UCSF study: AI model can spot early Alzheimer's risk factors
Add Axios as your preferred source to
see more of our stories on Google.
The UCSF Helen Diller Medical Center. Photo: Gado via Getty Images
Scientists at UC San Francisco say they have discovered an AI-powered method for predicting Alzheimer's disease up to seven years before symptoms appear.
Why it matters: Alzheimer's affects nearly 7 million Americans — nearly two-thirds of whom are women, per the Alzheimer's Association.
- Early detection and treatment are key for people impacted by the progressive and fatal disease.
How it works: The model analyzes patient records with machine learning to spot patterns in clinical data and identify risk as early as possible.
- Using UCSF's clinical database, the researchers examined co-occurring conditions in patients who had been diagnosed with Alzheimer's and those without.
- They found they could identify who would develop the disease up to seven years prior with 72% predictive power.
- The conditions that most influenced predictions included high cholesterol, as well as osteoporosis, specifically for women, according to their study.
Yes, but: The prediction model relies on health data that may not always be comprehensive or up to date.
- The researchers also acknowledged that data inputs can be influenced by other elements like clinician or patient behavior, sociological factors and implicit bias.
- "Due to changing patient demographics and societal factors, prediction models should be continuously trained, updated and evaluated ... to ensure effective utilization and account for biases that may have been learned from the data," the study notes.
The big picture: California has one of the highest rates of Alzheimer's disease in the country.
What to watch: Researchers plan to assess how the model can be used to help treat other diseases that are similarly difficult to diagnose, like lupus and endometriosis.