Remarkable for a Machine: Home Care Chatbots Among AI Tools Adopted by the Australian Healthcare Sector

Peta Rolls came to anticipate getting Aida's regular check-in each morning.

A routine morning call by an automated voice assistant wasn't initially included in the care package Rolls expected when she signed up for the home care but when they asked to be part of the trial four months ago, the 79-year-old agreed because she wanted to help. Even though, truth be told, her expectations weren't high.

Nevertheless, when the call came through, she states: “I was amazed by how interactive she was. It was remarkable for a machine.”

“The system would inquire ‘how are you feeling today?’ and that gives you an opportunity if you’re feeling sick to say you felt sick, or I might reply ‘I’m fine, thank you’.”

“She would go on to ask questions – ‘have you had a chance to step outside today?’”

The virtual assistant would also ask what the user had planned for the day and “it would reply appropriately.”

“If I would say I plan to go shopping, she’d say nice shopping or food shopping? I found it entertaining.”

AI Reducing the Administrative Burden on Healthcare Staff

This pilot, which has now wrapped up its first phase, is an example in which advances in artificial intelligence are being integrated in healthcare.

Digital health company the provider approached the care organization regarding the program to utilize its advanced AI system to provide social interaction, as well as an option for home care clients to report any medical concerns or issues for a caregiver to follow up.

Dean Jones, head of St Vincent’s At Home, explains the AI check-in under evaluation is not a substitute for any face to face interactions.

“Recipients continue to get a regular face to face meeting, but between these meetings … the automated system allows a routine call, which can then flag any potential concerns to either our team or a client’s family,” the director says.

Dr Tina Campbell, the CEO of the company, says there have been no any negative events reported from the St Vincent’s trial.

The company uses open AI “with strict safety protocols” to guarantee the conversation is safe and procedures are in place to respond to serious health issues quickly, Campbell states. As an instance, if a patient is reporting heart symptoms, it would be alerted to the medical staff and the conversation terminated so the person could dial triple zero.

Campbell thinks artificial intelligence has an important role given significant workforce challenges throughout the healthcare sector.

“The benefit very safely, using such systems, is lessen the admin burden on the staff so trained clinicians can focus on performing the duties that they’re trained to do,” she says.

AI Not as New as Often Believed

An expert, the founder of the national AI health alliance, explains established types of AI have been a standard part of medicine for a long time, frequently in “administrative functions” such as interpreting scans, cardiograms and pathology test results.

“Software that carries out a function that requires judgment in certain aspects is artificial intelligence, regardless of how it achieves that,” says the professor, who is additionally the head of the Centre for Health Informatics at Macquarie University.

“If you go the imaging department, radiology department or pathology lab, you’ll see software in machines doing just that.”

In recent years, advanced versions of AI known as “deep learning” – an algorithmic approach that allows algorithms to analyze very large sets of data – have been used to interpret diagnostic scans and enhance detection, the expert notes.

Recently, BreastScreen NSW became the nation's first population-based screening program to introduce machine reading technology to support specialists in interpreting a specific set of mammography images.

These represent advanced systems that still require a qualified physician to interpret the findings they might suggest, and the responsibility for a clinical judgment sits with the healthcare provider, Coiera emphasizes.

AI’s Role in Early Disease Detection

A research center in Melbourne has been working alongside researchers from UCL London who first developed AI methods to detect neurological lesions known as specific brain malformations from brain scans.

These abnormalities cause seizures that often cannot be controlled with medication, so surgical intervention to remove them becomes the only treatment available. But, the procedure can proceed if the doctors can pinpoint the affected area.

In research published this week in the journal Epilepsia, a group from the institute, headed by neurologist the lead researcher, showed their “AI epilepsy detective” could detect the abnormalities in nearly all of cases from advanced imaging in a specific form of the lesions that have traditionally been missed in more than half of patients (60%).

The system was trained on the images of 54 patients and then tested on 17 children and 12 adults. Of the 17 children, 12 had surgery and 11 are now seizure free.

The tool employs neural network classifiers comparable with the mammography analysis – highlighting regions of abnormality, which are still checked by experts “speeding up the process to reach a conclusion,” the researcher says.

She emphasises the researchers are still in the “early phases” of the project, with a additional research required to advance the tool toward real-world use.

Prof Mark Cook, a neurologist who was independent from the study, says MRI scans now generate such huge amounts of detailed information that it is hard for a human to review it accurately. So for doctors the challenge of finding these lesions was like “identifying the needle in the haystack.”

“This illustrates of how AI can support clinicians in making quicker, precise identifications, and has the ability to improve operation opportunities and results for kids with treatment-resistant seizures,” Cook comments.

Illness Identification in the Future

Dr Stefan Buttigieg, the vice-president of the European Public Health Association’s AI health division, explains deep neural networks are also helping to track and forecast epidemics.

The expert, who spoke recently at the Public Health of Australia’s conference in Wollongong, gave as an example Blue Dot, a organization set up by medical experts and which was an early detector to detect the coronavirus pandemic.

Generative AI is a additional branch of machine learning, in which the system can produce original material based on training data. Such applications in medicine include tools such as the virtual assistant along with the automated note-takers doctors and allied health professionals are adopting more.

Dr Michael Wright, the president of the national GP body, reports GPs have been embracing digital assistants, which records the appointment and converts it to a consultation note that can be included in the patient record.

Wright states the primary advantage of the scribes is that it enhances the quality of the communication between the physician and individual.

A medical leader, the president of the national doctors' group, concurs that AI note-takers are assisting doctors manage schedules and says artificial intelligence can also help to prevent repeated examinations and scans for their patients, if the {promised digitisation|planned digitalization