Patient Knows Best: Catching Early Cognitive Decline

Up to 90% of mild cognitive impairment goes undetected during routine primary care encounters, but streamlined digital assessments and closer attention to patients’ concerns could help close this diagnostic gap, investigators said at the 2026 annual scientific meeting of the Gerontological Society of America in National Harbor, Maryland. Findings from several studies highlight practical strategies to detect early dementia in busy clinics. Researchers spotlighted MyCog, a brief electronic health record (EHR)-integrated digital screening tool, alongside evidence showing that women’s self-reported cognitive changes correlate with measurable deficits and warrant clinical follow-up, not dismissal. “The benefit of primary care is that the majority of older adults are engaged with a healthcare provider, so it’s a great place as a frontline for mobilizing these care teams,” said Michael Wolf, PhD, MPH, director of the Institute for Public Health and Medicine at the Feinberg School of Medicine at Northwestern University in Chicago, a co-author on a study of MyCog. Wolf and his team developed a self-administered cognitive screening tool, MyCog, for use on an iPad in the office or at home using a mobile device. The tool was derived from the National Institutes of Health’s Toolbox for Assessment of Neurological and Behavioral Function.Two assessments are included in MyCog: measures of executive function and cognitive flexibility. Patients are asked to match pictures on the basis of color or shape, and reaction time is factored into scoring. The other assessment shows patients a series of images in a specific sequence. The pictures are then shuffled, and patients are asked to put them back in sequential order.The screening could be included in a Medicare annual wellness visit or upon suspicion of cognitive impairment. A medical assistant could administer the test, with results uploaded to the EHR. The program then provides interpretable results and recommendations for clinicians to review. Wolf and his team implemented one version of the tool in 2025 at 14 intervention sites, with uptake as high as 77%. Early results showed that the tablet tool, performed on 145 patients, yielded an area under the curve (AUC) of 0.890 out of 1.0 in identifying cognitive impairment. The smartphone model, tested in 277 patients to identify mild cognitive impairment only, yielded an AUC of 0.827. Both methods had a sensitivity of about 0.7-0.8. Their team is currently studying performance of the tool at 24 Oak Street Health sites and a mobile version for patients at home and at clinics affiliated with Northwestern University. Other CluesOther research presented at the conference suggests that older women who sense they have some cognitive deterioration are likely to be correct. Clinicians often do not take patients' complaints of decline seriously, especially after they have performed well in cognitive tests. Clinicians may attribute complaints to depression or stress, said Stephanie Cosentino, PhD, a professor of neuropsychology at the Taub Institute for Research on Alzheimer’s Disease and the Aging Brain at Columbia University in New York, who helped lead the research. “If you measure subjective cognitive difficulties appropriately, you may be picking up on early Alzheimer’s-related changes,” she said. “This is something that we need to be paying more attention to, especially in primary care settings where patients will say, ‘I just feel like I’m not myself anymore.’” Cosentino and colleagues tested a cognitive screening test, the Screener for Cognitive Problems in Everyday Life, on 342 women ages 45 or older. They were asked about concerns related to any changes to their memory and thinking, and whether any family or friends had concerns about their cognitive abilities. Participants also completed the Linus Health Digital Clock and Recall test, which includes a three-word recall and clock drawing on a digital device.About 36% of women reported concerns that only they noticed, while 11% reported corroborated concerns. Those who had cognitive concerns without peer corroboration showed lower mean scores than women without concerns. But they also scored higher than those with peer corroboration, indicating that self-perception of cognitive decline may precede external detection, Cosentino said.These simple cognitive tests could be incorporated into annual gynecologic checkups or primary care visits, Cosentino said.“The main take-home message is: Patient knows best,” said Jagger Kaye, a student researcher working with Cosentino. “In a clinical setting, keep in mind that if the patient reports concerns like this, it shouldn’t just be written off as something else — depression or anxiety or typical aging.” The speakers reported no relevant financial disclosures.Karen Blum is a freelance medical and science writer in the Baltimore area.