One of the most interesting things I saw at RSNA 2016 in Chicago this week was Philips’ Illumeo. Beside being a really slick radiology interface that they’ve been doing forever, they created a kind of “war room” like dashboard for the patient that included a bunch of data that is brought in from the EHR using FHIR.
When I talked with Yair Briman, General Manager for Healthcare Informatics Solutions and Services at Philips, he talked about the various algorithms and machine learning that goes into the interface that a radiologist sees in Illumeo. As has become an issue in much of healthcare IT, the amount of health data that’s available for a patient is overwhelming. In Illumeo, Philips is working to only present the information that’s needed for the patient at the time that it’s needed.
For example, if I’m working on a head injury, do I want to see the old X-ray from a knee issue you had 20 years ago? Probably not, so that information can be hidden. I may be interested in the problem list from the EHR, but do I really need to know about a cold that happened 10 years ago? Probably not. Notice the probably. The radiologist can still drill down into that other medical history if they want, but this type of smart interface that understands context and hides irrelevant info is something we’re seeing across all of healthcare IT. It’s great to see Philips working on it for radiologists.
While creating a relevant, adaptive interface for radiologists is great, I was fascinated by Philips work pulling in EHR data for the radiologist to see in their native interface. Far too often we only talk about the exchange happening in the other direction. It’s great to see third party applications utilizing data from the EHR.
In my discussion with Yair Briman, he pointed out some interesting data. He commented that Philips manages 135 billion images. For those keeping track at home, that amounts to more than 25 petabytes of data. I don’t think most reading this understand how large a petabyte of data really is. Check out this article to get an idea. Long story short: that’s a lot of data.
How much data is in every EHR? Maybe one petabyte? This is just a guess, but it’s significantly smaller than imaging since most EHR data is text. Ok, so the EHR data is probably 100 terabytes of text and 900 terabytes of scanned faxes. (Sorry, I couldn’t help but take a swipe at faxes) Regardless, this pales in comparison to the size of radiology data. With this difference in mind, should we stop thinking about trying to pull the radiology data into the EHR and start spending more time on how to pull the EHR data into a PACS viewer?
What was also great about the Philips product I saw was that it had a really slick browser based HTML 5 viewer for radiology images. Certainly this is a great way to send radiology images to a referring physician, but it also pointed to the opportunity to link all of these radiology images from the EHR. The reality is that most doctors don’t need all the radiology images in the EHR. However, if they had an easy link to access the radiology images in a browser when they did need it, that would be a powerful thing. In fact, I think many of the advanced EHR implementations have or are working on this type of integration.
Of course, we shouldn’t just stop with physicians. How about linking all your radiology images from the patient portal as well? It’s nice when they hand you a DVD of your radiology images. It would be much nicer to be able to easily access them anytime and from anywhere through the patient portal. The great part is, the technology to make this happen is there. Now we just need to implement it and open the kimono to patients.
All in all, I love that Philips is bringing the EHR data to the radiologists. That context can really improve healthcare. I also love that they’re working to make the interface smarter by removing data that’s irrelevant to the specific context being worked on. I also can’t wait until they make all of this imaging data available to patients.