It isn’t every day that the House of Representatives takes bipartisan action to reverse a policy that’s been in place for two decades. But that’s what happened last month, when Democrats and Republicans alike voted for a measure designed to address a perennial problem that undermines medical record-keeping, puts patients at risk, and costs our health care system billions of dollars every year.
Specifically, the House voted to repeal a 21-year ban on funding for a national patient identifier — a unique number or code comparable to a Social Security number that would be assigned to each and every American. As envisioned, this identifier would make it easier for health care providers to access accurate medical records anywhere, anytime — whether the patient is making a routine office visit in Boston or lying unconscious in a San Francisco emergency room.
A unique patient identifier could help ensure that people with the same name don’t get mixed up, even if they live on the same street or are born on the same day. Although electronic health records (EHRs) have helped improve care overall, health systems still struggle to link patients to the right record on every visit, and attempting to match records between hospitals can fail up to half the time. That harms patients.
At a minimum, trying to find the right records can delay needed care. Worse, inaccurate or incomplete records can lead doctors to order unnecessary tests or miss allergies, medical implants, dangerous drug interactions, and other relevant facts. The administrative costs for fixing mismatched and duplicate records can be enormous. The Mayo Clinic, for example, found that it cost at least $1,200 to untangle each mismatch and correct a patient’s record.
This House vote opens the door to establishing a national identifier. But the bill must still pass the Senate.
The original ban was written in 1998 by then-Rep. Ron Paul (R-Ky.), who believed that a national identifier represented an unwarranted government intrusion into private health care. Those who believe in a unique patient identifier argue that it is a sound public policy that would protect patients and save costs.
The debate about a unique patient identifier is worth having — and should be informed by international experiences with them. Although many developed countries use such identifiers in health care, the systems face challenges: Duplicate identifiers can be erroneously assigned, patients can misplace their identity cards, and numbers can be used fraudulently, among other problems. So even countries that use patient identifiers typically still use algorithms that leverage names and other demographic data along with the identifier to better establish identities.
If the United States were to establish a unique patient identifier system, policymakers would need to learn from these experience and recognize that identifiers don’t fully solve the problem of matching patients to their correct medical records.
Fortunately, a unique ID for the 21st century could take multiple forms, because technology has evolved in the two decades since Congress originally envisioned the concept. Smartphones and email addresses can be used as long-lasting identifiers, and we increasingly use biometrics such as facial recognition and fingerprints to unlock our devices or identify ourselves at airports, in banking, for security systems, and even at sporting events.
Biometrics are already part of our lives; when used in conjunction with demographic data, they make the risk of confusing one person’s ID with another’s almost nonexistent.
In a series of focus groups conducted by the Pew Charitable Trusts, respondents overwhelmingly chose biometrics as the preferred option for a medical identifier over smart cards, unique numbers, or usernames and passwords. As many participants noted, a smart card or ID number can be lost or forgotten, but a fingerprint or face scan could almost always be used, even when an individual is unconscious. If the United States goes ahead with the development of a system of unique patient identifiers, it should give serious consideration to using biometrics — and the technical and privacy issues they introduce.
Even without a national system, biometrics are now being used within individual health organizations, though not typically for exchanging data among them. The Office of the National Coordinator for Health Information Technology, the federal agency that oversees electronic health records, should begin exploring how these technologies are already being applied, the benefits they provide, the privacy considerations they raise, and the technological challenges to using them to improve rates of matching patients with their medical records.
In the meantime, there are near-term steps the government should take to improve matching. Today, hospitals use multiple data points to match records to the right patient, such as the patient’s name, birth date, address, and other identifying information. But different electronic health record systems don’t always record that information in the same format. Some, for example, use “Street” in addresses, while others use “St.”
There is one relatively simple solution for that issue: Hospitals and EHR vendors could format addresses using the U.S. Postal Service’s standard — the system that online retailers use. When a shopper enters an address online, the standard postal address pops up to streamline shipping and delivery.
A study by researchers at Indiana University found that applying the same standard for addresses in medical records could improve match rates by a small, but important margin — enough to correctly match tens of thousands or more records each day. These gains would accrue even without a national identifier and help support accurate matching if one is eventually established.
As part of new rules designed to aid the exchange of health data, the Office of the National Coordinator for Health Information Technology should require that the postal service format for addresses be used by all electronic health records when sharing patients’ records. In addition, it should require all EHR vendors to use data already in their systems but not typically used for matching — such as the patient’s email address — to link records.
Although the outcome of the debate on a unique patient identifier remains uncertain, there’s no reason to delay taking advantage of solutions that are available now to tackle the patient matching problem.
Shaun Grannis, M.D., directs the Center for Biomedical Informatics at the Regenstrief Institute. John D. Halamka, M.D., is professor of international healthcare innovation at Harvard Medical School. Ben Moscovitch directs the Pew Charitable Trusts’ health information technology initiative.