Research carried out by Prof Cal Muckley of UCD Michael Smurfit Graduate Business School has shown that money management difficulties can help identify individuals with early-stage dementia up to four years earlier than current methods allow.
The importance of that early diagnosis is difficult to overstate. “Within the last two years new drugs have been licensed that can reduce the progress of the disease by up to 30 per cent,” Prof Muckley explains. “That’s really remarkable and could result in several years of much better quality of life for the individuals involved. But the drugs work best when given early. Late detection is the standard at present and that’s a real problem.”
In those circumstances, a four-year head start could make a dramatic difference to people’s lives.
That research will be among the topics discussed at a symposium on AI for Good – A Transdisciplinary Approach to Tackling Dementia at the UCD Lochlann Quinn School of Business, Belfield on Wednesday, March 12th.
Along with Prof Muckley’s research, the event will feature health and financial experts discussing evidence that financial abuse is more likely if an individual has severe cognitive decline and how the biomedical sciences need earlier detection of dementia to have better visibility of the evolution of the pathogen.
Among the speakers will be Annmarie O’Connor, programme manager for Safeguarding Ireland and chairwoman of its State Payments Sub-Group who will address the development of good practice response to financial abuse. “She has some great insights into people who are experiencing financial abuse when they are suffering from cognitive decline,” says Muckley. “Those people are very vulnerable and not thinking clearly and that can result in abuse from different sources.”
Associate Prof David O’Connell of the UCD School of Biomolecular and Biomedical Science will discuss current developments towards the detection and treatment of Alzheimer’s disease. “He will look at the latest advances in early detection and the potential for adding financial indicators to the mix,” Muckley points out.
“Our keynote speaker is Prof Carole Roan Gresenz of the McCourt School of Public Policy and the School of Health at Georgetown University,” he continues. “She will address the financial consequences of undiagnosed memory disorders. Her research looks at the financial mistakes people make before diagnosis. She doesn’t use the data to predict at the moment but is about to do that. We are delighted to have her as our keynote speaker.”
Muckley’s research into elder financial exploitation goes back some years and in 2021 he published the findings of a 30-month project with a big global financial institution. Using data about people’s accounts at the institution, Muckley and his team were able to develop an alert model to predict, detect, and prevent instances of elder financial abuse.
His latest research builds on those findings. “I became interested in this area from a human rights and customer protection perspective,” he says. “In the first piece of research, it became clear that cognitive decline is a lead indicator of EDE. After that, I asked if money management difficulties would be an indicator of cognitive decline and dementia. If we could predict years in advance and identify a condition not yet diagnosed that would be of interest to financial institutions which have a duty of care to protect customers at risk of financial abuse.”
The project looked at established lead indicators for dementia and added money management. “We used AI and machine learning to look at it and the model worked very well. We were able to demonstrate that money management difficulties were the most important indicator apart from age out of 12 others.”
Financial institutions can use that information to enhance the protection of customers with dementia and, potentially, inform the inter-generational transfer of financial control to a reliable agent such as a relative or legal representative.
The predictions produced by the models are of very high quality. “They are accurate for 91 to 92 per cent of people in single-person households and for 71 to 72 per cent of the wider population four years in advance,” he says. “That’s a very promising true positive rate. In principle, this could enable a variety of financial and medical interventions but there are complex legal and data privacy questions to be addressed.”
While financial institutions have access to customer data, difficulties arise when it comes to sharing it. “You would probably have to set up an opt-in system for customers to allow the data to be shared with medical professionals, family members or others. This can help to ensure the financial autonomy of people for longer by providing an additional safety net. I’m so happy to be working in this space. It’s probably the best piece of scholarship I’ve ever done.”
- To register for AI for Good – A Transdisciplinary Approach to Tackling Dementia, see smurfitschool.ie
