Listen and Subscribe to Our Podcast
"The Dish on Health IT"
Engaging discussion around Health IT with perspectives from across the healthcare landscape. This informative and entertaining rotating panel of senior health IT consultants and their guests will keep you in the know about the latest innovations, policies and industry shifts impacting healthcare and point out the opportunities that lie within.
Subscribe on:
Apple Podcasts
Google Podcasts
EPISODE
Episode 9: Health Analytics and Improving Interoperability
Guest Dale Sanders, CTO of Health Catalyst, joins hosts Gary Austin and Ken Kleinberg to discuss the future of interoperability and health analytics and how decision support, population health, and easing provider burden can be improved with the right incentives to effectively unlock data to fuel more powerful analytics capability.
Gary begins the discussion with patient identification. He asks Dale why patients do not yet have a single healthcare identifier for their healthcare records. What is going to change this situation? Dale explains that the answer is a combination of a voluntary system for those in the commercial healthcare space combined with a mandatory system for those that are benefitting from Medicaid and Medicare. We need to commercialize the management of those patient identifiers, just like we’ve done with internet domains.
Ken states that TEFCA is looking at an infrastructure for a nation-wide health information exchange with a lot of these interoperability initiatives. He thinks we could do the same with patient identifiers if we could agree upon a dozen or so fields to try to do a better job capturing core data that could aid with patient identification.
Gary asks Dale why we can’t have data normalized across the board. Why is this such a difficult process? Dale explains that it boils down to economics. He suspects we’ve normalized enough to enable reimbursements but that’s it. If you think about it, does it matter to a single healthcare system to normalize to a national or international standard? Dale explains that it is probably not that valuable for individual institutions. They are normalizing to their own vocabulary, which works for them. If we’re interested in analyzing data beyond the boundaries of healthcare systems – as we need nationally for public and population health – we need to incentivize or mandate broader normalization of vocabulary and go way beyond LOINC codes.
Gary asks Dale if he sees the CDC mandating some of this for public health purposes post-COVID. Dale strongly believes it is time to mandate normalization. Ken notes that lab mapping is strikingly ineffective. There are physicians ordering tests by names that they know, but they don’t really know what tests they are actually asking for because what’s behind the scenes isn’t visible to them. There is no convention for how those things are named, which can be a problematic situation.
Gary asks Dale why there is not a standardized vocabulary. Why can’t the power of the computer be used to translate all of this in a normalized fashion? Is the power just not there?
Dale points out that human language alone is a very difficult thing to make sense of, especially the English language. If you layer on the complexities of clinical vocabulary, then it gets even more complex to turn that into computable, discrete elements. We can progress toward passive dictation, but there will always be a need for humans to make edits.
The discussion moved on to cover combinatorial data. Gary asks “How do you look at pulling this data together and rationalizing it across the two massive domains (administrative and clinical data) that are really driving healthcare in this country?” Dale explains the Health Catalyst data model. He says what you would see are domain and vocabulary-oriented data models that sit in between late binding and enterprise data modeling. There is this middle ground of curated data that is a manageable thing to keep up with and execute.
Gary asks Ken who he thinks is going to win by pulling all this payer and clinical data together. Is it the analytics companies, the EMR companies, big tech or who? Ken says there is this concept of a converged platform between payers and providers. Registries are an example of how pulling information together that a lot of people can use. The EHR vendors have taken a run at this, the analytics companies have taken a run at this. We’ve got the population health management vendors, some of which have been more on the business side of value based care, but there are also some that are on the tech side. It’s frankly a huge opportunity. The payers don’t have the same reliance on vendors that the provider side has. We have a very defined market of electronic health records vendors for providers, but how do you identify who the vendors are for payers? They are, in a way, a fortress to penetrate. I think Health Catalyst has a huge opportunity to sit in the middle and bring these two worlds together.
Dale suggests that in the future, payers will need to become providers and providers need to become payers. The payers are moving toward the provider space at warp speed, the providers are moving into the payer space at a snail’s pace. Providers need to be capable of working in the claims space, risk management and risk prediction space better than they are right now. The payers have to do the same working in the other direction in the clinical space.
Are the recent CMS ONC regulations around FHIR APIs going to make a real difference for healthcare moving forward? Dale believes the regulations are going to be transformative. However, for FHIR to be successful, he thinks we’re going to have to see more transformative changes in the EMR vendors. He’d like to see some aggressive re-architecting with the major software vendors to see if we can’t move that along faster.
Gary asks Ken to give a profile on consumer applications. Where are these things going? Ken thinks there are a couple of intersecting categories to consider. We have those apps that come under more FDA certification or approvals that are more medically oriented, and then others that will be more generally available that didn’t have to go through all those approvals. There is another category of apps associated very specifically with, for example, EHR vendors, health systems or payers. Then, there are other apps that are more independent. The interesting competitive angle here is that those independent apps that can potentially tap into provider systems and payer systems can put a picture together that any one provider or payer organization can’t equal. So it becomes more of a cross industry view and we’re talking about a longitudinal view and it’s usually longitudinal to the provider and the payer, but what it used to be was longitudinal to the patient. Whether those independent apps will succeed or not because they have a battle with the large players in their space. What CMS and ONC have delivered here is in fact that competitive environment far beyond the competition we saw in the past.
Gary asks Dale if he sees Health Catalyst getting into the consumer applications space. Dale says the data operating system they build has three missions of data, including analytics, workflow application development and interoperability. They can build applications on top of that data operating system. We have a care management platform that we call a data first application so it looks like an analytics engine but it is really about the workflow of care management. We’ll eventually put some patient-facing app on that as soon as the market calms down. That touchstone repository that I mentioned earlier is the national repository that I would eventually like to expose to the public so you, I and ken can query that national repository to other patients like me to see other patients like me where they are being treated, how they are being treated and eventually allow them to socially interact with other patients who are like them. We have to create social interactions between patients like all of us that’s what will keep people coming back.
Dale gives a primer about healthcare from a battlefield perspective. In the nuclear decision space, a lot of time is spent on subjective and objective data fusion as well as its reliability. The military has formalized decision making and the data required to support decision making. There’s a concept known as intelligent preparation for the battlefield, which can be overlapped conceptually onto healthcare. We can build the digital battlefield of healthcare that would support population health, public health, ambulatory care, the framework applies.
What is the biggest change we are going to see in healthcare next year? Dale says we will see a relaxation in some of the rules for FDA clinical trials. Additionally, he would like to see an increased emphasis on what we need to do to make quality measures more effective, less burdensome for clinicians.
