THE FEATURE - AI AND HIE
Here’s how AI is transforming health information exchange (HIE)
If you’ve been to the doctor anytime in the past five years (and we hope you have!), you know this is true: Electronic health records (EHRs) are everywhere.
And now that EHRs are so common, any provider can easily share and access their patients’ health data, regardless of where else those patients received care, right?
Some nations have developed broad, nationwide health information exchange (HIE) infrastructure—such as Taiwan.
But here in the U.S., things in healthcare are a little more regionally-fragmented. And privately-owned. That makes HIE (the concept) a harder task. And that’s an understatement.
Enter: HIEs (the organizations). In this context, HIE entities are critical to maintaining and creating pathways for health data flow.
But wait, what are they again?
For the HIE-unacquainted, think of them like health data post offices. If post offices also translated and organized your mail to make it more relevant to your recipient.
Now, more literally: These often regional health IT organizations are responsible for normalizing and organizing data flowing through them from different health systems and EHR vendors. Many of them also handle data analytics and public health reporting.
We’re excited about HIEs because they’ve now begun implementing AI in their operations and data analysis to improve their functionality. Today, we’ll be diving into 3 ways AI helps HIEs be more effective—on the population level, on the patient level, and on the financial side..
The Big Picture: Improving HIE public and population health analysis
In 2020, HIEs were right there with providers on the COVID-19 front lines.
That’s because, when it comes to what’s going on in public and population-level health in any given area, HIEs are often the ones to see it first.
Take a look at a handful of HIEs’ services, and we bet you’ll see public health dashboards—from COVID to Legionnaire’s disease.
But tracking outbreaks is not the only way HIEs compile and provide authorized users access to critical population-level health information.
Many HIEs have started using machine learning to predict health outcomes. Doing so, they, give providers and health systems better insight as to which patients of theirs are at risk of poor outcomes or hospitalization.
One such use case is in a grave area of American public health: battling the opioid crisis.
HIE predictive alerts help providers better target their addiction interventions, since "part of the behavior of people that have this addiction is they visit a lot of different EDs at a lot of different health systems, and it's very hard to put the puzzle together," said Don Woodlock, vice president of InterSystems’ AI-enabled interoperability platform, HealthShare.
The Patient-Level Picture: Making HIE data more clinically actionable
Speaking of HIE alerts—as you can imagine, these can be a game-changer for clinical decision-making.
Especially when it comes to getting a patient’s many providers on the same page, data-informed care coordination can make a huge difference.
For instance, New York-based HIEs Healthix and NYCIG have used HealthShare to more easily de-duplicate records. They also run AI-enabled real-time clinical alerts to patients’ providers that can make a difference in time-sensitive care decisions.
And with the help of similar predictive analytics, Rhode Island Quality Institute has reduced its own patient hospital readmission rate by 15 percent since implementation.
The Financial Picture: How AI helps improve HIE ROI
Of course, our readers already know: One of the biggest ways we can harness AI in healthcare is by having it help us manage our data.
Earlier this year, we discussed how AI can help us make better use of healthcare’s abundant dark data. You might remember talking about all the dollar signs healthcare companies flush down the drain when they store endless amounts of this largely unstructured, unanalyzed data.
Naturally, HIEs receive a lot of this data. They’re responsible for working with the healthcare company to normalize and validate the data. And ultimately, make it usable and accessible to other healthcare companies in the network.
For health systems already investing significant time and resources in HIE connections, this is a big value-add. Especially if it can help them better position themselves for value-based payment.
This is a tangible example of what we’ve been discussing in terms of AI’s big promises for healthcare data management.
Of course, as we make big strides in interoperability, the existence of regional HIEs does serve to remind us how fragmented healthcare data exchange still is in this country.
In our wildest dreams, we see a future where our health system’s data flows seamlessly and regional HIEs are redundant. But for now, they’re key to working toward greater interoperability, care coordination, and public health response.
AI making them more effective is a win for us all.
Plus, HIE data is an invaluable data source for AI researchers. Such as those who were able to use HIE data to build a model that could forecast individual COVID-19 hospitalization. When AI helps HIEs succeed, their data in turn continues to be available for these fascinating (and crucial) purposes.
Feeling inspired? Now, think big. What do you think HIEs should use AI for next? Reply back and let us know.