Australian Senate votes to abolish algorithm for assessing elderly care
The Australian Senate voted to abolish a computer algorithm that determines the amount of government funding for in-home care for elderly people. The government also allocated $18 million to review the algorithm and the Independent Assessment Tool, ABC News reports.
Questionnaire and algorithm
Health and Ageing Minister Mark Butler said that the Independent Assessment Tool, or IAT, was an important recommendation of the Royal Commission into Aged Care. At the same time, ABC notes that the IAT is a questionnaire completed by a clinician while assessing a person's needs, whereas the algorithm is a separate government formula for calculating funding. The Royal Commission did not recommend introducing this algorithm.
The IAT questionnaire was trialled with 22,000 people, and three working groups involving clinicians worked on its development. A detailed report on its development and testing is public. According to ABC, the algorithm was not tested in real-world conditions before its launch, and the Department of Health did not name the clinicians who approved it.
Concerns about the model
Professor Kathy Eagar of the University of New South Wales, who developed a classification algorithm for allocating funding in residential care, analysed the Support at Home system for the Four Corners programme. She concluded that most of the more than 500 IAT questions do not affect the algorithm's decision, while the system cannot process clinicians' detailed notes in more than 100 text fields.
More current news is available on the UA.News Telegram channel Telegram.
Eagar also said that factors such as health frailty, cognitive impairment, mental health and carer exhaustion are taken into account too late to affect the amount of funding. In her assessment, a difference of one point at a certain stage of the algorithm can mean a difference of almost $40,000 in annual funding.
Complaints and appeals
In the first five months of the system's operation, more than 1,000 people requested a review of the algorithm's decision — six times more than during the entire financial year before its introduction. The government said such requests accounted for less than 1% of all assessments. At the same time, the ombudsman is investigating the system following 800 complaints.
The government rejected a Senate amendment that would have allowed clinicians to review an algorithmic decision if they considered it erroneous. Instead, the authorities plan to introduce by law a new mechanism under which assessment agencies will be able to initiate reviews of decisions. Currently, appeals mostly have to be lodged by the elderly person or their family.
