Most hospital executives and administrators believe the accuracy of patient matching when information is exchanged between organizations with different EHR systems is subpar, according to a recent Pew Charitable Trusts report.
For the report, Pew partnered with the Massachusetts eHealth Collaborative to survey experts and executives from 18 U.S. hospitals, physicians' offices and health information exchanges. Participants were interviewed on their current experiences with patient matching, influencing factors that that could lead them to make adjustments and ways to solve the problem.
Five report insights:
1. Interview participants said they experience more issues and inaccuracy with external data exchange than internal exchange. Respondents said they consider match rates of 99 percent or higher ideal for interoperability.
2. Almost every participant presented difficulty when asked what their organization's current match rates are. Many participants said their organization only measures the number of records identified as duplicates, and not how often records obtained from other organizations can be linked to records on file.
Participants also said they don't know what percentage of their records are unlinked because they are not aware of all the records that should be related.
3. It's usually easier to match records when information is sent to facilities that are exchange partners or are expecting the data, participants said. Unsolicited inbound records or requests can present patient matching challenges because many have no record for the patient or have outdated information related to the individual.
4. Rural healthcare providers said their organizations experience a lesser need for matching records because their patient population often does not seek care at other organizations.
Urban healthcare providers said their patients have sought care in multiple places, which has enhanced their need for accurate matching records.
5. When presented with four solutions to enhance the accuracy of matching records — use of unique identifiers, smartphone use, standardize demographic data and use of referential data — interview participants said each option shows potential, but none are likely to solve the matching problem on its own.
To access the full report, click here.