How four promising trends could aid patient matching

The issue of inadequate patient matching and duplicate records has grown increasingly complex as more data is generated and more applications are introduced into the healthcare environment.

As the formation of clinically integrated networks and consolidation play a greater role in the business strategy toward accountable care, healthcare organizations must ensure records are correctly tied to the right individual for enhanced quality and safety, reliable reporting and analytics, administrative efficiencies and financial improvements.

Poor patient identification not only places patient safety and an organization’s bottom line at risk but serves as sizeable barrier to interoperability. According to a new report by Pew Charitable Trusts and Massachusetts eHealth Collaborative (MAeHC), today’s match rates are far below the desired level for effective interoperability and data exchange. The report, which details providers’ perspectives on the state of patient matching, also found that match rates are difficult to measure and that emphasis on care coordination places greater demand on positive patient identification

While providers often struggle to understand the magnitude of their internal duplicates, those surveyed did acknowledge challenges externally. Interviewees cited current match rates as low as 50 or 60 percent when exchanging data outside the organization, and that ideally, accuracy rates of 99 percent or higher would be needed for interoperability.

Today’s complex IT environment demands that healthcare organizations engage in more comprehensive patient matching approaches. As data sharing matures and the industry pivots toward value, an enterprise view of patient information is essential for informed clinical-decision making, better outcomes and a seamless patient-provider experience during every encounter.

While use of an enterprise master patient index (EMPI) remains the most robust tool for centralizing patient identity and facilitating fluid data exchange, there are some promising new trends that are helping to extend one’s EMPI and get HCOs closer to 100 percent matching accuracy.

Location intelligence
Location intelligence is a powerful new trend in patient matching helping healthcare organizations proactively manage, detect and eliminate data quality issues. Location intelligence including address verification and geocoding allow organizations to standardize and authenticate address information in real-time using data from the U.S Postal Service to avoid duplicate record creation and identity fraud.

In fact, a study published in the May issue of the Journal of the American Medical Informatics Association (JAMIA) found that standardizing patient addresses using the USPS format in EHRs improved match rates by as much as 3 percent. When combined with other demographics for patient matching, it ensures the address information is consistently formatted to avoid data errors at the point of capture. The use of such a tool has the added benefit of geocoding the address data which enables location-based searches of the patient population and maximizes the likelihood of successful communications via conventional mail delivery.

Referential matching
While use of sophisticated patient matching algorithms will continue to be the safest, most effective approach for automated record matching, there will always remain a small subset of records where an external source of information is needed. This typically applies in cases where the records being compared represent widely separated snapshots in time. In such cases, it may be possible to relate these records together by comparing them to a reference data provider that can correlate older demographic elements like address and phone information.

While by no means a substitute for an EMPI, in conjunction, third-party data can be an added tool to help organizations associate demographic data changes. A sampling by NextGate involving 13 million actual record pairs found reference data to be helpful in resolving approximately 35 to 40 percent of the 3 percent of duplicate record pairs not matched by the EMPI.

However, limitations with referential matching do exist. For instance, identifying and linking health records belonging to minors is difficult to achieve with referential matching, which relies on demographic information from public records that do not exist for those under the age of 18. This includes income and property taxes, utility bills, licenses, loans, voter registrations and court and criminal records.

According to Pew and MAeHC survey, providers found referential matching was seen as helpful, but expressed uncertainties regarding the reliability and accuracy of the data. Additionally, interviewees raised concerns about the potential challenges of entering into business agreements to enable a referential data approach.

Machine learning
While technologies like biometrics and blockchain have made inroads, the next wave of change in identity management will be driven by machine learning. At the heart of the technology is the ability to learn, perform and recognize patterns, thereby fine tuning the probabilistic matching process. As part of an augmented patient matching strategy, machine learning can be leveraged to detect common manual data remediation sequences and behaviors.

This rule-based intelligence, which continually improves over time, can not only further record matching accuracy but alleviate considerable data reconciliation burdens for HIM departments charged with piecing together fragmented, disconnected sources.

The challenge in using this kind of information is in the sheer number of human interactions required for an algorithm of this type to truly outperform human remediation. This is because the system must be able to detect broad patterns where users consistently take an action of marking a pair of records unique or as a match. Training, however, is greatly simplified in a cloud environment where usage statistics across many implementations can be gathered to produce a highly intelligent record resolution algorithm.

Consumer smartphones
As federal officials at ONC and CMS continue to push for patient data access and ownership, the opportunity for individuals to play a role in monitoring their personal health data is substantial. With the appropriate data governance controls in place, patients can be effective in managing and updating their own health record as they happen, such as moving to a new address or changing a phone number.

This self-administration, in which personal data is controlled and maintained by the patient using their smartphone device, can support patient matching efforts at key stages where errors often occur—during enrollment and at the point of registration.

When asked about consumer smartphones as an additional opportunity to improve record accuracy, the Pew and MAeHC survey found providers generally in favor. Potential drawbacks included rural access to reliable broadband and the security and integrity of the patient data into the provider’s IT systems, in leu of hackers. However, as smartwatches and other wearable devices with the same capability continue to become more of a commodity, individuals will likely expect these devices to be involved as part of the care process.

While no one solution exists to solving the patient identity crisis, an EMPI approach—backed by one or more of these technologies—can be a powerful advantage in furthering match rates across various settings.

Ultimately, in today’s transformative digital healthcare landscape, HCOs are going to have to look beyond their EHRs to manage patient identity. To effectively capture, record and share accurate demographic data elements across systems, HCOs will be best served by integrating a vendor-agnostic EMPI platform, leveraging one or more of these trends, and implementing a sound data governance strategy across the enterprise.