National Patient ID – All Talk, No Action. Why??

We have been talking about a national patient identifier in the USA for decades.  And well-respected organizations like Rand, ASTM, Sequoia, ONC, and many others have studied the patient matching problem, issuing a series of reports that confirm what we all know — there are serious problems associated with mismatching including unnecessary testing, medication errors, and even death; negatively impacting quality, efficiency, and patient/clinician satisfaction across the entire healthcare system.

We all agree…but few are willing to actively pursue a solution.  Why?

  • It is complicated and any proposed fix is costly. Virtually all healthcare organizations have multiple IT systems and dozens of them may be involved in the care of any one patient, including billing.  Getting all vendors to support a single approach and then upgrading these systems is a massive effort.  So, it’s easier to put an “enterprise patient matching index” (EMPI) on top of them to match patients using personally-identifiable information (PII) such as name, address, birthdate, etc.
  • EMPIs fail about 8-10% of the time. Multiple studies have shown that patient matching is fraught with difficulties.  Different spellings of names, changes of address, changes of gender, and the fact that matching is often done under time pressure to get the patient moved to where they can get care – all of these contribute to matching failures.
  • Interoperability makes things worse. Despite federal initiatives to increase data sharing among healthcare organizations, these providers generally do not share patient identifiers and, therefore, must rely solely on matching PII to ensure that patient records from different organizations belong to the same individual.  In these cases, the identification error rate can skyrocket to 40% or even more.  And with increasing pressure to “interoperate”, these errors will only increase if we don’t solve the patient ID problem.
  • More data for matching leads to more problems. A classic response is that “maybe we just need more data.”  The assembly of massive databases of PII – while they can lead to some reduction in patient mis-identification incidents – raise enormous problems in terms of privacy, manageability, maintaining proper updates, and cost.  Identity theft and massive data breaches are just two of the complications that result.

This first in a series of GPII blog posts is published in hopes of drawing attention to the possibilities for a real solution to the patient ID problem.  We need a solution that can protect existing health IT investments; take synergistic advantage of both legacy technologies and emerging ones (e.g., biometrics); avoid large data bases and security risks; support patient empowerment and privacy; be implemented at a reasonable cost; and ensure providers and their patients that the information they’re using to make important healthcare decisions is the right information.  Watch this space!  And go to for a more in depth discussion of patient identification issues.