How Adding Biometric Data to the EHR Can Drive Patient Matching

Adding biometric data to the EHR can enhance patient matching, but it presents unique data security challenges, according to a Pew Charitable Trusts report

Patient health data is often spread across several EHRs from doctors' offices, hospitals, and health systems. Healthcare providers must match those files to get a complete picture of their patients' health history, but errors in matching records to the correct patient are common.

According to a 2012 survey, nearly one in five hospital CIOs indicated patient harm in the previous year due to record mismatches. These kinds of errors cost the healthcare system about $6 billion annually.

RTI International, a nonprofit research institute, and The Pew Charitable Trusts conducted 12 interviews between April and July 2020 to identify how biometrics could enhance patient matching.

The researchers then held work group discussions with 29 experts from health systems, insurers, biometric and digital identity technology developers, health information exchange (HIE) platforms, EHR vendors, patient and privacy advocates, health policy advisers, standards organizations, and ONC.

The workgroup participants agreed that facial imaging is an optimal type of biometrics because it is relatively inexpensive and contactless. However, the technology raises data privacy concerns.

Storing encrypted biometrics on patients' devices is preferable to keeping them in a single national repository or across health systems' databases.

"Additional protections may be needed to explicitly safeguard biometrics stored on patients' personal devices and to ensure that patients can provide informed consent before participating in biometrics-enhanced patient-matching programs," the report authors wrote.

Additionally, using biometrics for patient matching could present challenges for health equity.

"Small hospitals, rural health care clinics, stand-alone practices, and other providers may not have the resources to buy biometric technology and hire staff to operate it, or they may be situated in a region with poor internet connectivity," the report authors explained.

Some biometric technologies require patients to own smartphones, and people earning lower incomes are more likely to experience illness and less likely to have smartphones.

Further, the authors added that there will always be populations for whom a given type of biometrics does not work.

For example, fingerprinting does not work well for older adults or people with certain skin conditions like eczema. Facial imaging also tends to be less accurate for people with dark complexions due partly to algorithm development bias.

Workgroup participants also noted that the industry must create standards to support the use of biometric data nationwide for patient matching.

"There are hundreds of government-certified EHR products but no national technical standards for capturing biometric images, encrypting biometric data in storage or transmission, or comparing and matching biometrics—all of which are necessary to facilitate the exchange of biometric information between different systems, known as interoperability," the report authors wrote.