Cell phone numbers could provide a fast-track solution to matching patients with their health records

Featured in Stat News

When the U.S. House of Representatives voted in August to overturn a ban on using federal funds to establish a national patient identifier — a unique health ID number for every U.S. resident — the move was widely lauded by health care groups.

A national patient identifier could be the unambiguous thread that would tie all medical records across all health care organizations to the correct person. It would provide the basis for improved care coordination and enhanced patient safety by ensuring that physicians have access to comprehensive medical data at the point of care.

It’s certainly a step in the right direction. Today, 30% of patient records are linked to the wrong people for a variety of reasons, such as when data in health care databases is incomplete or out of date. A universal patient identifier (UPI) would help ensure that the right records are matched to the right patient.

Even if the Senate also threw its support behind a universal health ID, it would be five years or more before we’d see an improvement in patient-matching accuracy because of the length of time it would take to create, implement, and distribute these identifiers. But there is a data source that could more quickly provide a fast-track solution to patient-matching challenges: cell phone numbers.

As someone whose company helps health care organizations accurately match their patients with their health records, I see great value in using these now-ubiquitous numbers in the health care setting.

Why a UPI isn’t enough

The pandemic underscored the critical need for universal patient identifiers as breakdowns in patient matching made it difficult to track patients who tested positive for Covid-19 or make them aware of their test results.

But there are numerous reasons why such an identifier should not be the only patient-matching solution used in the U.S.:

Implementation won’t be simple and it won’t be cheap. One estimate suggests a UPI could cost $3.9 to $9.2 billion to implement. take five years to distribute to 50% of the population. Achieving 80% coverage could take 10 years.

UPIs won’t deliver perfect match rates. The unique identifiers already issued on a national level — think Social Security numbers — do not achieve perfect match rates. It’s not uncommon for numbers to be transcribed incorrectly or misremembered. They are also stolen, forgotten, and mishandled in processing. Some experts predict universal patient identifiers could fail to identify patients 2% to 10% of the time because of human error.

The process of tying UPIs to patients’ medical records won’t be quick. It will take at least a generation for universal patient identifiers to permeate into medical records at thousands of health care organizations across the country.

The question is not whether universal patient identifiers would be useful. The question is, how much should we invest in them?

How cell phone numbers could make the difference

On their own, universal patient identifiers won’t strengthen patient matching. They must be paired with demographic information to ensure that health care providers have a way of contacting patients after care is delivered or lab tests have been performed. Without a way to verify patients’ demographic information — such as through referential matching, which spots errors in patient information by comparing the data against a reference database of U.S. identities that is continually updated — providers will be unable to reach patients quickly when minutes count. It’s a situation that became clear during the pandemic, when commercial labs were missing as much as 40% of demographic information, making it difficult, if not impossible, to notify some patients who tested positive for Covid-19.

Cell phone numbers could achieve most of the value of a government-distributed universal patient identifier at virtually no cost, and 96% of adults and 83% of teens already have them.

Cell phone numbers are unique. They are also “sticky,” meaning they follow their owners when they move from one place to another. Not only are cell phone numbers an effective identifier, but they can also be used to contact people directly. And, because cell phone numbers are 10 digits long, they can easily be captured at the point of registration in the same way that a nine-digit Social Security Number can.

When consumers change their cell phone numbers, the U.S. has an existing management facility — the Number Portability Administration Center — that can make the appropriate adjustment. The same HIPAA parameters that would apply to a patient’s UPI or Social Security Number would also apply to a cell phone number used for medical record identification.

While it would take five to 10 years for universal patient identifiers to be implemented for the majority of the population, using cell phone numbers to match patients to their records could be achieved in two years or less, given the high percentage of the population that already has a cell phone number and the infrastructure available for cell phone verification.

Would using cell phone numbers to match patients to their medical records be a cure-all for patient-matching challenges? Not on its own. Any enumerator — whether it’s a Social Security number, driver’s license number, cell phone number, insurance card number, or universal patient identifier — cannot be relied upon to be perfect. Demographic information will still be needed to round out the verification process, and a patient’s demographic information must be verified using referential matching. But if leveraging cell phone numbers to improve patient matching could yield marked improvements in 2021 or 2022 — not 2026 — by using a highly reliable identifier that is already nationally available.

Combining cell phone numbers with referential matching could reduce patient matching errors to levels we’ve never seen before. This is the type of research that the federal government should undertake before undertaking a lengthy, expensive move to universal patient identifiers.