Neil Sebire, Professor and Chief Research Information Officer (CRIO) at the hospital, recently told delegates at a London-based Microsoft event that the organisation is partnering with the technology vendor to create a research platform on Azure.
Sebire explained that whilst Great Ormond Street is very research active, where last year it did more than 1,000 clinical research projects, it is beginning to think about how it can better use technology to provide superior care for its patients. His job is now essentially to understand how the hospital can maximise the use of technology and data in the healthcare environment. He said:
This really started a couple of years ago when we started to re-look at the whole digital strategy for the hospital. And we did something a little bit different. We basically said, let’s ignore healthcare. Because healthcare always thinks it’s very special and different. We wanted to look at technology. What are the technologies that are going to impact society? And what would we really like to be able to do?
One of the striking things that we came up with was the realisation was that medicine in general, but particularly for hospitals like this that are very specialist, we are going to become a data analytics business. This is not something that’s going to be done on the side.
Sebire said that whilst healthcare is very early on in this journey compared to other industries, progress is starting to take place and things are starting to change. However, he added that up until now, healthcare institutions have been using technology to make doctors more efficient - slightly better at what they do now.
Instead, healthcare should be thinking about how technology can be used to enable doctors to do things that they cannot do now. Sebire explained:
The real bonus is when we start to become augmented. We don’t want to be a little bit better. What we want as doctors is to be able to do things in the future that doctors cannot do now.
I’m a diagnostic pathologist. I spend my time looking down a microscope, looking at biopsies, to see if it’s cancer, for example. We know that AI systems now are better than humans at pattern recognition, they’re better at recall and looking up lots of information and making algorithmic diagnosis. It’s just we haven’t taught them yet. These systems will be better than any human. And that’s really where we want to be.
How do you get there?
Sebire said that a huge amount of work needs to take place to enable this kind of change. Great Ormond Street Hospital assessed its IT estate and realised that it had 400 different clinical systems in place, meaning that there was no common data model, no common place to put that data. These systems are now being replaced by a single patient record system, which captures all the data and stores it in one place. Sebire explained:
However, that’s not enough. Those systems are not designed for research, they’re designed for managing patients. So in addition to that, we’ve actually implemented a cloud-based dedicated research platform, which sits on Azure. That means we can take data from our systems and under our own governance we can de-identify that data, we can audit the data, we can control it, and apply tools on top.”
Undertaking a project like this, one with high ambitions, doesn’t come without its challenges though. Sebire added Great Ormond Street doesn’t want to be creating its own technology, it wants to work with technology partners that adopt a platform approach. He said:
We don’t want to be inventing technology. We want to be leveraging technology. Why have we chosen Microsoft? We went through a long process of looking at what we required and with healthcare data the most important thing is security. We do not want to give our data to somebody else to do the analysis. We want to keep the data. We are the academics. We know what we want. We want tools to apply our data that stays within our control. We want the ability to be able to proof of principle testing really quickly. The things that are good, we want to be able to scale them out into clinical practice quickly.
Similarly we don’t want the people working to have to learn new tools for every single different cognitive agent we want, we want a platform approach. Therefore, Azure fits very well with that. Finally, we are well aware that we are very early on this journey here. But when we move on to IoT and edge computing, which is clearly going to be for a lot of the stuff that we do with high volume data devices, we don’t want to have another platform for that. It’s this ability to scale up further.
Maintaining control and authority of the data and technology, however, creates a skills problem for the hospital. One that Great Ormond Street is looking to address. Sebire explained:
One of the biggest other problems, is that everything I’ve talked about so far is nobody's job in healthcare. And that’s the single biggest issue. How do you bring together commissions, doctors, nurses, computer scientists, technology?
This is going to be a whole new thing in medicine. One of the things we are doing to address that is to develop a digital research unit, specifically to bring these people together.