Press "Enter" to skip to content

Health Groups Press for Solution to Patient Matching Problem

WASHINGTON — Will 2020 be the year that Congress authorizes the development of a unique patient identifier (UPI)? The Health Innovation Alliance is hoping so.

“Patients are the biggest beneficiaries of a UPI,” said Joel White, executive director of the Health Innovation Alliance, a group of 75 different organizations — including 22 patient groups — focused on using technology and data to make healthcare work better. “We see a lot of medical errors in the system, we see a lack of care coordination, we see medication mismatches that result in not only patient harm, but higher costs and worse outcomes. A UPI will allow us to match patient records in a way that prevents those errors and produces better outcomes while hopefully lowering costs for individuals.”

UPI is a good idea for both patients and providers, White said. Mike Martz, chief information officer at Mount Nittany Health, a 260-bed hospital and health system in State College, Pennsylvania, said that at a previous hospital at which he had worked, a patient who was very ill came into the emergency department; he ended up having his gallbladder removed — a pretty routine procedure.

“His insurance denied him coverage, and the reason he was denied was because according to their records, he had had his gallbladder removed 5 years ago,” Martz said. “As it turned out, the man’s brother had used his insurance coverage to get his gallbladder removed 5 years ago. Privacy is not just something that impacts the patient, because guess who was left holding the bill? It wasn’t the patient; it was us. We had to eat those costs … The need for privacy is just as important to the provider community; it’s just as important to us as it is to patients, and we need to make sure the solutions we have ensure that not only is it accurate, but it does not get misused.”

The Current System

How good is patient matching currently? Martz noted that his own state of Pennsylvania has five different health information exchanges (HIEs), and “most HIEs I’ve talked to in the country are very proud of the matching rates they have using algorithms and trying to match for same last name, phone number, or email — they get up to 90% or 95%.”

Think of a state like Pennsylvania or Ohio, say, a state with 10 million people. “If we’re doing a matching rate with 95% — probably the best you’ll find in the country — it means we’re misidentifying 500,000 citizens in that state every year. No industry would find that kind of error rate acceptable,” he said. “It means for one out of every 20 people, we’re getting it wrong. People have been harmed, they have not been given treatment they should have had because we didn’t have all the information on them. They’ve been given treatment they shouldn’t have had because we had the wrong information on them. I know in this country people have been killed because of misidentification … It probably is happening almost daily.”

Katherine Lusk, the chief health information management and exchange officer at the Children’s Health System of Texas, in Dallas, said that when her hospital was beginning to get involved with three different HIEs and began sending them data, “all three came back and said, ‘You have a 20% duplicate rate. We’re going to merge these patients.’ We do not have a 20% duplicate rate; we have a 0.1% or 0.2% duplicate rate.” Lusk showed them that in fact, many of the records they were calling “duplicates” were actually separate patients who were “multiple births” — twins or triplets, for example — who had the same birthdate, same address, same parents, and same insurance.

“What would have happened had I not shown them exactly what was wrong?” she said. “I made a difference in my population alone; that’s not enough. We need to all be aware of this.”

The History of a UPI Ban

The problem for getting a nationwide UPI began in 1999, when then-Rep. Ron Paul, MD (R-Texas), a libertarian, inserted a ban on any government money being spent on a UPI into an appropriations bill. “There was some concern that the government was going to identify each one of us and take our information and watch us or control us,” White explained. Since then, Paul’s son, Sen. Rand Paul, MD (R-Ky.), has taken up the cause, and has argued against the idea. The House passed a measure last year to get rid of the ban as part of an appropriations bill, but the provision didn’t make it in the Senate, White told MedPage Today in an interview.

Congress needs to allow the Department of Health and Human Services to spend money to develop a unique patient identifier, said Joel White, executive director of the Health Innovation Alliance. (Photo by Joyce Frieden)

Martz disputed the idea that a UPI would be bad for patient privacy. “We can’t protect peoples’ privacy; we can’t protect their information, unless we can be certain we’ve got the right person,” he said. “It’s the exact opposite of the argument he’s making.”

In the meantime, the private sector has been working on a solution. Lee Ann Stember, president and CEO of the National Council for Prescription Drug Programs, in Scottsdale, Arizona, explained that there are three different types of matching methodologies that are commonly used:

  • Deterministic matching: This is what has been used historically; systems use demographic data such as information like name and date of birth to match up a patient. “But in the case of ‘Bob Smith,’ ‘Robert Smith,’ and ‘B. Smith,’ unless there’s an exact match, the record wouldn’t be identified as a potential match,” Stember said.
  • Probabilistic matching: “This is smarter technology” that takes into consideration that some names don’t match exactly, she said; it uses algorithms and statistics to determine possible matches.
  • Referential matching: This matching system bounces information off of more reliable databases of demographic data, and since those databases are continually updated, “it’s easier to identify someone who got married or changed their address,” she said.

“No single approach gives an ironclad guarantee of a match, but the combination of probabilistic and referential matching provides the highest matching rates,” she said. Stember’s organization has been working with Experian Health on the issue, and uses its Universal Identity Manager to help prevent misidentifications — with a 99.9% matching rate. It also has the ability to set up a UPI for each patient. Last year, the pharmacy industry approved use of her organization’s UPI “to be used behind the scenes, transmitting the patient identifier along with the patient data,” making it possible for a prescription to be transmitted electronically from the doctor’s office to the pharmacy, she said, adding that Experian Health has enumerated 100% of the U.S. population — an identifier for 328 million Americans.

Number Shouldn’t Be Public

Stember emphasized that a UPI “is not a number that’s known to patients or even healthcare providers. We believe if the number is known, there’s an opportunity for the number to be abused. The UPI is passed through the healthcare transaction using industry standards.” Stember’s group is making a no-charge offer to industry to use the Experian technology to identify patient matches and assign a UPI to each patient in their files.

Martz liked the idea, but expressed some caution. “I’m always suspicious of a single-vendor solution for everything; I like diversity in everything,” he said, adding that although the solution may likely be with a single database or structure, “it doesn’t need to be locked down to one vendor.” He also wondered for how long Experian’s system will remain free to use.

image
Biometrics will likely be part of the patient matching solution, especially for pediatric patients, said Mike Martz, chief information officer at Mount Nittany Health. (Photo by Joyce Frieden)

Biometrics also are likely to be a key part of the UPI solution, Martz said during a question-and-answer session, noting that the technology for that continues to get less expensive. “When you get into the pediatric space, biometrics may be the only solid way you can identify a newborn baby,” because other data like financial information don’t work very well in that population.

Although the private sector has been making strides, “It takes big gorillas in any industry to make significant change happen,” said White. “In healthcare, there’s only one big gorilla and that’s the federal government.” That includes having the Congressional Budget Office (CBO) estimate how expensive a UPI system would be. “Congress lives and dies by the CBO estimate; we want to make a case to CBO that this saves lives but also saves dollars.”

“This year we need Congress to step up; we need the ban to not be a part of the appropriations process; we need HHS [the Department of Health and Human Services] to go forward with a solution here as the largest payer of healthcare in the United States,” White said.

1969-12-31T19:00:00-0500

last updated

Source: MedicalNewsToday.com