Matching data from multiple disparate information sources to the same unique individual has been and continues to be a challenge for achieving high-value analytics based on high-quality healthcare data. This patient matching process and its associated database is referred to as an enterprise master person index (EMPI) system. EMPIs are utilized across the healthcare arena by hospitals, health plans, and accountable care organizations (ACOs) with varying degrees of accuracy and automation. As healthcare payment models continue to shift away from fee-for-service and towards the more data-reliant model of value-based care (VBC), EMPIs are crucial to ensuring patient data is up to snuff and not a roadblock to proper contracting and performance tracking.
The Value of EMPIs
Patient identification and matching issues are a source of data integrity problems within electronic health record (EHR) systems. Insufficient patient matching impedes healthcare quality improvement by complicating information exchange, care coordination and, in some unfortunate cases, can lead to patient mortality as a direct result of medical errors stemming from incorrect EHR data.
In the United States, about a third of all deaths are due to medical errors. That’s over 400,000 patients a year and more than 1,000 each day. Despite best practices in patient access and medical record management, duplicate records continue to be a significant problem in the delivery of high-quality healthcare services.
Hospitals with an EMPI on average see a reduction in patient mismatches, and as a result, fare far better than those without an EMPI. Those who forgo EMPI have match rates of only 24% when exchanging externally. Organizations without an EMPI also reported an 18% duplication rate, which on average costs $1,950 per patient per inpatient stay and over $800 per emergency department visit. Hospitals with an EMPI reported an average 85% match rate when exchanging internally, and organizations overall reported consistently correct patient identification and matching in 93% of registrations.
When many provider organizations successfully implement EMPI systems internally, sharing patient health information between organizations becomes more effective as well. This reduces patient matching efforts for health information exchange (HIE) organizations, who almost universally implement strong EMPI systems.
Challenges of EMPI
Twins present a unique challenge and are found within all data sources. Twins have the same last name, date of birth, and address, but may also share the same gender. Additionally, parents tend to give twins similar names (e.g., Chris and Crystal, or, Taylor and Tyler), which can make patient matching a more arduous task. While our proprietary data matching engine has competence in distinguishing individual patients in twin cases, we also include additional steps in the algorithm between the matching step and the final EMPI ID assignments in order to ensure that the two babies are given different EMPI IDs.
All data sources have their fair share of missing data, such as missing addresses, Medicaid IDs, and, in some cases, even a name. What happens when the baby isn’t named yet? All data sources use placeholders like “BABY,” “BABY BOY,” “BABY BOY OF JANE” in the first name field. Dummy birthdates are also commonly used to satisfy “mandatory” EHR system fields so the data entry of registration records can be completed.
Amanda Hood, PhD, our Principal Data Scientist, and EMPI subject matter expert, observed one case where a medical claim record with mixed information of both mom and baby were included, and where the baby’s medical record number (MRN) was used but the rest of the demographic information was that of the mother. In this particular case, the coded diagnosis was for a newborn condition and the medical information used for billing was for the baby despite the fact that the mother’s name, gender, and date of birth were listed. No doubt such a claim would be denied by the payer, causing a need for the provider’s medical record to be reworked, as well as delaying payments.
Why EMPIs Are So Important
The Health Insurance Portability and Accountability Act (HIPAA) of 1996 had initially set up a remedy to the universal ID problem, but patient privacy concerns in the following years delayed any solution on this front. While existing unique identifiers, such as a Social Security Number (SSN), can account for a large chunk of the healthcare patient population, there are many who receive treatment that will remain unaccounted for (visiting non-residents, undocumented immigrants, etc.). An EMPI solution is essential to ensuring high-quality healthcare in lieu of a universal patient ID to uniquely identify individuals across the entire US healthcare ecosystem. With value-based care models on an upward trajectory, having proper access to the most accurate patient data is a must for all parties involved in such risk-based contracting and EMPIs are a necessary part of ensuring that level of accuracy.