Using data in electronic health records, University of Florida researchers have developed an artificial intelligence system that can predict which patients will develop Alzheimer’s disease up to five years before receiving a diagnosis.
There are currently no early detection tests for Alzheimer’s disease, which is not diagnosed until patients develop symptoms. By this time, the disease has already caused significant brain damage.
Real-world clinical data, like electronic health records, have the potential to be powerful screening tools for diseases like Alzheimer’s disease, said Jiang Bian, Ph.D., chief data scientist at the University of Florida Health and a professor of biomedical informatics into the College of Medicine. Many known risk factors for Alzheimer’s disease, such as obesity, high blood pressure and high cholesterol, are routinely recorded in patients’ electronic health records.
“We are developing and testing models that use artificial intelligence to extract this type of information from patients’ medical records and predict which patients will develop the disease many years before a diagnosis,” Bian said. A study team led by Bian and Jie Xu, Ph.D., a new UF AI collaborator and an assistant professor in the Department of Health Outcomes and Biomedical Informatics, published their findings this week in Alzheimer’s and Dementia, the journal of the Alzheimer’s Association.
The researchers reported that both of the AI models they tested scored in the range of “excellent” to “outstanding,” using standard performance measures that determine how well an AI model performs a given task.
“More testing is needed before these AI tools are available to doctors and their patients,” Bian said.
However, the study shows that it is possible to use patient information in electronic medical records to search for Alzheimer’s. More than 6.5 million Americans age 65 and older are living with Alzheimer’s disease and related dementias. This neurodegenerative disease leads to progressive memory loss and declining cognitive function. Over time, the disease deprives older adults of the skills they need to live independently.
Research suggests that changes in the brain that lead to Alzheimer’s and other forms of dementia begin much earlier than previously thought – perhaps as early as middle age.
“The sooner doctors and high-risk patients can intervene, the better the chances that those interventions will work,” Xu said.
The study team used real patient data, sanitized from patient identifying information, from approximately 16.8 million Floridians housed in the OneFlorida+ Data Trust repository to identify nearly 24,000 patients over the age of 40 diagnosed with Alzheimer’s or dementia . These patients served as a “case” group. Almost 1.04 million patients over the age of 40 who were not diagnosed with dementia served as a control group.
The OneFlorida+ Data Trust is a repository of secure electronic medical records for millions of patients in Florida and select Southeast cities.
The team tested two predictive models and asked each to scan nearly 10 years of patient data and identify patients they knew would later develop Alzheimer’s. The computers were rated on the accuracy of their predictions at four time points: at the time the patient was diagnosed and one, three, and five years before diagnosis.
The team’s knowledge-based model based its predictions on current scientific evidence, including known risk factors for Alzheimer’s disease, such as health conditions, behaviors, lifestyle habits and medications known to contribute to Alzheimer’s, as well as prescriptions for medications indicated for the disease Treatment of Alzheimer’s diseases and related forms of dementia are permitted.
The team’s data-driven prediction model used the same scientific evidence as the knowledge-based model when making predictions, but had the flexibility to consider other available data in medical records that could contribute to Alzheimer’s disease.
“We wanted to see if the data-driven model could identify risk factors and social determinants of health in the data that experts weren’t even aware of,” Bian said.
Both models performed well overall, but the more flexible data-driven model significantly outperformed the knowledge-driven model for predicting Alzheimer’s disease both before and at the time of diagnosis.
Using a machine learning metric known as “area under the curve,” or AUC, to measure how well their AI models predicted Alzheimer’s disease prior to a diagnosis, the study team reported that both of the AI algorithms they tested AUC values of 0.85 to 0.95 were achieved, which are considered “excellent” to “superb”.
The data-driven model topped the list with “excellent” scores of 0.939 for predicting Alzheimer’s disease at the time the patient was diagnosed and 0.906 a year before diagnosis. Although its performance diminished somewhat as the prediction window increased, the data-driven model still achieved “excellent” scores of 0.884 in predicting Alzheimer’s disease three years before diagnosis and 0.854 five years before diagnosis.
The data-driven model also identified several potential risk factors that the knowledge-driven model failed to identify, including muscle weakness, mood disorders, malaise and fatigue. In addition, the data-driven model found that women who receive health care, including regular medical exams, pelvic exams, and mammography screenings, have a lower risk of developing Alzheimer’s disease than women who do not receive such care.
Detecting Alzheimer’s early using screening tests like these is a critical first step in developing effective treatments and improving patient outcomes.
“Interventions initiated during the incubation period of the disease are likely to be much more effective in maintaining or improving cognitive function, delaying symptoms, or even preventing Alzheimer’s disease altogether,” Bian said.
Media contact: Diana Tonnessen, [email protected], 352-294-5972 (office), 352-665-9331 (mobile)
