Health十一月 20, 2023

Improving voice technology diversity in healthcare with inclusive interfaces and conversational AI

Voice technology has advanced patient service efficiency in the healthcare landscape, but the 42.2% of the U.S. population identifying as non-white tends not to be represented by the voices used in healthcare interfaces.

An expert view of voice user interfaces

While voice technology can help reduce wait times for patients and allow them faster access to the health information they need, a lack of diversity in voice technology can negatively impact patient engagement and outcomes, according to Freddie Feldman, Director of Voice and Conversational Interfaces for Wolters Kluwer, Health. He spoke on “Closing care gaps with conversational AI, inclusive interfaces, and meaningful patient engagements” for a recent Scottsdale Institute webinar.

Although voice interfaces are prevalent in many aspects of our lives — from film and television to Siri, Alexa, and Google Assistant — white voices dominate, Feldman observes in the webinar. When you do hear non-white voices, they are usually a caricature of the way an ethnic group speaks, which doesn’t necessarily engender trust. “In many ways, this example is a shockingly insensitive portrayal that sheds light on a crucial issue that often goes unnoticed, meaningful patient engagement with diverse populations.” Implementing voice user interfaces (VUIs) with diverse voices helps more patients and members relate to the content and vital health information being shared, Feldman explains.

The importance of voice technology diversity in healthcare

“On average, black individuals and other people of color experience shorter lifespans than their white counterparts,” Feldman says. “They face higher mortality risk due to treatable illnesses, increased maternal mortality rates, greater susceptibility to severe pregnancy-related complications and elevated infant mortality rates.”

The disparity between health outcomes for people of color and whites suggests that the healthcare industry could do more to improve patient engagement and adherence. Meaningful patient engagement is critical to improving patient outcomes and experiences, he says.

What is the role of voice user interfaces in healthcare?

In today’s healthcare environment, VUIs are ubiquitous when accessing information, scheduling appointments, and navigating customer service menus, Feldman notes. Patients often must trust the virtual assistant with personal, sensitive, and sometimes embarrassing information to get the needed services or information.

Why are racial and cultural representation in voice user interfaces important?

Feldman poses the question of whether you can actually hear someone’s race in their voice? “You bet you can. Especially if that race matches your own. You can tell. It’s like radar. And, now, we’re not talking necessarily about the words someone is saying. We’re talking about the voice tone. It’s inflection and accent and the timber or sound of the voice. We’re getting at racial and ethnic disparities in healthcare that are historically pervasive in the United States.”

Trust can be hard to come by when the patient can’t identify with the virtual assistant’s voice, he explains. Utilizing a racially inclusive VUI that patients can identify with forges a bond between the patient and the virtual assistant’s persona, increasing trust and improving patient engagement and adherence.

Wolters Kluwer has used a female voice actor for its UpToDate® patient and member engagement (formerly Emmi®) English voice programs, an approach that is a step up from a synthetic voice and gives the program a human quality. However, although she connects with customers more deeply than an artificial voice, Feldman observes that because she is identifiably white, many users could have difficulty identifying with her voice, creating an unintentional care gap.

What are some examples of diverse voice options in VUIs?

Since 2020, Feldman and his team have been working to bridge the care gap in the VUI, designing interfaces that build trust and rapport in healthcare communication. A critical element of their approach is ensuring racially inclusive voices, beginning with a new Black female voice for programs with the UpToDate Outreach, UpToDate Journeys, and UpToDate Engage solutions. The team also developed a campaign-specific Black male voice for a hospital conducting prostate cancer screening outreach.

The racially inclusive voice user interfaces in the UpToDate engagement programs are among the first in the healthcare ecosystem. Indeed, VUI diversity is rare across industries and environments. As Feldman notes, the only voice choice in most commercial user interfaces is gender, though a couple of platforms are exploring a racial options.

We envision a future where proactive prevention and patient engagement takes center stage in healthcare delivery. Patients will be empowered with knowledge to make informed decisions about their health by voices they can better relate to and trust.
Freddie Feldman, Director of Voice and Conversational Interfaces for Wolters Kluwer, Health

The possibilities for using conversational AI in healthcare VUIs

Technological developments often lead to rapid and significant changes in the healthcare industry. Conversational AI is one such development that has the potential to transform information delivery systems and improve the patient experience. During the webinar, Feldman dove into the possibilities for using conversational to enhance the diversity of voice user interfaces in the healthcare ecosystem.

Committing to diversity in voice technology is essential because “our world is beautifully diverse,” Feldman says. “People come from different backgrounds, cultures and experiences. Each individual’s health journey is influenced by these factors. If our healthcare systems and solutions don’t reflect this diversity, we will inevitably miss out on addressing some crucial needs. And that’s where care gaps pop up.”

How does conversational AI work?

Conversational AI is adaptive technology that utilizes machine learning, artificial intelligence, and natural language processing (NLP) to understand human language and user intent. NLP allows a computer system to interpret voice or written language, deciphering its meaning without relying on correct grammar and syntax. It’s why you can input a few words into a search engine search box and receive results that match your search.

How can conversational AI improve the patient experience?

You’ve probably already encountered conversational AI in healthcare without necessarily realizing it, Feldman says. Today’s healthcare consumer often turns to technology to access information. When implemented well, conversational AI allows patients to access relevant information quickly. Currently, patients can use the technology to:

  • Find answers to health questions
  • Assess symptoms and offer advice on next steps
  • Manage medication schedules and prescription refills
  • Schedule appointments
  • Track and follow post-treatment care

Rapid, automated responses and access to accurate and relevant information quickly provide patients with what they need. Additionally, they don’t need to insert specific keywords in the system to get the right results. Conversational AI’s NLP interprets and understands the language people actually use.

Combining conversational AI with racially inclusive voices for voice user interfaces can improve user engagement, as system responses align with the patient’s natural vocal patterns. As Feldman suggested in an example during the webinar, when Mia, an elderly African-American woman living alone, hears a calm, soothing voice that reminds her of her daughter’s way of speaking, she feels comforted and seen. She is more likely to trust what she hears and adhere to her treatment.

What are the privacy and ethical concerns with conversational AI?

The idea of AI often elicits either excitement or fear, and there is cause for both, Feldman says. Responsible use of conversational AI provides numerous opportunities in healthcare. However, the privacy and ethical concerns must be addressed.

On the patient side of the system, the biggest concern is maintaining patient privacy. “How do we put guardrails in place to limit the breadth of information where the patient can go? Also, you know, on a privacy level, how do we prevent PHI [personal health information] from getting passed in and getting sort of trapped within the brain?” Feldman asks.

Feldman indicates that the healthcare industry can’t create these tailored, racially inclusive VUIs without implementing the same safeguards utilized to protect patient privacy in other healthcare systems. VUI designers must not forego privacy and personal health information security for the sake of personalized voice user interfaces.

Another ethical concern is the potential for racial profiling based on the sound of a caller’s voice. Furthermore, with guardrails in place, a malicious actor could access voice recordings in the system and co-opt those voices for nefarious purposes.

“So, you [have to] be careful with what you feed these tools [when] giving it samples of your own personal information, but also…try to stay up on it because otherwise it can get away from you. And then one day you’re like, I don’t totally understand this. And in my hands it’s dangerous a little bit, right?” he cautions.

Watch the webinar

To learn more, watch the webinar, “Closing care gaps with conversational AI, inclusive interfaces, and meaningful patient engagements.” This webinar was first presented to the Scottsdale Institute Members.
Back To Top