Critical Considerations for Voice-based Applications


Telling Alexa that you’re 5’6” and have blue eyes. That’s easy. But diagnosing and treating patients requires access to much more in-depth information.

Voice recognition technology in healthcare has the potential to make care more accessible to consumers and help clinicians save valuable time in their practice; from simple surveys, to adherence and compliance, to new forms of diagnosis and therapy.

Amazon has recently made massive strides in making this potential attainable with an update that makes their leading voice assistant Alexa compliant to the U.S. Health Insurance Portability and Accountability Act of 1996 (HIPAA). It’s only a matter of time before the “invite-only” arrangement with early developers will scale to a myriad of healthcare solution providers creating innovative new ways to access care and support. From there, it’s only a matter of time before we see voice-first or even voice-only healthcare applications.

Voice-based Apps Have Some Problems

Nonetheless, today most of the compliant services on offer use Alexa for transmitting requests back and forth, without much in the way of added intelligence or processing by Amazon. That is left to the solution providers. A lot of what Amazon can and can’t do is buried deep inside agreements between Amazon and the healthcare service or application provider. This is why innovators at payer, provider and pharma companies need to keep the following critical factors in mind when it comes to voice recognition technology in healthcare:

Embarrassment. Health is personal—and people can be embarrassed by certain health issues, for good reason. Asking Alexa for the name of the generic equivalent of Tylenol isn’t embarrassing. Recording the color and frequency of bowel movements and having Alexa replaying that information when other family members or friends are present is different. Alexa is social, not socially aware. Healthcare applications need information to provide personalized advice, suggest actions, or look up valuable information. However, the design of those applications needs to be done with extra care.

Privacy. Health information is private and privacy concerns are ever-present. With over 1 billion voice devices sold, consumers worry about who is listening to their interactions and what’s being done with the information they provide, especially their health information. These devices tend to wait on a “wake word” to start listening, however, there are cases of unintended activation, leading to an invasion of privacy. Google and Amazon have already “messed up” royally with voice technology privacy.

Control. Many patients, especially those participating in clinical trials, are quite open with sharing their information if the service or application provides utility and efficacy – impacting their health in more personal ways than before. However, managing consent and ensuring patients know how much control they have over their interaction is important. How many of us knew that Amazon has thousands of people – real humans – listening to our Alexa interactions to train the algorithm?

Security: Voice, and by extension voice related health data, should be treated similarly to any other personal information that needs to be protected. In some instances, a user’s voice can be the equivalent of a biometric password, but it is far from mainstream technology. If user information isn’t stored and transmitted securely from these voice-enabled devices, security fallouts can, and have, happened.

Managing Some Pitfalls

There are multiple strategies at our disposal to make use of a technology that is the fastest growing search type, according to research firm Gartner, and fosters accessibility for a segment of the population that has limited mobility.

Start with applications that are not dependent on Personal Health Information (PHI). Many valuable solutions have already been implemented in hospitals and care facilities and at home that do not depend on PHI use. These applications have provided insights into usage and behavioral patterns – paving the way to much better-designed interactions and experiences.

Utilize existing technology & assets. Voice, on its own, isn’t going to move the needle on applications that drive outcomes. That’s why you need to leverage existing technology assets when developing voice recognition technology in healthcare. Northwell Health has its Alexa skill running off of its content management system taking advantage of the same APIs that the website and digital signs use, according to HealthcareITNews. By doing this they are able to quickly get the product developed, demonstrate to stakeholders what the technology can do for them in the healthcare space, and understand how the users could interact with the technology.

Leverage alternative voice platforms. For example, Dr. Rand Hindi founded and created Snips AIR, a decentralized implementation of AI voice technology that guarantees “privacy-by-design”. The platform utilizes blockchain technology to ensure that user data never gets sent to the cloud which is where the security concerns exist with the current voice technologies.

In Summary

When strategically implemented, voice recognition technology in healthcare offers an accessible and engaging healthcare experience. Therefore, healthcare leaders must consider voice-based apps. Ready to learn how voice-based applications can help your organization achieve your big picture goals? Let’s talk. We strive to solve the toughest problems in healthcare.

Contact Us

We’re Here To Help

Whether you’re ready to talk to someone right now or just want general information

Global HQ

240 Elm St., 2nd Floor Somerville, MA 02144