PRESCRIPTION MONITORING PROGRAMS (PMPs), also known as prescription drug monitoring programs (PDMPs), continue to be important sources of information in the fight against opioid abuse and overdose deaths. While much of the harm is now attributed to illegal drugs instead of prescription drugs, policymakers continue to seek additional ways to use PMP data and increase the availability of that data to prescribers and dispensers of prescription drugs. The recently released 2019 National Drug Control Strategy report describes PMPs as proven means to increase accountability in opioid prescribing and prevent adverse interactions with opioids from multiple providers. Therefore, it is increasingly important to understand the data that goes into a patient’s PMP report and is sent to prescribers and dispensers. In my last article, I identified the patient, the prescriber, and the dispenser as critical data fields reported to the PMP that can lead to errors or discrepancies in the reports you request.
I return to the premise that all data in the pharmacy system is based on a hand-written prescription, a telephone conversation, or a facsimile document that is transcribed, or is transmitted electronically from a prescriber. The data from all of these sources can be amended as needed during prescription processing to create a legally complete record.
DATE OF BIRTH
While date of birth seems fairly straightforward and unchanging, data entry can result in transposed digits or “fat-finger” errors. Fat fingers is jokingly used to describe the instance where a finger accidentally strikes an adjacent key either on the keyboard or on a numeric keypad. It can occur at either the prescriber or dispenser level. Such errors are fairly easy to recognize and usually easy to correct.
The most difficult errors in the date of birth field arise when the patient’s third-party payer has recorded an incorrect date of birth. I’ve seen these errors, and there seems to be no explanation or pattern for how they occur. Moreover, these errors are difficult to correct because the pharmacy has to use the payer’s data in order to be reimbursed. These errors then become part of the pharmacy’s permanent record.
PATIENT ADDRESS AND TELEPHONE NUMBER
The issues with patient address are obvious — people move. While this cannot be programmed, there are some things that pharmacists can do to minimize the number of address issues.
First, confirm the patient’s current address at each visit. While this seems obvious, based on my experience it doesn’t always occur when prescriptions are picked up.
The other action that is beneficial is to be aware of default settings in the pharmacy system. I’ve seen too many instances of the same patient’s address being attributed to multiple locations. For example, a patient living in a tri-state area might have the following addresses on his or her PMP report:
123 Main Street, Some Town, Pa.
123 Main Street, Any Town, Ohio
123 Main Street, Another Town, W.V.
Further investigation often reveals that a pharmacy system has a default setting in the patient’s address field that auto-populates the city, state, and ZIP to be the same as the pharmacy’s city, state, and ZIP. A better practice would be to remove such defaults or train pharmacy staff to ensure that the city, state, and ZIP information is accurate for each patient.
Patient phone number errors have similar characteristics. Today, many people do not change their phone number, especially cell phone numbers, so that helps. However, the pharmacy system may insert a default area code. Since some patient data-matching algorithms rely heavily on phone number to distinguish among multiple James Smiths, having a correct phone number, including area code, is vital.
Back in the old days, most patients got their new prescriptions filled on the same day that the prescriptions were written. In addition, the date written was usually of little concern, as the date of processing was the primary concern for most purposes. Many pharmacy systems were designed to insert today’s date by default into a prescription record. If not changed at the time of processing, the pharmacy record and the resulting PMP data will show the date written to be the same as the date of processing. While often correct, an error is created if these dates are actually different.
The date written is being utilized today to determine prescriber compliance with state prescribing laws. For example, when a state has mandated that prescribers query the PMP prior to prescribing an opioid, the prescriber’s regulatory authority may request data from the PMP to determine whether their licensees requested a report for every opioid prescription written. The date writfeature: PMP Update
ten should match the prescriber’s patient record. In the long run, it takes less time to enter the date written as the prescription is processed than to search later for prescriptions that match up to prescriber records in response to the agency’s request or subpoena.
And then there’s the problem of prescribers who don’t date the prescription. While this used to be a common occurrence with hand-written prescriptions, it is much less frequent with electronic or computer-printed prescriptions.
DATE SOLD VERSUS DATE PROCESSED
Another change from the old days: Pharmacies used to fill prescriptions and deliver (relinquish possession) of the drugs on the same day. Today, the pharmacy landscape is very different. E-prescribing, central-fill, auto-refill, and delivery processes frequently result in a processing date that is different from the date that the patient takes possession of the drugs (date sold).
Since the purpose for creating a PMP is to provide information related to a person’s possession of the drugs, the date of actual receipt by the patient/patient’s agent is important. Some pharmacy systems can separate the date that a pharmacy processes the prescription from the date that the patient takes possession, but apparently many cannot. If the pharmacy has a point-of-sale system and can report to the state PMP both dates or only when a patient picks up the drugs, that is ideal. If the pharmacy cannot separate these instances, it is critical that the pharmacy at least correct the PMP records by submitting a deletion to the PMP whenever a prescription is not picked up or is transferred to another pharmacy after reporting that it was processed. The latter could happen when the pharmacy processes the prescription, and then two days later, the patient decides he or she wants it transferred to another pharmacy. In the meantime, the pharmacy may have reported this prescription to the PMP, especially in states that require reporting within 24 hours. It would be a best practice to amend the PMP record whenever the pickup date is different from the processing date, but this would require drastic changes to pharmacy workflow.
Fortunately, date-written and date-sold errors are relatively minor in most cases. But they contribute significantly to the overall perceptions that PMP data is rife with errors and that changes in data collection are warranted.
Errors related to the drug that leaves the pharmacy are relatively rare because most PMPs rely on the NDC number to identify the drug, strength, and form. Problems can arise, particularly with an electronic prescription where the drug form field has been truncated in the software. According to pharmacists, prescribers may select the drug and strength correctly, but the form of the drug may be hidden or abbreviated. If the error is not detected until the prescription is delivered to the patient (a few days later), the PMP report may include the NDC of the drug prescribed, not the NDC of the one finally processed and sold.
The other drug-type problem occurs when the drug has been repackaged. The PMP software systems use national NDC databases to convert the NDC number to a drug name, strength, and form for the patient’s PMP report. If the repackager applies a new NDC number to the labeling (as required by the FDA), the new NDC number may not be recognized. The PMP may accept the data and show a “drug unknown”’ message on the patient report, or the PMP may reject the record.
One of the most perplexing debates in PMP circles today is about veterinary prescriptions. Historically, many PMPs did not address veterinarians, since these prescriptions would likely be reported by the pharmacy. Other PMPs recognized that veterinarians can be and are scammed by pet owners to obtain drugs for personal consumption, and so required reporting of controlled substances dispensed from stock at the veterinary office, hospital, or clinic. One of my favorite law enforcement officers has related his first-hand experience with a dog that was taught to cough on demand, resulting in a prescription for hydrocodone. Obviously, the problem was the animal’s owner and not the animal. There are a few issues that arise. The debate is over these issues.
❏ What name should be on the prescription for the animal? Many states’ pharmacy laws mandate a format, but I’ve never seen it enforced. Often the pharmacy owner will decide, or the processing pharmacist will develop a format that works for him or her.
❏ How does the pharmacy identify the prescribing veterinarian if he or she has no DEA number? Also, as I understand it, vets cannot get an NPI (National Provider Identifier) number. Some states now require the PMP to have a database of state veterinary license numbers that is added to their PMP software and is continually updated. This works for in-state queries but may not work for out-of-state queries.
❏ Can the veterinarian query the PMP on the animal’s owner? The owner is not a patient of the vet, so PMP laws may not allow it. HIPAA allows providers to access a patient record, but the vet is not treating the owner, so that is problematic.
❏ If a state’s PMP law allows or requires the vet to consult the PMP before prescribing for an animal, the PMP must be able to separate prescriptions for the animal from prescriptions for the owner. If the PMP report is on the patient, how are veterinary prescription identified? The ASAP (American Society for Automation in Pharmacy) standard allows reporting whether the prescription is for a human or an animal, and if an animal, then the name of the animal can be reported as well. The pharmacy systems must make provision for the capture of this information at the time of processing the prescription. In addition, the PMPs must have a way to include prescriptions for animals on the owner’s reports.
❏ What about the controlled substances or drugs of concern that are dispensed by the vet? If the state requires that these be reported to the PMP, a vet’s computer system could be programmed with the ASAP reporting standard, just like pharmacy systems. Then the controlled substances file can be submitted to a PMP electronically. Depending on cost and volume, vets always have the alternative of writing prescriptions (just for these PMP drugs) to be filled at a pharmacy instead of dispensed at the clinic.
TO SUM UP
All of these discussion topics reinforce the fact that a PMP is not a panacea for the opioid problem, nor is it a repository of completely factual data. It’s a tool for gathering additional information that must be combined with nonprescription information in order for a healthcare professional to assess a patient’s situation. It’s not easy. The best thing that pharmacists can do is strive for accuracy in the prescription records that are reported to their PMP and share their knowledge about PMPs with prescribers and policymakers. CT
Danna E. Droz, J.D., R.Ph., is the prescription monitoring program liaison for the National Association of Boards of Pharmacy in Mount Prospect, Ill. She can be reached at firstname.lastname@example.org.