Researchers at the Indiana University School of Medicine, in collaboration with the Regenstrief Institute, Eskenazi Health, the University of Miami School of Medicine, and Lamar University, have developed a groundbreaking artificial intelligence (AI) tool aimed at improving the early detection of Alzheimer’s disease and related dementias (ADRD) in primary care settings. This fully digital, cost-free method requires no additional time from healthcare professionals, addressing a significant gap in timely diagnosis of cognitive impairments.
The innovative approach combines the Quick Dementia Rating System (QDRS) with a Passive Digital Marker (PDM), a machine learning algorithm that utilizes natural language processing to analyze electronic health records (EHRs). By identifying indicators such as memory issues and vascular concerns, this system enhances the screening process for dementia.
In a randomized clinical trial involving over 5,300 patients, the researchers demonstrated that their dual approach increased the rate of ADRD diagnoses by more than 31% within 12 months. This was achieved without requiring extra time or effort from clinical staff, a crucial factor in busy primary care environments.
Transforming Dementia Detection
The team highlighted that this method represents a significant advancement in integrating AI and patient-reported outcomes into routine clinical practice. “This is the most scalable approach to early detection that I know of,” stated Malaz A. Boustani, MD, a research scientist at Regenstrief and co-developer of the PDM. He emphasized that traditional screening methods typically demand at least five minutes of a clinician’s time and often incur licensing fees. In contrast, the new method operates without any clinician time or cost.
Co-developer Zina Ben Miled, PhD, an affiliate scientist at Regenstrief, added that embedding these tools directly into EHRs ensures equitable access to early detection for all patients, regardless of their background. The findings were published in the JAMA Network Open, under the title “Digital Detection of Dementia in Primary Care: A Randomized Clinical Trial.”
The research indicated that over half of older adults in primary care settings do not receive timely diagnoses of ADRD. The limited time available for clinicians, combined with the focus on immediate health concerns and the stigma surrounding dementia, often hampers recognition of the condition.
Enhancing Clinical Efficiency
Digital and paper versions of cognitive performance tests pose scalability and sustainability challenges within primary care. Many existing detection methods depend on clinicians conducting data collection through direct interviews or assessments. Although the FDA has approved a blood-based biomarker for detecting Alzheimer’s disease, other forms of ADRD lack such biomarkers.
The researchers proposed that patient-reported outcome (PRO) methodologies, alongside advancements in EHR data and machine learning algorithms, could help overcome these barriers without imposing additional demands on clinicians.
The trial incorporated the QDRS, a 10-question patient-reported tool, alongside the PDM. All participants were aged 65 and older and had no prior diagnoses of mild cognitive impairment or dementia. The study, conducted across nine federally qualified health centers in Indianapolis, embedded the QDRS and PDM directly into the Epic EHR system. Patients were automatically invited to complete the QDRS survey through their patient portal, while the PDM continuously analyzed clinical data to identify those at risk. Results were promptly available to clinicians, triggering further evaluation only when necessary and requiring no additional manual screening.
The findings revealed that the combination of these two tools led to a 41% increase in follow-up diagnostic assessments, including neuroimaging and cognitive testing. This suggests that the new approach could facilitate earlier access to dementia care for underserved populations.
The QDRS was designed to empower patients and families to report changes in cognitive function accurately and swiftly. James E. Galvin, MD, a professor of neurology at the University of Miami, underscored the significance of using digital tools like the PDM to scale early detection effectively.
Boustani concluded, “This work represents the next phase of our half-century legacy at Regenstrief—using data, innovation, and compassion to transform healthcare delivery.” The researchers highlighted the potential for AI and patient-reported outcomes to redefine clinical practice, making early detection of dementia accessible and efficient.