Upcoming Study Aims to Make Electronic Health Record Matches More Accurate

For clinicians to seamlessly coordinate care for the patients they share with other practitioners, they must be able to accurately link medical records across facilities. An upcoming study will shed light on a key barrier to that linkage: the collection, formatting, and use of address data, phone numbers, and other demographic data that is typically compared to identify individuals.

The study, which is being conducted by The Pew Charitable Trusts, the Center for Health Information Partnerships at Northwestern University, and the Regenstrief Institute, aims to ascertain which demographic data health care organizations most frequently use for this process—known as patient matching—and how it is entered in electronic health record (EHR) systems. Using the same data points and formatting them in the same way can allow different systems to better match records among shared patients.

In the first part of the study, facilities will complete a questionnaire about the demographic data elements they collect, as well as how those elements are documented and formatted. The researchers will subsequently run an analysis on a subset of organizations’ databases to assess how often these facilities collect this data in practice. For example, if an organization’s policy is to collect a patient’s previous mailing address or email address, this step will aim to determine the percentage of records that contain this information—without sharing or exposing identifiable information. The responses to the questionnaire (which will be coded so they do not identify specific organizations) will be used to develop recommendations for which data elements should be used for matching and which formats should be used when entering them into EHRs.

The inconsistency in collected data as well as a lack of uniform standards for recording it hamper effective sharing and record matching. Although nearly all hospitals and most office-based practices use EHRs to house patients’ health information, one estimate found that attempts to link patients’ records held in different places may fail up to half the time. Variations in how or whether systems record and exchange demographic data complicate those matching efforts.  

Information from this study will inform policymakers as they work to improve the accuracy of patient matching. Recent federal rules included a basic set of information—known as the U.S. Core Data for Interoperability (USCDI)—that all EHRs should be able to share in a standard manner; this dataset already includes some critical data for patient matching. In addition, the federal government continues to develop policies to support nationwide data exchange, called the Trusted Exchange Framework and Common Agreement (TEFCA). Future versions of the USCDI and TEFCA could include additional requirements for demographic data standardization—for example, using the U.S. Postal Service format for addresses—to make data sharing easier and more effective.

The study will also examine how atypical identity information is documented, including biometrics such as fingerprints or photos. In the longer term, these emerging technologies could also help improve matching, but more research is needed on how to collect and share biometric data in a manner that makes data exchange easy but also protects privacy.

The results of this study could provide some clarity for policymakers on the current patient matching landscape, including how to make this process easier and the results more accurate so that patients can benefit from more coordinated care.

Ben Moscovitch directs The Pew Charitable Trusts’ health information technology initiative.