A Unique Patient Identifier Decreases Adverse Drug Interactions

Developing an enhanced patient identification strategy, such as a National Patient Identifier (NPI), across separate data sources can boost patient safety and limit adverse drug-to-drug interactions, according to a study published in the Journal of Medical Internet Research.

To further increase patient safety, providers should enhance patient data exchange (e.g. prescription history) and interoperability.

Patient matching issues have surrounded the healthcare industry for decades. Patient matching, or the ability to connect the correct medical data to the appropriate patient, is a key patient safety issue complicated by limited interoperability and data governance.

Patient matching aims to connect patient records across different medical providers or facilities. The same patient visiting two separate doctors or two different hospitals should always have the correct medical record brought up. However, this does not always happen, and a mistake can be lethal to the patient and carry a heavy financial burden for the health system.

One potential fix that has been particularly controversial, yet brought up on occasion, is the NPI. In short, the system would assign each American citizen a unique healthcare number. This unique number would eradicate the chances of overlapping a patient health record when an individual shares the same name or birthdate.  

Between 2016 and 2019, researchers aimed to analyze the frequency of harmful drug-to-drug interaction alerts and estimate the number of drug to drug interactions at retail pharmacies that use a unique patient identifier. Researchers said this is likely due to a lack of patient data exchange.

Researchers compared override, abandonment, and replacement alerts to the prescription type to determine whether the unique patient identifier recognized the crossover alert. Using previous data, the research team hypothesized a high rate of pharmacy benefit managers missing the drug-to-drug interaction alerts.

Researchers analyzed nearly 50 million patients and 1.5 billion total claims during this three-year period.

From those claims, researchers identified 242,646 serious drug-to-drug interaction alerts (16.5 percent) and 2,388 stemmed from crossover alerts. Additionally, the serious drug-to-drug alerts would not have detected roughly 1 percent without a unique patient identifier.  

Overall, researchers estimated the insured population has roughly 458,000 yearly serious drug-to-drug alerts.

Following the study, researchers estimated that roughly 10 percent of serious drug-to-drug alerts were left undetected by pharmacy benefit managers due to a lack of a unique patient identifier. As a result, up to 6,000 patients across the country are receiving harmful medication.

However, researchers acknowledged potential limitations.

One is that the researchers do not directly link the unique patient identifier to increased health outcomes as the serious drug-to-drug alerts may have also been detected by the pharmacy, with or without the patient identifier. While it does enhance the chances of identifying the drug interaction alerts, it may not be the sole reason.

If improperly implemented, a unique patient identifier could misidentify a patient or result in a false positive.

“Although progress is being made in US health care systems toward more comprehensive interoperability, fragmented information silos remain the status quo,” wrote the study authors. “When patients transition to a new insurer or pharmacy benefit manager, their identity and historical prescription data do not seamlessly follow. Subsequently, pharmacy benefit managers may lack both identifying information and historical data.”

Even though privacy and security can improve to implement an NPI in the future, researchers said the study data proves consistent identification can help identify other drug-to-drug interactions.

“Given the volume of opportunities to improve patient care, the health care system should choose the most accurate identification strategy possible,” concluded the researchers. “We hope that others will conduct similar studies in other areas of the health care ecosystem to forecast benefits from patient identification and patient-record sharing.”