“Alexa, tell the nurse…”
Three healthcare facilities in Australia and New Zealand have deployed a new voice-controlled system to help patients notify nurses when they need assistance.
The digital replacement for the call button was developed by Deloitte and is being used in Prince of Wales Hospital, an aged care facility in Sydney and Christchurch’s Burwood Hospital.
The solution, DeloitteAssist, uses an Amazon’s Alexa skill to identify what patients are asking for. That information is passed on to the ServiceNow platform which applies business rules to the data and alerts nurses via an ipad at the nurses station or their mobile device.
If patients are not able to reach the traditional call button, they can wake up the smart speaker with the command “Alexa, tell the nurse” to ask for assistance for things like using the bathroom, requesting a glass of water, or to say they’re in pain.
Rather than a generic alarm, the system provides important contextual information to nurses so they understand what the patient is asking for before responding to a call.
The business rules can also be configured at a ward level to determine priority or route the request to the right person, for example “I need a blanket” could be routed to a nursing assistant as opposed to a nurse.
Nick White, principal at Deloitte, who leads to work on DeloitteAssist told Which-50 the solution was designed to update the antiquated technology systems that carers rely on to assist patients.
White said Prince of Wales in Sydney, which went live with the technology in its spinal ward 12 months ago, has reported a significant drop in complaints regarding nurse wait time and improved patient experiences.
By using a consumer-based device as the smart speaker, patients also get access to other Alexa skills which let them music or white noise, tell the time or ask about the weather.
“For somebody who has had an accident or an injury and has ended up in hospital it can be incredibly isolating. To be able to have access to connect back to your daily life by listening to the radio or drowning out the sounds of the ward, it really helps to combat that isolation,” White said.
White said Deloitte spent a lot of time thinking about the way people speak and building out the complex model behind the scenes to allow it to work.
“We wanted to make sure we didn’t deploy a system into a clinical environment that had a low [word] recognition rate because we would essentially get rejection from the users,” he said.
To ensure the system wouldn’t be rejected by patients, the team made sure it achieved a 95 per cent recognition rate in a lab environment before deploying it in a clinical environment.
“We’ve found is that even people with accents are fine using the system, they may have a lower recognition rate but the way the system interacts with them it hasn’t caused rejection of the system,” he said.