those who are next to the problem.
