For years and years, hospitals and health systems have grappled with linking patients to the right record for every visit. The impact of not being able to connect disparate patient records across different providers is significant, as mismatched data could result in faulty or unnecessary medical care.
Many advocates for a nation patient identifier, including prominent healthcare associations such as HIMSS and CHIME, assert that a unique identifier would help improve patient privacy and safety, while propelling interoperability forward since data going from system to system would be far more accurate than it is today. However, research has shown that healthcare facilities fail to link records for the same patient as often as half the time.
This issue has generated interest among federal policymakers, and while the HIPAA law passed in 1996 required the creation of a unique patient identifier—which could address poor matching rates—time after time lawmakers have included language in appropriations bills blocking any federal spending on developing or finalizing standards for such an approach for more than two decades. The U.S. House of Representatives voted to remove that prohibition in July of last year, but the Senate has not yet followed suit.
In the meantime, it’s up to the private sector and individual health systems to figure out patient matching themselves. The 24-hospital Texas Health Resources is one organization that has taken the challenge head-on. Because of growth and new business relationships, like many integrated delivery systems, Texas Health had many separate systems generating and maintaining consumer data. So the organization turned to Verato, a healthcare technology company that develops a cloud-based enterprise master person index (MPI) platform for its end-user customers.
“The way we look at information on our patients is that people can visit [us] multiple times at multiple facilities, [creating] all different touch points,” says Michael Parris, vice president of data integration and analytics at Texas Health Resources. “So it’s about taking care of our consumers and making them healthier. In order to do that you really have to know your consumer and have to be able to meet them at a [particular] location or touch point they’re coming to you at. We saw that as a gap in our organization; when someone called on the phone, we pulled up their record, verified who they are, and half the time that conversation started from scratch,” he recounts.
As such, the goal for Texas Health leaders turned to continuing the conversation from the last time a patient “touched” the organization, whether that was via Epic’s MyChart patient portal, through the web, the phone, a clinic or facility, Parris explains. “We needed to pull all the identities and come up with a single version of a person to be able to attach what I call ‘attributes,’ and transactional data so we can then continue that conversation. From a customer perspective, it’s really about us knowing who you are and being able to serve you in a better manner,” he says.
Solving data quality issues
One of the core challenges for Texas Health and other health systems as they tackle patient matching is making sure that the data going into the MPI system is of high quality. Poor quality data makes it that much more difficult to match patients to all of their records. Parris notes that throughout patients’ healthcare journeys, they are seen by a variety of providers over the course of their lives. So over a span of just a few years, a person can change his or her address, last name, and even gender. “That makes tying a certain person together difficult,” he acknowledges.
For example, Parris offers, there could be a situation in which patient Mary Smith lived in an apartment on the eastern side of Dallas-Fort Worth when she visited Texas Health last, and then a year later, comes visits the health system and is now Mary Joseph living in northern Dallas. “How do I make certain I connect those records, especially when you first come in and give your name, address, and phone number, and those don’t connect [with each other],” he asks. “With Verato, we were able to piece that record together, not just across our venture but across our joint ventures, too. That allowed us to get a better picture of somebody.”
In the past, explains Parris, that person might have existed—especially with the traditional MPI—as two separate people, because there’s very little identifying a person except for the first name, “and you don’t want me to merge every Mary together.” But unless a patient tells the healthcare entity that his or her name changed and the record should be updated to reflect that, there wouldn’t be a way to know it happened.
Of course, when a patient steps foot into a healthcare facility, the last thing he or she is worrying about is back-end data. “They’re worried about their health,” says Parris. “So from the perspective of an MPI, being able to reference a data set that’s outside of Texas Health that has someone’s name, previous names, and previous addresses all pulled into one data set, allows us to pull Mary Smith and Mary Joseph together into the same person. That enables us to connect that person’s identity, allowing us to merge the electronic health record [EHR], and merge it across other systems to continue that conversation,” he says.
Healthcare is littered with “disjointed places where information might be given, but it also coincides with other information we’re bringing in to understand you and your needs better. There is that same need to match information up,” says Parris. He believes that’s where the value of an MPI comes into play, in that it performs referential matching—explained by Verato as matching the demographic data from each record to a continuously-updated database of identities spanning the entire U.S. population, as opposed to directly comparing the demographic data from two patient records—to pull these identifies into the same version of the person.
For Texas Health, the result of these efforts has been the creation of a patient ID for each of its 7 million patients, allowing the organization to get a full, comprehensive view of every patient. “Now we can understand how customers are journeying through our system,” Parris says. For example, he offers, going down the path of needing a knee surgery, does a patient call Texas Health directly, use the system’s website, or book an initial appointment using the MyChart portal? Once the patient has multiple consultations, gets imaging done, and schedules the procedure, there is an ability to look at just that person or thousands of people who have that orthopedic surgeon, he notes.
“So you can look down that chain and say, here are the 1,000 that actually had the surgery, and the 10,000 who did not. Where did they fall off on that journey?” That would not be possible without initially linking that data from all of the health system’s separate information sources. “So we can now connect [data] to get folks back on the journey to improving their health,” Parris emphasizes.
