Mismatch Game: PMP Errors Could Harm Patient-Physician Relationship



Oct_20_TM_Law_PMP

Now that Texas prescribers must check a patient’s history in the state’s prescription monitoring program (PMP) before prescribing opioids, plus three other drug classes, the errors are becoming more apparent.

When physicians do check, there’s a chance they’ll unwittingly miss a patient’s existing entry in the database. They might even find themselves looking at the wrong patient and the wrong drug history on their computer screen, which might make the person sitting in their office wrongly look like a drug abuser.

Such errors can happen, for example, if physicians search a patient’s common nickname instead of his or her given name – like “Bill” instead of “William” – which won’t allow them to see their patient’s PMP data and prescription history under “William.”

Reconciling such discrepancies is called patient-matching, and the PMP includes technology to do so. Nevertheless, physicians on the Texas State Board of Pharmacy’s (TSBP’s) PMP Advisory Committee say the inaccuracies pop up in the state’s monitoring program, PMP Aware, often enough to raise concerns. And the implications amount to more than just a temporary setback and annoyance.

San Antonio orthopedic surgeon Adam Bruggeman, MD, a member of the advisory committee, says those mistakes can damage the patient-physician relationship. He’s run into various matching issues in the past few years and approached the Texas State Board of Pharmacy (TSBP) to correct them. The pharmacy board administers PMP Aware.

“Imagine if you are a patient and you were getting ready to have surgery … and you need a pain medicine, and the PMP mistakenly put somebody else’s record in yours because of some mismatches. And now it looks like another Karen Smith, born on the same day as you, her records are in your record,” Dr. Bruggeman said. “Now the surgeon’s saying, ‘I have this chart in front of me, and you’re using medications.’ And the patient says, ‘No I’m not, I’ve never filled. I’ve never taken the medications before.’ And the surgeon’s getting frustrated and saying, ‘Well, you’re lying to me, you’re drug-seeking.’

“You can set up a scenario in which the patient-physician relationship could be really harmed by information that’s gleaned from the prescription monitoring program.”

Although Appriss Health, the vendor that developed the PMP, says those mismatches are statistically rare, the vendor and TSBP agree with physicians that improvements are possible, such as a standardized data entry system and search function.

“A mess”

The Texas law mandating that prescribers check the PMP before writing a scrip for opioids, barbiturates, benzodiazepines, or carisoprodol went into effect on March 1. But the law did nothing to resolve the problems with patient-matching and data consolidation, says Mesquite pain physician C.M. Schade, MD, presiding officer of the PMP Advisory Committee.

“That’s one of the aggravating things about the legislature just thinking that mandatory checking equals goodness. It doesn’t,” he said.

Dispensing pharmacies, not prescribers, create entries in the PMP. Patient mix-ups can start off innocently enough, for instance due to a typo, by physicians searching a patient by a different first name, a father and son with the same name, or someone who changed his or her last name after getting married.

But without standards for searching and entering patient data within the PMP, “it’s a mess. And it’s widespread, and it’s real,” Dr. Schade said.

Allison Benz, TSBP’s executive director, acknowledges the board has encountered problems with the PMP system merging patients with the same last name, similar but slightly different first names, and other similar characteristics, such as hometown.

“We’ve seen that with twins, parents that name their kids [similarly],” she said. “When you have that super-similar name, and they always have the exact same date of birth, same address possibly, or same city, the system may think they are the same person.”

Those mismatches, she said, sometimes lead prescribers to confuse the patient with someone else with a concerning drug history.

Through the state’s Patient Access Program (pharmacy.texas.gov/patient-access), patients can obtain a copy of their PMP record. When errors get brought to TSBP’s attention by a prescriber or patient, the board works with Appriss Health to resolve them.

“If we get a person that is grouped that shouldn’t be, and [the patient is] contacting us, we tell them we can investigate it. And if we can confirm that they’re not [correctly matched], then we can get them [ungrouped],” Ms. Benz said.

Oct_20_TM_Law_PMP_Sidebar

Appriss: Errors are rare

Appriss Health told Texas Medicine that its research shows patient-matching issues in its PMP software are uncommon, both nationally and in Texas. Brad Bauer, Appriss’ senior vice president of new business development, said the company runs 43 PMPs in the U.S. and its territories.

“One instance is too many, obviously,” Mr. Bauer said. But he said with the vendor’s “patient-matching protocols that are in place today, we feel we do a good job, and we continue to enhance what we’re able to, with what we have to work with.”

Krista Whalen, PhD, data science leader for Appriss, says the Texas PMP has 114.6 million prescription records. Since 2019, the system contained just under 11 million instances in which it grouped different patient records in the PMP as being the same person, company data show. Of those, 231 entries required correction, which occurs when TSBP notifies Appriss of the need. That amounts to about two patient groups requiring correction for every 100,000 groupings in Texas, she said.

Dr. Whalen added that when the PMP receives new data on dispensed medications, “we do pretty extensive data validation and error-checking to make sure that anything that we do end up [matching] is quality data. So there’s quite a few extensive checks there, and every once in a while, something does creep through, either through a typo in the data entry process, [or the system] inappropriately linked somebody.”

As for how to move patient-matching errors closer to zero, the pharmacy board’s Ms. Benz says legislative or regulatory action could be the answer.

“We need to somehow get everyone to have a standardized way of entering the data, the names,” whether by law, rule, or simply by educating prescribers and pharmacists, she said. “So if your name is William, and that’s your legal name, the doctor needs to write William, and the pharmacy needs to put it in by William, and that will definitely help.”

Mr. Bauer of Appriss agrees some type of standardization would help, saying it could be beneficial for Texas and other states to collaborate on the process. Appriss says Texas’ PMP is interconnected with 34 states.

Dr. Whalen added, “Usually, it’s the point of the pharmacy that has a lot of the data quality checks. When you pick up the prescription, usually they confirm your phone number, your address, that kind of thing. Those kind of checks at the point of pickup are useful to ensure that the data is good quality and up to date.”

Dr. Bruggeman agrees Appriss is “really working hard” on the patient-matching issues, based on information the vendor has presented at PMP Advisory Committee meetings.

However, he said: “There is some disconnect between the practice of medicine and what they’re doing, and that’s why we’re getting [problems]. At the end of the day, [it’s] just a recognition that it’s not a perfect system. It’s the best system we have, but it’s just not perfect, and our goal should be to continually improve it and get closer to 100% accuracy each year.”

Tex Med. 2020;116(10):38-39
October 2020 Texas Medicine Contents
Texas Medicine  Main Page