As their condition progresses, a person with dementia finds understanding what they hear increasingly difficult. They may also lose their speech completely, or be limited to repeating a few words or crying out.
This communication breakdown can be hugely frustrating for sufferers, their loved ones and their carers. It also makes it a major challenge to assess whether the individual is suffering from pain – which regularly affects up to half of people with dementia – as they are unable to self-report it.
Dementia Support Australia (DSA) – a government backed partnership led by HammondCare – estimates that more than 70 per cent of its clients are experiencing under-treated or undiagnosed pain which significantly impacts their quality of life.
“The ability to communicate or be understood by caregivers can be impacted by someone living with dementia,” explains associate Professor Colm Cunningham director of the Dementia Centre for HammondCare.
“When that person is in pain, it can sometimes be confused with aggressive behaviour when the reality is much different.”
Now, following a hugely promising pilot in Western Australia and South Australia in recent months, DSA is rolling out an app which can determine if someone is experiencing pain using automated facial recognition technology.
See your pain
The tool – by ASX-listed company PainChek – runs analysis on a 10-second video of an individual’s face, to detect pain related expressions including brow lowering, cheek raising, tightening of eyelids, wrinkling of nose, raising of the upper lip, parting lips and closing eyes.
After capturing the video the carer then completes a five point checklist while observing the individual on their voice (for example selecting that they are groaning or sighing), movement (selecting that they are sitting abnormally or guarding a part of their body), behaviour (are they acting confused, aggressively or are they verbally offensive), activity (altered routines or not sleeping at night) and body (whether the individual has a painful recent injury or medical condition).
After submitting facial assessment and form, the carer is given a score for the individual based on the pain they are experiencing.
“This technology allows consultants who have been called to assist with someone where behaviour impacts on their care to assess and identify if that person is in pain,” said Cunningham.
A pilot study was run on 118 clients aged between 57 and 98 with various types of dementias or cognitive impairments, in September last year and April this year.
The study found strong indications the app performed better than the commonly used, Australian standard for pain assessment (the Abbey Pain Scale).
“When pain was identified via PainChek at the initial consultation and where subsequent pain-related recommendations were implemented, there was a significant improvement in their behavioral scores at the time of discharge from the DSA service,” PainChek said.
“This supports the usefulness of the tool across community and residential aged care settings,” the company added.
The tool – which was initially developed by researchers at Curtin University’s School of Pharmacy and Biomedical Sciences and known – is now being issued to DSA’s 150 consultants in six states, who provide care to around 5,000 people with dementia. They have conducted around 400 assessments so far.
PainChek is now seeking to adapt the app to cater for other groups that have trouble communicating; including infants and pre-verbal children.
“Twenty per cent of children have chronic pain, with common causes being headaches and gastrointestinal or musculoskeletal conditions. And that pain can produce a whole range of issues, such as behavioural problems, poor interaction with others and avoiding school. Most people think that little kids don’t feel pain the way adults do, but we’re learning this isn’t the case,” says Curtin Professor Jeff Hughes, who continues to lead the research effort as Painchek’s chief scientific officer.
Much like the adult app, the children’s app will include the carer questionnaire, but the facial recognition element will be more useful due to the fact that children typically use more pain-associated facial expressions than adults.
In recent months Hughes’ team has been capturing videos of children who are in pain – mainly from footage of them being immunised – with each video contributing to a database of coded images.
A preliminary algorithm has already been written, Hughes said, and PainChek plans to have the first prototype available for trialling in 2018.