How these technologies can solve common data problems.
Accurate patient matching is the foundation of almost every critical strategic initiative within today’s health systems—from patient engagement and patient experience initiatives to clinical and quality analytics, from population health analytics to interoperability and improved revenue cycle management. In fact, every aspect of a health system’s operations and care management requires an accurate and complete view of every patient.
But gaining this accurate and complete view has become nearly impossible for electronic health record (EHR) and enterprise master patient index (EMPI) technologies. This is because the sheer amount of patient data stored by healthcare organizations is exploding and is entering organizations from new sources and through new channels—such as newly acquired hospitals, newly consolidated EHRs, patient portals, and personal health record apps. In short, there are much more data, much more variance in the data, and much lower quality data, creating a huge challenge for EHRs and EMPIs, which must link all this data to the correct patients.
This challenge is evidenced by rising duplicate rates. According to a recent Black Book Research survey, the average duplicate rate within healthcare organizations is 18%, which is up significantly from the average of 8-12% measured by the American Health Information Management Association just 10 years ago.
And these duplicates have real, substantial, and growing economic costs to health systems. According to the same Black Book survey, each duplicate record costs healthcare organizations more than $800 per emergency department visit and more than $1,950 per inpatient stay due to redundant medical tests and procedures. In addition to these costs, the survey found that 33% of all denied claims were a result of poor patient matching, costing the average hospital $1.5 million annually and the healthcare industry $2 billion per year. A different study conducted by the Ponemon Institute is not as optimistic, asserting that patient misidentification costs the average hospital more than $17.4 million annually in denied claims.
Clearly, EHR and EMPI patient matching technologies are coming up short.
The fundamental problem with these patient matching technologies is that they use algorithms that are only as accurate as the underlying patient demographic data they are matching—and patient demographic data are notoriously error prone, frequently incomplete, and constantly changing.
Patients move and change addresses. They change their names after marriages and divorces. Some patients prefer to use nicknames or a middle name instead of their given first name. Some patients have similar names to their parents, children or spouses. And all demographic data are subject to errors and typos, or can be left incomplete on a form, or might be entered into a registration system with a default value (such as “01/01/1900” for a birthday).
But no algorithm—no matter how sophisticated—can definitively determine that 2 health records belong to the same patient when those records have very different demographic data.
Referential matching is a new, powerful, and fundamentally different approach to patient matching. Rather than using algorithms to directly compare the demographic data from 2 records, referential matching technologies compare the demographic data from those records to a comprehensive and continuously updated reference database of identities. These reference databases contain identities spanning the entire US population, and each identity contains a complete profile of demographic data for a person—including nicknames, aliases, maiden names, common typos, and old addresses.
The reference database is essentially an answer key for demographic data, enabling referential matching technologies to make matches that EHR and EMPI technologies could never make. For example, even if a patient record has a person’s maiden name, old address, and phone number, and another record has that person’s married name, current address, and birthdate, both records will match to the same identity in the reference database.
Organizations can simply plug in cloud-based referential matching services to their EHR or EMPI technologies to instantly improve those technologies’ matching by automatically resolving the toughest matches that would otherwise have to be manually resolved.
This enables organizations to reduce duplicates, reduce clinical costs, reduce the costs of data stewardship processes, improve care and patient safety, improve revenue cycle, and finally get the most out of their EHR or EMPI investments.
The healthcare industry needs a fundamentally different approach to patient matching. We have relied on the same patient matching technologies for decades, yet the landscape of patient data has dramatically changed. As a result, duplicate rates are rising, and the costs of duplicates are skyrocketing. If we do not change our approach to patient matching, we will be left with half-complete patient records, half-correct analytics, half-efficient information exchange, half-quality care, and half-satisfied patients. Organizations should look to referential matching technologies as a much-needed and much-welcomed solution to their patient matching challenges and should consider immediately harnessing these technologies to instantly improve the matching of their EHRs and EMPIs.
Mark LaRow is CEO of Verato
Get the best insights in healthcare analytics directly to your inbox.