Biometrics Are Only One Piece of the Patient Matching Puzzle

Mark LaRow, CEO, Verato
NOVEMBER 16, 2018
referential matching,biometrics healthcare,patient matching,patient identification
Referential matching technology could improve biometric solutions.

According to a landmark report published last month by The Pew Charitable Trusts, biometric technologies are one of four cornerstone opportunities to improve patient matching and interoperability nationwide. The report, which was the culmination of two years of research by Pew, also found that patients themselves are excited about the prospects of biometric technologies.

>> WATCH: Machine Learning: The Future of Medicine?

Another of the four cornerstone opportunities highlighted by Pew is “referential matching,” and it is the combination of biometric technologies with referential matching that finally offers an end-to-end solution to the patient matching challenge. Let me give three examples to demonstrate why every biometric solution should be complemented by referential matching technology in order to fully solve a health system’s patient matching challenges.
 

1) Biometric solutions will not help reconcile, match and link existing medical records

While biometrics do a great job of identifying patients, they cannot resolve the duplicate medical records that have accumulated in a health system’s electronic health record (EHR) system — which are 10-20 percent of all records. Nor do they help link existing records across a health system’s silos.

Think about it: If you already have three duplicate records for Jane Smith in your EHR, scanning Jane’s fingerprint at registration the next time she comes in won’t help you find or link those three existing records to the new record that has Jane’s fingerprint.

Instead, the only way to correctly associate duplicate medical records to Jane is by using patient matching technologies that use the demographic data in Jane’s records to link those records together. Yet we cannot expect the patient matching technologies in existing EHRs or enterprise master patient indexes to resolve the duplicates — those technologies created the duplicates in the first place. Instead, a next-generation demographics-based patient matching technology is needed, and that is referential matching.


2) Biometric solutions will not help link patient records across facilities that use different biometric modalities

Consider two facilities in a health system. One uses a fingerprint scanner, and the other uses an iris scanner. If Jane Smith gets her fingerprint scanned at the first facility and her iris scanned at the second, how can the health system link those two records together? The answer is that it will have to rely on demographics-based patient matching technologies that use Jane’s name, address, birthday and other data to figure out that Jane with fingerprint X and Jane with iris scan Y are really the same person.

This becomes especially important when health systems acquire hospitals or facilities that use different biometric modalities to identify their patients — or that don’t use biometrics at all. In these circumstances, demographics-based patient matching technology becomes the glue that identifies common patients shared with newly acquired facilities and prevents new duplicate records arising as a result of integrating new patient data sources. Only referential matching technology can do this with the unprecedented accuracy, speed and cost-effectiveness needed for organizations undergoing mergers and acquisitions.


3) Biometric solutions will not help link patient records across organizations that use different biometric modalities

Even if an entire health system adopts the same biometric modality and vendor across all of its facilities, that health system will still have to find and identify patient records in other health systems during health information exchange. And unless every organization in the U.S. adopts the same biometric modality from the same vendor, health systems will always need to fall back on demographics-based patient matching to enable interoperability and to find records during health information exchange.


So, what is the patient matching solution?

Ultimately, demographic data will continue to remain the “lowest common denominator” health systems will use to identify, match and link patient data within their systems and with other organizations. Which means that no matter how sophisticated or robust their biometrics strategy, health systems will need to complement it with an equally sophisticated demographics-based patient matching strategy.

Referential matching provides the end-to-end solution for patient identity and patient matching when deployed in conjunction with biometric technologies. It is the next-generation approach to patient matching, and it has become the new gold standard for patient matching technology after being adopted by the largest providers, payers and health information exchanges in the nation to dramatically improve their patient matching success.

Referential matching is truly revolutionary. It combines big data and cloud technologies with sophisticated new algorithms and a reference database of demographic data that it leverages as an “answer key” during matching. This means it is a quantum leap more accurate than other patient matching technologies, requiring no tuning and resolving more duplicate records with 50-75 percent less manual effort. It also means it can be deployed in weeks and is significantly more cost-effective than other patient matching technologies — it saves organizations hundreds of thousands of dollars in operational costs, on top of the IT cost savings of being a cloud-based technology. When combined with biometric technologies, referential matching offers a true end-to-end solution to solve health systems’ most pressing patient matching challenges.

Mark LaRow is CEO of Verato

Get the best insights in healthcare analytics directly to your inbox.

Related

Referential Matching: The Solution to Our Patient Matching Challenges
Does Your Healthcare Organization Have the Chops for Machine Learning?
4 Steps Toward Using Real-World Data

SHARE THIS SHARE THIS
21
Become a contributor