Patient Matching: How the Lack of a UPI Affects Interoperability
Understanding the complexity around matching patients’ health records and brainstorming solutions
Imagine you’re John Doe, a patient who had a fairly routine visit with your primary care physician. Post-visit, your insurance processes, and one to two months later, you receive a balance bill stating the patient responsibility is over $1,000, which sends you into a panic. Something is very wrong here.
Come to find out, you, John Doe, have suffered a case of mistaken identity. We’re not talking about identity theft of the credit card or financial fraud variety, like the Capital One commercials — but a medical records mix-up, which, in worst-case scenarios, could be fatal.
Turns out, the lack of a Unique or Universal Patient Identifier (UPI) could be to blame.
Many hospitals employ Enterprise Matching Patient Index (EMPI) tools, to eliminate duplicate and inaccurate records, with success rates as high as 93 percent — versus a scant 24 percent success rate without EMPI tools — according to a Black Book Research study, and when limited to in-house registration and record exchanges. But this isn’t how patients typically receive services. Rather, patients are treated by a variety of physicians and specialists, in different care settings and locales, over the course of their lifetime. When patient data is exchanged across different healthcare organizations, up to 50 percent of patient records are not matched correctly.
In the U.S., inaccurate patient identification is estimated to cost the healthcare system more than $6 billion annually.
“The real challenge of identity management and parsing together a longitudinal health record has to do with integration and interoperability,” according to Doug Brown, at Black Book. “Many systems still do not communicate and store data in disjointed architectures and an upsurge of identifiers continue to be created.”
Many liken lack of UPI to trying to get health IT data systems to “speak” the same language — whether it be EHR, CRM, HIE and other third-party systems, interoperability in this mindset serves to translate. While UPI misalignment is truly impacted by the inability to connect systems, we must step back and ask a larger question: how do we, as a country and culture, make significant progress toward interoperability when we can’t guarantee each patient’s records will land on the same page?
“Our customers are affected by the lack of a UPI because even in the simplest workflows, patient admissions for example, you still have room for errors and duplicates. You don’t want incorrectly matched patient information to contaminate the system of truth, which is usually the EHR,” said Josh Douglas, Chief Technology Officer at Bridge Connector.
Interoperability in its current state: how we got here
When electronic medical records (EMRs) emerged around 20 years ago, one of the original purposes was to bill — “to collect money more quickly,” according to Douglas. “Matching a patient across different systems wasn’t needed to simply file a claim, so it wasn’t a concern at that time,” he said.
As EMR functionality grew, so did EMR vendors’ potential to create walled gardens, making out-of-network referrals less appealing, if even possible, for both providers and patients.
EMR functionality has increased by necessity since then, to encompass the total health of the patient. Thus, the resulting systems have been increasingly referred to as electronic health records (EHRs), to reflect the industry’s need to “reach out beyond the health organization that originally collects and compiles the information” — promoting the sharing of information across multiple healthcare organizations.
EHR adoption was initially slow but picked up speed in the late 2000s due to the American Recovery and Reinvestment Act.
As of 2015, and as re-stated in Office of the National Coordinator (ONC) for Health Information Technology’s 2018 Report to Congress: 96 percent of non-federal acute care hospitals and 78 percent of office-based physicians have adopted “certified health IT.”
But there is a huge difference between certified health IT and interoperability of health information systems.
“What you have to realize is that each certified health IT system is, in effect, a stand-alone system,” said Douglas. “Patients receive care from multiple providers, and this will only increase as our population ages. The need for specialists in different areas becomes more real, and of course each specialist is using their own separate system.”
Here’s how discrepancies can appear for just one patient, across their network of care providers:
To provide true, quality patient care, you need to have a patient-centric focus across all of those systems — a feat only possible through health data integration.
We will continue to make a case for interoperability of health information systems as a means to accomplish both patient-centered and value-based care. And we have also explored interoperability as one of the best ways to leverage clinical data to solve some of the major business problems in healthcare, with referral intake management integrations as an example.
Despite HIPAA’s original mandates in 1996, one of which was to establish “a standard unique health identifier for each individual, employer, health plan, and healthcare provider for use in the healthcare system,” Congress has denied funding to establish a Unique Patient Identifier ever since. They’ve cited “Big Brother” concerns at every step and continue to stand behind the 1998 law they created, prohibiting the creation of a UPI “until legislation is enacted specifically approving the standard.” It’s the ultimate game of cat-and-mouse, which has thwarted real solutions.
How pervasive is the problem — the lack of a UPI?
A real-life patient example, using the name “Maria Garcia,” in “Harris County, Texas”, was created by the Bipartisan Policy Center, to illustrate just how prevalent patient-matching issues can be. There were 2,488 real patients with that name in the county, 231 of whom had the same birth date. Meaning, there were 69,807 pairs of patients who share both names and birth date — just in Harris County, Texas. “Extrapolate this to the entire nation,” the report suggested, “and the importance of accurate patient identification becomes starkly apparent.”
But let’s get real. “Even though a UPI would solve a lot of the current patient matching challenges,” added Bridge Connector’s Director of Analytics, Matt Wimberley, “for a UPI to be effective, that also assumes that the patient has been admitted under non-trauma circumstances, and that their information has been recorded correctly.”
There is no such thing as a 100 percent, fool-proof solution. But that’s not to say there haven’t been more solutions discussed — or requested — in recent years.
In early 2017, the ONC identified that we need to enact other creative solutions to create a UPI , particularly because they aim to encourage more interoperability across systems. They created the Patient Matching Algorithm Challenge to this end, where developers could win $75,000 in prize money.
And in early 2019, the ONC called for a Request for Information (RFI) for patient-matching solutions, while the Government Accountability Office revealed findings from its report, “Approaches and Challenges to Electronically Matching Patients’ Records across Providers.” The report identified the following possible ways to improve matching:
- Implement common standards for recording demographic data (example: ONC could require these standards as part of its responsibility for certifying EHR systems, or the ONC could call for the voluntary adoption of the same)
- Share best practices and other resources
- Develop a public–private collaboration effort
But in summary, the report reaffirms that no single effort will solve the challenges around patient record matching.
Here are some solutions, by topic, that have been discussed, however:
Blockchain has become synonymous with “decentralized” data, and “isn’t practically possible in the U.S., without some sort of reform to HIPAA, which I don’t see happening anytime soon,” said Douglas.
Use of SSNs as UPIs
“First of all, not everyone who lives in the U.S. and receives healthcare here has a social security number, nor is your SSN required to receive care,” said Wimberley. “Not to mention, people are rightly hesitant to share their SSN with the outside world. Identity theft and financial loss have been more closely tied to SSN misuse.”
As technology improves, possible solutions to the UPI problem have “bubbled to the surface,” according to Douglas, the most interesting of which he believes is the acceptance of a genetic marker as a patient ID that all systems would use.
As for how a “genetic marker” UPI system might work, consider the at-home genetic testing kits such as AncestryDNA and 23andMe. You might receive an oral swab kit, similar to how these services currently operate, send it back to a larger processing facility, and they create a profile, unique to you. If you could somehow tag a person’s DNA to them as a “patient,” that would become their UPI — unmistakable, plus it addresses privacy concerns around linking medical records with other personal or financial data.
Douglas spoke at a Nashville Post Vitals panel in late 2018 that creatively addressed UPI solutions, including the genetic marker idea and the idea of a government mandate to use one patient ID across EHR and ancillary systems.
“It’s clear from our clients’ experiences, that even without a national UPI of some variety, our integrations between systems still provide all of the value they were seeking,” Wimberley said. “It just makes it harder for all of us in health IT to do our jobs — a shared pain point that really everyone in the industry has experienced for the last 20 years.”
“It’s part of our jobs at Bridge Connector to establish best practices for our clients across a wide set of criteria — including being aware of advances with patient-matching strategies,” added Douglas. “We just keep doing what we’re the experts in, building solutions that connect disparate data systems. But we’re always going to handle that data in different, unique ways, because that’s how we keep delivering unprecedented, maximum value to each of our clients.”