By Gregg Church

Poor data quality undermines the efforts of healthcare organizations to provide people with the best care and to get the most value out of their data. Even something as basic as matching a patient to the right medical records can be extremely difficult.

The reason provider and payer organizations struggle with patient matching is clear: Lack of a national patient ID. The American Medical Association and other medical organizations long have advocated for a universal patient identifier (UPI), but Congress in January rejected a bid to lift a ban imposed in 1999 on funding to promote or implement a national patient ID. Without a UPI on which to rely, hospitals, clinics, labs, imaging centers, specialist groups, behavioral health practices and other providers end up creating their own unique medical record number for each patient.
If a patient sees a primary care provider in one city and then relocates, that patient’s new provider group will create its own medical record – which may not have the same data used to identify the patient as the previous provider. The lab where the patient underwent tests also may have a separate medical file for the patient.

These duplicate patient records are fed into the databases of health information exchanges or transmitted from one provider to another. As a result, any authorized entity accessing a patient’s medical records may not get the full picture because they are unaware there are duplicate records with different information on that patient.

When patient data is siloed, it can lead to clinicians making point-of-care decisions based on incomplete or inaccurate information. And that both endangers patients and increases the costs of care. A 2021 Black Book research survey revealed that patient ID errors cost nearly $2,000 per inpatient stay and more than $1,700 per emergency department visit. On an annual basis, patient misidentification costs the average hospital $2.5 million and the U.S. healthcare system $6.7 billion, Black Book writes.

If your healthcare organization doesn’t have a master patient index (MPI) looking at multiple incoming data feeds containing different medical record numbers for a patient to verify, it’s virtually impossible to successfully match patients to their full and accurate records. Even within a single hospital or medical center, the facility’s lab or radiology center may have different identifiers for a patient.

To identify and address problematic data sources and flag invalid patient identification attributes, it’s critical that healthcare organizations have the ability to quickly assess what patient information is valid in the data sets streaming into and throughout their electronic health record (EHR) systems. One way to do this is to use probabilistic algorithms that assign weighted value to identifying attributes.

Health IT Homepage