Beware Using Telemedicine for Voice and Speech Therapy

As a result of the coronavirus pandemic, people across the world have experienced how teleconferencing platforms like Zoom help folks stay connected - playing games with friends, hosting virtual weddings, and even visiting a doctor. But when it comes to telemedicine, not all medical care is easily translated to a remote format.

In a virtual world, voice therapy presents a unique challenge because clinicians must rely on acoustic recordings of voice to evaluate the effectiveness of their treatments. But many teleconferencing platforms distort sounds in their efforts to eliminate background noise.

Boston University graduate researcher Hasini Weerathunge wanted to find out if popular teleconferencing platforms used for telemedicine could capture sounds accurately enough for clinicians to successfully treat and evaluate patients with voice and speech disorders. Weerathunge, a graduate student fellow at BU's Rafik B. Hariri Institute for Computing and Computational Science & Engineering and a Ph.D. candidate in biomedical engineering, does research in the lab of Cara Stepp, a College of Health & Rehabilitation Sciences: Sargent College associate professor of speech, language, and hearing sciences.

"Although the COVID-19 crisis appears to be waning, telepractice popularity is here to stay," Stepp says.

Weerathunge and Stepp teamed up with other BU researchers to put five different HIPAA-compliant teleconferencing platforms to the test: Cisco Webex, Microsoft Teams, Doxy.me, VSee Messenger, and Zoom.

As the pandemic unfolded and lockdowns moved much voice and speech therapy online, "there was no consensus among [voice and speech] clinicians [who were] trying to convert to telepractice therapy, and we wanted to determine the accuracy of the acoustic measures they can get through telepractice," Weerathunge says.

Although voice therapists had sometimes conducted telepractice sessions with patients before the pandemic, evaluations of the effectiveness of treatment were always carried out in person. During that process, a patient goes into the clinic and sits in a soundproof booth outfitted for speech recordings. The patient repeats sustained vowel sounds, like "aaa" or "ooo," or reads a short passage that reflects a wide variety of sounds and mouth movements in the English language. The recordings of the patient's voice are then evaluated by algorithms that measure acoustic properties, including the acoustic correlates of perceived pitch and loudness of the voice.

In-person voice evaluations came to halt, however, at the height of the COVID-19 pandemic. Voice evaluations moved to a virtual format, but until now, the accuracy of those evaluation procedures done online has never been examined.

In a soundproof room, the team recorded voice samples from 29 patients, aged 18 to 82, that had a variety of speech or voice diagnoses. These recordings were then played back to researchers through an external speaker over the teleconferencing platforms, simulating telepractice conversations.

The team quickly learned that each platform has its own audio enhancement algorithm that affects the quality of the sound. Zoom was the only platform that enabled users to turn off these audio enhancement features, allowing the researchers to test the platform's original audio.

Despite the enhancements, the team predicted the ability to measure vocal fundamental frequency (pitch) and vocal intensity (loudness) through teleconferencing platforms is not significantly affected.

But the researchers discovered that all the teleconferencing platforms did a poor job at capturing many measurements needed for accurate and clinically meaningful voice evaluations. Pitch varied significantly on all the virtual platforms compared to the real-life recordings. This might be due to internet connection or bandwidth issues that affect how and when sounds get transmitted through the platforms, the researchers say.

They also found the dynamic range of the vocal loudness measured over telepractice was very different from live recordings. "This was the biggest surprise for us," Weerathunge says. The effect was even true for Zoom, where the researchers could turn off the audio enhancements.

Overall, "Microsoft Teams performed the best, in that all our voice measures were the least affected in that platform," Weerathunge says.

Because many of the voice metrics collected from virtual platforms had clinically significant differences from those collected in person, Weerathunge and the team urge caution for voice and speech therapists using telepractice.

"This work is likely to have substantial impact on clinical practice, providing crucial information about the effects of these telepractice platforms on clinical voice evaluations," Stepp says.

Hasini R Weerathunge et al.
Accuracy of Acoustic Measures of Voice via Telepractice Videoconferencing Platforms.
Journal of Speech, Language, and Hearing Research, 2021. doi: 10.1044/2021_JSLHR-20-00625

Most Popular Now

AI-Powered CRISPR could Lead to Faster G…

Stanford Medicine researchers have developed an artificial intelligence (AI) tool to help scientists better plan gene-editing experiments. The technology, CRISPR-GPT, acts as a gene-editing “copilot” supported by AI to help...

Groundbreaking AI Aims to Speed Lifesavi…

To solve a problem, we have to see it clearly. Whether it’s an infection by a novel virus or memory-stealing plaques forming in the brains of Alzheimer’s patients, visualizing disease processes...

AI Spots Hidden Signs of Depression in S…

Depression is one of the most common mental health challenges, but its early signs are often overlooked. It is often linked to reduced facial expressivity. However, whether mild depression or...

ChatGPT 4o Therapeutic Chatbot 'Ama…

One of the first randomized controlled trials assessing the effectiveness of a large language model (LLM) chatbot 'Amanda' for relationship support shows that a single session of chatbot therapy...

AI Tools Help Predict Severe Asthma Risk…

Mayo Clinic researchers have developed artificial intelligence (AI) tools that help identify which children with asthma face the highest risk of serious asthma exacerbation and acute respiratory infections. The study...

AI Model Forecasts Disease Risk Decades …

Imagine a future where your medical history could help predict what health conditions you might face in the next two decades. Researchers have developed a generative AI model that uses...

AI Distinguishes Glioblastoma from Look-…

A Harvard Medical School–led research team has developed an AI tool that can reliably tell apart two look-alike cancers found in the brain but with different origins, behaviors, and treatments. The...

Smart Device Uses AI and Bioelectronics …

As a wound heals, it goes through several stages: clotting to stop bleeding, immune system response, scabbing, and scarring. A wearable device called "a-Heal," designed by engineers at the University...

AI Model Indicates Four out of Ten Breas…

A project at Lund University in Sweden has trained an AI model to identify breast cancer patients who could be spared from axillary surgery. The model analyses previously unutilised information...

Overcoming the AI Applicability Crisis a…

Opinion Article by Harry Lykostratis, Chief Executive, Open Medical. The government’s 10 Year Health Plan makes a lot of the potential of AI-software to support clinical decision making, improve productivity, and...

Dartford and Gravesham Implements Clinis…

Dartford and Gravesham NHS Trust has taken a significant step towards a more digital future by rolling out electronic test ordering using Clinisys ICE. The trust deployed the order communications...