Biometrics in healthcare: use cases, benefits, and things to consider

Recording the physical or behavioral characteristics of choice together with the person’s name and/or identification number

Storing the selected elements after converting them into a graph or a line of code

Every time users attempt to log in, the system compares the characteristics they present to its existing records to either grant or deny access

Fingerprints: this is the oldest and the most well-developed method of biometric authentication. It is also the most commonly used type. Fingerprints can be collected using optical or thermal techniques. The resulting print is enhanced and processed with sophisticated algorithms to build a digital biometric template. We all see this identifier in phones and door panels.

Voice recognition: the way our voice sounds depends on the physical structure of our mouth and throat, and our habits. A voiceprint (or spectrogram) is used to compare voices and identify the speaker. It is created by analyzing the sound’s frequency, duration, and amplitude. Telephone-based service portals often use this recognition modality to identify customers. For this system to work, users need to say a predefined phrase or provide an extended speech sample.

Facial recognition: a facial structure is what people usually use to recognize each other. This is also a common biometrics technology suitable for surveillance applications. You can build a simple facial recognition system using a digital camera and specialized software. For more complex setups, you can add an infrared light emitter. Facial recognition applications analyze over 80 points on the human face to decide if it’s a match.

Iris scan: iris-based recognition is performed using an iris scanner, a digital camera, and image processing software. To make this identification more reliable, one can add extra features, such as liveness detection (for example, the user is asked to blink). An iris has 200 reference points that are matched during identification instead of only 70 in fingerprints.

Retinal scan: retinal patterns are located deep inside the eye, a spot which a regular camera can’t reach. This recognition method requires a retinal scanner, which scans retinal blood capillaries with near-infrared light. With all the technical complications surrounding this method, it is rather accurate. retina-based identification is 70 times more accurate than iris-based systems and 20,000 more accurate than fingerprints identification.

DNA matching: forensic investigations extensively use DNA profiling. One can obtain a DNA sample through saliva, blood, buccal smear, etc. Even though this biometric modality is viewed as a reliable form of authentication, Israeli scientists demonstrated back in 2009 that it was possible to fabricate DNA. However, the same team conducted a test that could distinguish between fake and real DNA by analyzing DNA methylation patterns.

Hand geometry: hand geometry readers capture a hand’s image and measure different features, including fingers’ length, width, thickness, and curvature. The performance of this identification system is not affected by skin conditions, but a hand’s shape can change when the subject ages or loses weight. Hand geometry is primarily used in attendance tracking and physical access applications.

Signature recognition: biometric signature verification is not limited to letters’ shape. It also analyzes features such as writing pressure, inclination, and special coordinates. This method is popular in banks and other financial institutions.

Body odor recognition: this is a relatively new recognition modality. Researchers at the Polytechnic University of Madrid started to explore body odor as a unique identifier in 2014. Dietary changes and diseases don’t impact this identification method. This technique is still in its infancy, but it shows promise. Body odor is impossible to replicate, making the technique highly secure. Also, simply passing by a sensor is less intrusive than the eye or fingerprint scanning.

A reliable method of identification that cannot be stolen or forgotten like passwords

Easy and convenient to use, as there is no need to carry a physical object

A cheaper identification method in the long term, as it helps eliminate costly duplicates. Commenting on the promise of health biometrics, RightPatient Cofounder Michael Trader said: “Compared to manual methods of identification that lead to an 18% average duplicate record, $1.5 million annual losses in claim denials, and a significant impact on patient safety, health systems should not see any limitations in implementing biometrics to address these issues”

Suitable for people with special needs. One study has shown that illiterate people are content to use biometric identification as they don’t have to reveal the fact that they can’t write

Medical organizations need to be careful about data security issues. If sensitive information is exposed, the consequences can be severe. If someone’s password got hacked, the person can easily replace it. But if biometrics data got exposed, you cannot change someone’s eyeballs or fingerprints. You will likely need to replace the identification system or exclude the victim

It is often costly to acquire biometrics identification technology

Physical damage can render the body part involved in biometrics unusable

Biometric authentication requires additional software, which must be compatible with the hospital’s EHR platform and other applications involved. Otherwise, technical issues can arise like with Windows Hello, a biometrics-based technology that grants access to apps, devices, etc. It wasn’t widely adopted because it is limited to Windows 10, and when it rolled out, many hospitals were still upgrading their operating systems to the latest version.

Assess the potential impact of the new biometric solution on your available IT infrastructure, including your network capacity and integration with existing systems.

Choose a reliable healthcare biometrics vendor who will be available for after-sales support and help fix any negative impact the system might have on your other applications. Also, it is best if your vendor is already familiar with the healthcare domain and the regulations around it.

Protect your medical biometrics data. Biometrics fall under the category of protected health information (PHI). Medical facilities need to safeguard this data following the Health Insurance Portability and Accountability Act (HIPAA), the Health Information Technology for Economic and Clinical Health (HITECH) Act, and other regulations applicable in their country of operations. Five US states have passed the Biometric Information Privacy Act (BIPA) at the moment of writing this article. Other states are debating it. So, watch the news carefully to ensure compliance.

Train your employees on proper usage of this technology so that they can maintain the system and secure the data.

If you decide to opt for biometrics identification, ensure that every individual is adequately enrolled and has given explicit consent. You might consider enhancing the quality of biometric scans during the enrollment phase to achieve better results in identification. For example, in the case of scanning fingerprints, there is a prescan treatment called “milking the finger”, which helps raise fingerprint ridges.

Biometrics don’t offer the ultimate security. Deloitte cites studies where US-based researchers designed copies of fingerprints that can unlock around a third of existing smartphone models. In another instance, a UK-based organization experimented with hacking facial recognition systems using 3D-printed heads. To strengthen the system, one can use multifactor authentication, such as combining biometrics with PIN codes.

This technology can be biased. For example, facial recognition algorithms show higher error rates, up to 34%, when attempting to recognize dark-skinned females than white males.