Speech Recognition Application for the Voice Impaired (SRAVI) V1.0
Research type
Research Study
Full title
A speech recognition application as a communication aid for critical care patients with tracheostomies
IRAS ID
307199
Contact name
Bronagh Blackwood
Contact email
Clinicaltrials.gov Identifier
Duration of Study in the UK
1 years, 5 months, 30 days
Research summary
Summary of Research
Patients in critical care (intensive care or high dependency units) often undergo a tracheostomy. A tracheostomy is an incision at the front of the neck to insert a breathing tube directly into the airway. The tube sits in place in the airway using an inflated air-filled cuff. This means that no airflow is directed up and out past the vocal cords through the voice box, and speech is not possible. Being unable to speak can cause distress to patients and may place them at an increased risk of harm if they are unable to express their wishes or needs. It can also increase stress for relatives and healthcare staff as they try to understand what patients are trying to say. Usually when patients cannot talk, staff use different items to help, like a pen and paper. A new communication device that runs on a smartphone or tablet has recently been developed. It is for patients with tracheostomies and works by reading lip movements and translating them into words on the device screen.The aim of this study is to find out if providing adult critical care patients who have a tracheostomy with the use of this lip-reading device could improve how they communicate. This study will include:
1. Using the lip-reading device in critical care to test if it helps patients with tracheostomies to communicate better.
2.Interviews with patients, relatives and focus groups with staff to find out their views on communication including the use of the new lipreading device.
3. Follow-up with patients approximately 3-months after hospital discharge to complete some further questions about their physical and mental health.The study will take place in three critical care units in Northern Ireland and is expected to last 18 months. The study has been funded by the Public Health Agency Research and Development Division.
Summary of Results
A study in Northern Ireland recruited 31 patients, with a median age of 61 years, across four study sites between January 2023 and April 2024. Most participants (84%) were from critical care, and the most common reason for admission was neurological conditions (35%). Among critical care patients, 87.5% experienced at least one episode of delirium. Researchers evaluated the Speech Recognition Application for the Voice Impaired (SRAVI), which recorded 468 videos of patient communication. The first version of SRAVI, which relied on a set phrase list, had an accuracy of 21.8%, while the second version, allowing for free speech, achieved 34.6% accuracy. Other communication aids were used 1,338 times, with nearly half (48.8%) relying on non-verbal methods such as lipreading, gestures, and facial expressions. Half of eligible participants completed a three-month follow-up.Interviews were conducted with 35 healthcare professionals, nine patients, and five relatives to assess SRAVI’s acceptability, using the Theoretical Framework of Acceptability (TFA). The most frequently discussed themes were the effort required to use the app (burden) and its perceived effectiveness. While attitudes toward SRAVI were generally positive, challenges such as patient acuity and Wi-Fi limitations affected its usability. Suggested improvements included optimising the app for smaller devices and further refining the second version. Analysis of broader communication experiences identified six key themes, including the value of communication, emotional strain from communication difficulties, and the importance of a patient-centred approach.
The study concluded that a larger trial of SRAVI in acute and critical care settings is not currently feasible. Despite some progress with Version 2, persistent accuracy issues highlight the need for further refinements. While iterative development could enhance SRAVI’s design, the complex physical and cognitive challenges faced by critically ill patients may limit the app’s ability to fully support communication in this population.
REC name
HSC REC B
REC reference
22/NI/0109
Date of REC Opinion
20 Jul 2022
REC opinion
Further Information Favourable Opinion