Educational content on VJDementia is intended for healthcare professionals only. By visiting this website and accessing this information you confirm that you are a healthcare professional.

Share this video  

AD/PD 2022 | Deep-learning tool can screen electronic health records to identify undiagnosed cognitive impairment

Sudeshna Das, PhD, Massachusetts General Hospital, Harvard Medical School, Boston, MA, talks on work aiming to develop a deep learning tool capable of screening electronic health record (EHR) data for signs of cognitive impairment. Information regarding cognition is often found in unstructured clinician notes and may therefore be overlooked by automatic tools assessing structured data. A deep learning based natural language processing algorithm was trained on an initial seed set and applied to these unstructured data. The model’s performance compared to that of a tool using structured data. The deep learning model increased the sensitivity of dementia detection compared to the baseline model. Over 16% of patients predicted as positive did not have a dementia diagnosis, suggesting automatic processing of EHR could identify patients with cognitive impairment who could benefit from further evaluation. This interview took place at the AD/PD™ 2022 Conference in Barcelona, Spain.