Many adults over the age of 65 complain of changes in memory or cognitive functioning. Typically the symptoms are mild, with nearly one out of every five people over 65 reporting symptoms of what we call “mild cognitive impairment” — slight changes in memory, language and cognitive functioning. While some people with mild cognitive impairment will go on to develop dementia, the majority do not.
At the present time, we cannot reliably predict who will go on to develop dementia, and typically by the time we confirm the diagnosis, it is often fairly late in the progression of the illness. If we were better able to predict which people would go on to develop dementia, including Alzheimer’s disease, we might be able to develop and initiate interventions that either prevent or slow down the progression of the illness.
Researchers from the Center for Quantitative Health at MGH, including Dr. Roy Perlis, the Director of the Center, and Dr. Thomas McCoy, the Director of Research, are now exploring how the electronic health record (EHR) may be used to facilitate the early identification of patients at risk for dementia. In a recent study, longitudinal electronic health records from two large academic medical centers were analyzed using a natural language processing tool. Natural language processing or NLP can be used to extract data from the narrative portion of the EHR which may yield information that is not reflected in the diagnostic or treatment codes documented in the EHR. In this analysis, NLP was used to estimate cognitive symptomatology at the time of hospitalization. Survival analysis was then used to examine the association of identified cognitive symptoms with subsequent diagnosis of dementia.
This retrospective analysis included a total of 267,855 hospitalized patients with 1,251,858 patient years of follow‐up data; 6516 (2.4%) of the patients ultimately received a diagnosis of dementia. In both younger and older patients, increasing cognitive symptom scores were associated with earlier diagnosis of dementia (HR 1.63; 1.54–1.72). The higher the cognitive symptom score at the time of discharge, the earlier the subsequent diagnosis of dementia. Looking at the individuals with cognitive symptom scores in the highest quartile at discharge, they were 7.4 times more likely to receive a diagnosis of dementia after five years than individuals with cognitive scores in the lowest quartile.
Based on the findings of this study, the researchers concluded that natural language processing of discharge summaries could be used to dementia risk up to eight years before diagnosis. By combining this technique with information regarding family history, risk factors, and biomarkers, we might be able to further refine this tool.
McCoy TH Jr, Han L, Pellegrini AM, Tanzi RE, Berretta S, Perlis RH. Stratifying risk for dementia onset using large-scale electronic health record data: A retrospective cohort study. Alzheimers Dement. 2020 Mar;16(3):531-540.
