Patient matching is the quiet failure point in most interoperability programs. Published research has found match rates within a single organization commonly around 80 percent and rates as low as approximately 50 percent when records are exchanged between organizations, which means a technically correct query can still return an incomplete record. Improvement comes from demographic data capture at registration, standardized address and name handling, measured match performance, and a named owner for duplicate resolution — not from changing transport standards.
A correct query can still return the wrong record set
Interoperability programs are usually measured by connectivity: is the endpoint reachable, does the query return a 200, does the document parse. None of those measures tell you whether the record you received belongs to the patient in front of the clinician, or whether half that patient's history stayed behind under a second identity.
ASTP/ONC defines patient matching as identifying and linking one patient's data within and across health systems, accomplished at minimum by comparing demographic fields such as name, date of birth, phone number, and address. Every weakness in how those fields are captured shows up later as a missed link or a false one.
What the published evidence shows
Work published by The Pew Charitable Trusts in 2018 found match rates within a single organization commonly around 80 percent, and rates as low as roughly 50 percent when records were exchanged between organizations. An earlier report prepared for ONC reached similar conclusions about the role of demographic data quality and the absence of standardized capture practices.
Read those figures as a planning assumption rather than a benchmark to hit. They mean that any program relying on external data — TEFCA queries, payer-to-payer exchange, trial pre-screening — should be designed to expect incomplete linkage and to make that visible instead of silent.
Where intervention actually pays
The highest-yield changes sit upstream of any matching algorithm, in registration practice and field standardization.
- Registration capture: required fields, verification steps, and staff guidance for nicknames, hyphenated names, and shared addresses.
- Standardization: consistent address normalization, name parsing, and phone formatting applied before comparison.
- Measurement: track duplicate creation rate, potential-duplicate queue volume and age, and manual override frequency.
- Resolution ownership: a named team that works the potential-duplicate queue on a schedule, with an audit trail for merges and unmerges.
- External exchange handling: define what the workflow does when an inbound query returns partial, conflicting, or no results.
Governance decisions you cannot defer
Matching thresholds are policy choices with clinical consequences. A looser threshold raises linkage and raises the risk of a wrong-patient merge; a tighter one keeps records apart. Those trade-offs belong to a group that includes health information management, clinical leadership, privacy, and IT — documented, with the chosen thresholds and the review cadence written down. Where a program treats thresholds as a vendor default, no one owns the outcome.
How BlockInterop supports this work
- TEFCA and QHIN Readiness consulting and implementation — Assess organizational readiness, define participation strategies, support onboarding, and operationalize nationwide health information exchange.
- Testing and Implementation Support consulting and implementation — Support interface validation, workflow testing, deployment readiness, issue resolution, and production optimization.
Sources
Every figure cited in this article links to its public source. BlockInterop publishes no client names, outcomes, or internal performance statistics.