IMT commends Pew Charitable Trusts for accelerating the patient identity record matching conversation. Their recent report, Enhanced Patient Matching Is Critical to Achieving Full Promise of Digital Health Records, nicely summarizes opportunities for improvement in four areas. Perhaps most importantly, Pew emphasizes the need for a nationwide strategy and comprehensive solution to truly, finally address the patient matching challenge.
IMT sees the critical need for change through our work with 180+ clients. Every day, we see the stakes of accurately identifying and linking patient records at the point of registration and beyond. Creating accurate patient identity information during the first encounter empowers interoperability, value-based care initiatives, research, patient safety, and virtually every facet of data sharing.
But how do we get there?
The Pew report reviews the decades-old patient record matching challenge, including a vital discussion about algorithms. IMT uses all the data elements a client provides for record matching, including historical names, phone numbers, and addresses. Additionally, common issues — character transposition, typographical errors, default values, and so on — are carefully and systematically evaluated and scored by a probabilistic algorithm that can better detect linkages. Alas, the majority of healthcare technology solutions still use deterministic or rules-based matching that inhibits true data interoperability and strategic and operational goals. One of our clients reported that prior to deploying the IMT solution, 36% of lab results failed to connect to the right patient’s EHR, with serious implications.
The 4 Opportunities
The Pew report identified four key opportunities for improvement:
One: Unique Patient Identifiers. A unique identifier could take many different forms, such as biometrics or a government/private entity issued identifier. While Congress has prohibited funding to explore these options, the private sector has advanced biometrics slowly in healthcare. But a lack of standards for biometric data interoperability inhibits their widespread use. The consumer focus groups organized as part of the Pew study preferred unique identifiers to resolve identity inconsistencies, with biometrics being the first choice. Provider focus groups indicated support while recognizing biometrics as just one more data point to assist in matching — and they also saw the lack of biometric standards as an impediment. The report discusses several options for moving forward, including having the soon to be formed entity for the Trusted Exchange Framework and Common Agreement (TEFCA) lead the discussion for biometric standards.
Two: Patient-Empowered Solutions. The August 2018 RAND report, funded by Pew, explored potential solutions to patient matching that are empowered by patients themselves. A patient-empowered approach might entail having patients validate their cellphone numbers and other key attributes — an innovative method of potentially boosting both consumer engagement and data quality. Consumers and providers both raised many concerns during the focus groups, namely loss of cellphone or population segments lacking cellphones or not adept with technology. Pilots will be key to testing the strength of this approach and addressing concerns.
Three: Demographic Data Standardization. Pew advocates for standardizing the data elements that impact patient matching. During the Pew webcast, Dr. Shaun Grannis noted that standardized names and addresses improved match rates by double digits in research conducted by Regenstrief Institute. Numerous bodies have made this recommendation over the past five years, including ONC, ECRI, Project Sequoia, and others, and could be advanced by TEFCA when the coordinating entity is formed.
Four: Referential Matching. Referential matching overcomes some of the existing challenges by using data from credit bureaus, public databases, and key retail sources that heavily invest in high quality, current data. The specific data varies between vendors, but all have cited very high match rates and the ability to scale to a national solution. Data limitations exist, particularly among minors, undocumented immigrants and homeless people. Consumer focus groups expressed concerns, and Pew advises an independent analysis of referential matching’s accuracy and improvement claims.
IMT sees merit and benefits to all approaches, but sees Opportunity 3, Standardization, as offering the broadest potential. So many organizations rely on deterministic or rules-based matching (rather than probabilistic algorithms) that would benefit from standardized data elements.
We have worked in countries (and states and provinces) that use standardized unique identifiers for payment or healthcare delivery purposes, as well as in the government, human services, and justice sectors. Yet many of these same entities view standardization as just the first step towards a solution. The most successful entities layer in a mix of technologies and identifiers to achieve their goals.
We’re excited by the discussion sparked by the Pew report, and we want to be part of the solution. By blending probabilistic technology and strong matching practices with new types of data (whether biometric, patient-empowered, referential, or a standardized element) we can help finally crack the patient identification and record linkage challenge once and for all.
