Allscripts CEO Paul Black talks AI, machine learning and precision medicine

Chief executive shares his vision for the future of genomics and offers a glimpse of the AI-based EHR the company is exploring.
By Mike Miliard
09:56 AM
Allscripts CEO Paul Black

In part one of our two-part Q&A with Allscripts CEO Paul Black on Thursday, he explained how the acquisition this year of McKesson’s health IT business has vastly increased the company's scale nationwide, and how he plans to capitalize on that to position the company for a new era in healthcare – focusing on innovation on an array of fronts and having some "fun" helping its clients solve an array of challenges.

In part two, Black looks toward 2018 and beyond, discussing how Allscripts aims to continue making inroads with precision medicine, artificial intelligence and more.

[Also: Allscripts CEO Paul Black claims the EHR vendor will pull ahead of the pack in 2018]

Q. Let's talk about precision medicine. I spoke at length recently with Joel Diamond, MD, CEO of Allscripts subsidiary 2bPrecise, about the near-term future for genomics and personalized treatment. Where do you see things heading here? How far off are we from it becoming the standard of care? What will providers and vendors alike need to start doing differently?

A. I think it is here today, but not necessarily at volume. But the fact that there are some 50,000 orderable procedures that you can order through LabCorp or other lab houses, the fact that those genetic tests are out there – not necessarily for a full genetic sequence, but also could include a two-panel test for a very specific condition you might have. To the extent that those are already out there, I would say it's already mainstream today.

But to the question of when it will be best community practice, I think that will be specialty type by specialty type. Clearly ASCO is pushing that for oncology and there's a lot of efficacy for getting genotyped for breast cancer and for other types of cancer. So the importance of getting the chemotherapy right based on the way your body is going to metabolize it is extraordinarily important, because you want to make sure you're on the right regimen. Different specialties will bring this to the forefront at different rates. But there is so much being published about it, so many studies being run, so many countries doing serious work based on certain segments of the population to understand why they are sick and how to best treat that in genomic sequencing in some form or another – that's all playing a very major role today.

I think 2bPrecise is very well positioned in that it's EMR agnostic, in that it surveils the database to actually look for people who should be sequenced, based on family medical history, based on current diagnosis, based on current medication regimen. It will actually say, This person might benefit from, and other people like them have benefited from, a genomic test. That then leads to the question of will insurance companies, will CMS become players in this. And I think the answer to that is yes and in many cases they already have. On the macroeconomics, the health economists will say these tests will pay for themselves, because of the efficacy, because you minimize the delay, minimize the effect of having people on the wrong medication, so you have the same or smaller spend.

[Also: Rx precision medicine tool latest to be integrated into clinical workflow]

So I think the combination of consumers demanding it, of insurance companies understanding the economics, the reality of the price performance – that the tests are no longer $10,000 to have a genome sequenced, the Moore's Law effect, if you will. All of that comes together to say the time is really great for this to be not only best community practice but also standard operating practice clinically in the next three to five years.

Q. What are some of the technology innovations you're particularly excited about? How do machine learning and AI factor into your plans, for instance?

A. We have talked about a new EMR that would be AI-based, machine learning based. That is something we have been working on. And that solution is predicated on the machine doing a lot more work for the physician with regard to how you practice, practice patterns, condition of payment, the medical history you have on a person and what sort of questions should be asked the patient is in the room – what kind of things have you asked of this patient that you either have or have not yet documented.

And so that component of a true assistant – something that truly adds value to the experience for the physician as well as the patient, is work that has been underway for the past three-plus years inside the company. We also have a substantial amount of data that we collect and have in a database of some 40 million different lives. That we have, in a deidentified way, a set of capabilities to look for patterns in that data that we might historically not have seen without the advent of an AI data crawler or capabilities where the data are being correlated in different ways that a PhD scientist may not have had a thesis around, historically.

So you're not as reliant on the smartest guy in the room to look at data as you are on the machine to bring you revelations based on correlations that exist in the data that you were not aware of. And, importantly, make that actionable because you have an EMR capability that will allow that to occur. So it's another extraordinarily exciting set of capabilities that exist today that weren't here a decade ago. And it's absolutely another top performer, from my perspective, in the 'fun' category.

Healthcare IT year in review

This was one of our most popular stories of the year.

Twitter: @MikeMiliardHITN
Email the writer: