Development and Validation of a Deep Learning Model for Earlier Detection of Cognitive Decline From Clinical Notes in Electronic Health Records

Authors:
Open Access
Publisher website
Google Scholar
Abstract
There are nearly 6 million people diagnosed with Alzheimer disease (AD) at the stage of dementia in the US, and the prevalence increases dramatically with age.1 Mild cognitive impairment (MCI) and subjective cognitive decline (SCD) represent precursor stages that can serve as targets for early treatment.2-4 Early detection of cognitive decline can facilitate enrollment in clinical trials and early interventions.5,6 The US Food and Drug Administration recently approved aducanumab, a drug directed at the underlying pathologic characteristics of AD that clears amyloid plaques in the brain, to treat patients with AD.7 However, detecting patients with cognitive decline is challenging. There is an insufficient number of specialists with the necessary expertise (behavioral or cognitive neurologists, geriatric psychiatrists, geriatricians, and neuropsychologists) to see all at-risk patients. Instead, primary care physicians and other nondementia specialists have direct contact with these patients but not necessarily the time or tools needed for diagnosis.8

This publication has 20 references indexed in Scilit: