The Future of Voice User Interface in Healthcare
The potential for voice-user interface (VUI) technology has immense potential for applications within the healthcare field. With over 24% of U.S. households using either Alexa, Siri or Google, users are familiar with using voice commands and interacting with the voice-user interface (VUI). With voice technologies, healthcare providers can optimize their time and resources, and patient care can be maximized using artificial human responses.
Future uses for voice technologies in healthcare are wide-ranging and varied. They include:
• Selecting a doctor/scheduling appointment
• Patient care (e.g., diabetic care)
• Searching for health information and treatment options
• Charting by healthcare personnel
• Managing drug adherence
• Maintaining healthy behaviors
• Surgery and the intensive care unit procedural checklists
• Preoperative organ transplant validation
• Patients providing data between visits
• Assisting the patient to prepare for procedures
• Standardizing care information given before/after treatment
• Eliciting responses during the care process
Presently, interactive phone calls enable healthcare stakeholders to automatically reach out to their entire patient population. However, in the future, advanced voice technology will be more than a convenience. Mixed with Natural Language Processing (NLP), Artificial Intelligence (AI), and Machine Learning, voice-based applications will exceed the current command-and-answer functionality. NLP is a branch of AI that helps computers understand, interpret and manipulate human language. NLP, along with machine learning, help computers “understand” human speech, which is incredibly complex and nuanced.
The process starts with VUI designers mapping out numerous responses for each state of patient interaction. Designers then identify various conversation flows and decision trees and word choices. Conversation flows also incorporate virtual memory so that subsequent conversations with patients start with past information accounted for.
A developer then builds code to follow the various paths that the voice designer has mapped out. The first step here is to transcribe voice into text (voice recognition). While voice recognition is relatively advanced and established, it cannot “understand” what has been said. For nuanced and personal interactions, once a machine has converted voice into text, it then has to “understand” what has actually been said. Here is where NLP comes in. The output of the NLP process is called “semantic interpretation”.
Taking things from the spoken word to the semantic representation enables the programmer to define how the computer should respond to each of the semantic categories (what actions to take). The computer seems to simulate the “understanding” of spoken speech, but that “understanding” is based on the programmed rules. The rules may come from training efforts using machine learning or substantial rule-coding efforts. Both approaches work.
One of the significant limitations to using voice technology today in healthcare is the lack of HIPAA-compliance. Currently, patient protected health information can't be integrated within the voice ecosystem. Appropriate safeguards will need to be included to meet the requirements of the HIPAA Security Rule. Compliance with HIPPA will be required to allow for more personalized experiences.
However, this limitation is being addressed: Alexa has recently announced six HIPPA Compliant skills: the new skills allow patients to schedule appointments, find urgent care centers, receive updates from their care providers, access their latest blood sugar reading, and check the status of their prescriptions
Voice technologies, along with compliance to HIPPA, are sure to create a demand for the use of voice. Once the technology is more advanced and protections are in place, voice technologies are sure to bring greater efficiency to healthcare, and to personalize patients’ experiences across the complete health value chain.