‘Game-changing’ AI breakthrough could transform dementia diagnosis, uncover thousands of missed cases

A “game-changing” new method for improving dementia detection using artificial intelligence (AI) could pick up thousands of undiagnosed people living with the disease and help to determine the true prevalence of dementia in Australia.

Currently, 50 million people worldwide are estimated to be living with dementia, and this number is expected to triple by 2050, according to the World Alzheimer Report. However, routine health data currently used for tracking dementia likely underestimates the actual number of people affected.

While hospital admissions are a key source for dementia statistics, many potential patients remain undetected. This is because identification relies on information gathered from medical records by coders, who often struggle to sift through the vast amount of written data.

Additionally, there is no single authoritative data source for deriving dementia prevalence in Australia, so many of those living with the condition can be missed by health workers.

To tackle this problem, researchers from Melbourne’s Monash University (MU) and the Victorian state government’s NCHA’s Healthy Ageing Data Platform group undertook a study which investigated how AI could improve our ability to diagnose those with dementia missed by current tests and data gathering methods.

Using traditional data approaches with AI in electronic health records, the researchers poured over the records of over 1000 participants aged 60 and over in the Frankston-Mornington Peninsula, Victoria.

Publishing their findings in the Alzheimer’s & Dementia Journal, the research team showed that algorithms combining traditional methods with a particular type of AI called natural language processing or NLP (which is applied to written text in medical records) significantly enhanced dementia identification.

To test the efficacy of NLP, lead author Dr Taya Collyer said the researchers separated the participants’ data into two groups. The first group contained the data for those with dementia who were diagnosed by specialists, while the second group comprised records for those without a dementia diagnosis.

“Accessing high-quality curated electronic health records from our Healthy Ageing Data Platform helped assemble the data efficiently to address this problem,” Dr Collyer said.

“Special software was used to harness the large amount of free text data in a way that NLP could then be applied.

“We then developed dementia-finding algorithms through a traditional stream for usual structured data and an NLP stream for text records.”

For the first group, information was obtained about the participants’ demographics, medications, socioeconomic status, in-hospital events such as confusion or distressed behaviour and emergency and clinic health appointments.

Then for the NLP group, the team used clinical experts to guide the analysis and ensure its clinical relevance.

Research lead, Professor Velandai Srikanth said the findings are exciting as the efficient identification of people with a high probability of dementia will get them the assistance they need but who may get missed otherwise.

“Given that clinical recognition of people diagnosed with dementia presenting to hospitals is poor, using this new approach we could be identifying people earlier for appropriate diagnostic and clinical care,” Prof. Srikanth said.

“I am sure that many people are missing out on good care because we are not very good at identifying them or their needs.”

In addition, Prof. Srikanth said this new method gives healthcare professionals a new digital strategy that captures and combines cues in written text such as descriptions of confusion or forgetfulness, or alerts for distressed behaviour, which they can flag for suitable care.

“Responsibly using AI in scientific research and dementia identification is potentially game-changing,” Prof. Srikanth said.

“The NCHA’s Healthy Ageing Data Platform, an Australian-first initiative, has been able to bring together various sources of data from electronic health records, safety and governance, and the technical capacity to enable such high-value projects.”

IMPORTANT LEGAL INFO This article is of a general nature and FYI only, because it doesn’t take into account your personal health requirements or existing medical conditions. That means it’s not personalised health advice and shouldn’t be relied upon as if it is. Before making a health-related decision, you should work out if the info is appropriate for your situation and get professional medical advice.