HIT Think

Why EMPI is crucial for current patient identification needs

The depth and breadth of patient identification errors that currently exist across the U.S. healthcare system demand that stakeholders develop more comprehensive approaches to accurately linking individuals to their health data.

Because the safety of the patient starts with consistently identifying and matching them to their records, there lies greater pressure on the industry to improve EHR interoperability and to strengthen Enterprise Master Patient Index (EMPI) systems. This is critically important as healthcare organizations become more dependent on initiatives such as ACOs, population health, precision medicine and health information exchanges (HIEs), all of which rely on accurate and easily accessible patient data.

Despite newly directed federal efforts to encourage information exchange and patient data empowerment, the industry still struggles to recognize the foundational role of patient identification in achieving the interoperability goals of the administration.

Fragmented patient records are most often the result of multiple name variations, data entry errors and a lack of data standardization processes. Typos or omissions in key demographic attributes, such as the patient’s birthdate, address or phone number, make the matching process very complex. Additionally, patients move, marry, divorce and visit multiple providers in their community—thus, new records are created and the potential for misidentification grows.


To appreciate the full extent and severity of the problem, consider the following statistics. A 2016 Ponemon Institute survey found that 86 percent of medical errors occur as a result of patient misidentification. Repeated medical care because of duplicate records cost an average of $1,950 per patient per inpatient stay and more than $800 per emergency department visit, according to a recent Black Book survey. Further, the report estimates denied claims as a result of inaccurate patient identification cost hospitals $1.5 million annually.

As healthcare technology advances and new innovations take hold, the need to accurately identify patients will remain an urgent priority. Making improvements in the matching accuracy will require a combination of tactics. While concepts such as referential matching have been touted by some as a game-changing approach to solving patient matching errors, there’s ample evidence to suggest that it will take more than one tool or approach to fully manage the patient identification problem. For this reason, there is no doubt that reports of the demise of the EMPI are premature and they are not an accurate depiction of the current state of patient matching.

That is not to say that technologies such referential matching of data don’t have value. Many healthcare organizations have looked to referential data from public records, including credit or financial information, that provide demographic information like current and past home addresses, nicknames and phone numbers. A combination of demographic information with patient identifiers in EHRs and EMPI systems are used to match patients to their records, and for certain cases this process provides an opportunity to verify data, provide context, improve accuracy and reduce the number of duplicate and unmatched records.

But there are significant limitations and complexities that need to be considered with such an approach. For example, identifying and linking the records of patients who are not of adult age is difficult to achieve with referential matching, which relies on demographic data from public records that most often does not exist for minors. This includes income and property taxes, utility bills, licenses, loans, voter registrations, court and criminal records, and more. Nearly one out of every six discharges from U.S. hospitals in 2012 was for children age 17 years and younger, the majority of whom were infants and newborns. Even if the data existed, searching a child’s personal information via public records raises privacy, ethical and legal questions.

Like other tools that were promoted as an answer to solving patient identification errors (such as EHRs), referential matching in healthcare can only be as good as the information the system collects and the accuracy of the data it is compared against.

While critically important, demographic information is often incomplete, inaccurate, inadequate or irretrievable, and approaches for collecting the data is unstandardized. For referential matching to work effectively, the records being matched must have enough information to match the local data to the reference data. For this reason, sparsely populated records, or records with too many data entry errors, will not result in a confident match.

NextGate recently examined a large sampling of ambiguous matches at a health information network on the West Coast and discovered that when historical address or phone information was present, the adult patients in the set could be matched with the aid of reference data in more than 60 percent of the cases. While this is a promising statistic, it highlights the fact that reference data is just one part of a set of strategies that need to be employed to improve match rates.

Because of the importance of address information when using reference data, one additional approach to maximizing the benefit of reference data is to clean up the address data prior to performing the query. This can be done with the aid of address cleansing tools that leverage a complete database of all the known delivery points known by the United States Postal Service.

While such tools do not associate a specific patient to an address, they do ensure the address is consistently represented and this improves the likelihood of a match to the reference data. The use of such a tool has the added benefit of geocoding the address data which enables location-based searches of the patient population. It also maximizes the likelihood of a successful communication with the patient when the delivery of documents is involved.

Currently, a plethora of other technologies are being tested to improve patient matching accuracy and the management of demographic data. For example, machine learning can better leverage the existing data by fine tuning the probabilistic matching process more closely than ever before. Healthcare organizations are also finding that the use of biometrics can help complement their patient identification efforts. Leveraging an EMPI along with biometric applications can empower organizations to drive data security into patient matching and verification systems across its IT infrastructure.

While a powerful tool in increasing patient identification accuracy, biometric solutions also come with many large-scale deployment and technological demands. Storing images, palm scans, and fingerprints of millions of patients requires significant infrastructure that needs to be scalable. User anxiety associated with invasive sensors are also still common, since palm, fingerprint and iris scanners require patients to touch or interact with the hardware.

Another emerging technology poised to impact patient matching is the use of blockchain. While data governance questions still act as a barrier toward its adoption, blockchain may provide the technical underpinning to allow healthcare organizations to securely store patient records in a distributed manner while maintaining a single auditable version of the truth.

Aside from leveraging advances in technology, healthcare organizations are engaging stricter information governance processes and policies for data stewardship, from standardized registration processes to defined roles and responsibilities within the organization.

With the appropriate data governance controls in place, patients can also be actively involved in managing and updating their health records. For example, personal data can be controlled and maintained by the patient using their own smartphone device. This method of self-administration would help support patient matching efforts at key stages where data errors often occur; during enrollment and at registration.

As healthcare IT executives struggle to design a system that reduces duplicate patient records and patient identification errors, they’ll find there is no silver bullet or one-size-fits-all solution. Healthcare organizations are best served by enhancing their core EMPI and EHR systems and implementing a multi-layered strategy across the network that will seamlessly match individuals to their medical records while giving providers timely, accurate access to their patient’s health.

For reprint and licensing requests for this article, click here.