Leveraging Natural Language Processing and Clinical Notes for Dementia Detection

Early detection and automated classification of dementia has recently gained considerable attention using neuroimaging data and spontaneous speech. In this paper, we explore the problem of dementia detection with in-hospital clinical notes. We collected 954 patients' clinical notes from a local hospital in Melbourne and assign dementia/non-dementia labels to those patients based on clinical assessment and telephone interview. Given the labeled dementia data sets, we fine tune a ClinicalBioBERT using filtered clinical notes and conducted experiments on both binary and three class dementia classification. Our experiment results show that the fine tuned ClinicalBioBERT achieved satisfied performance on binary classification but performed poorly on three class dementia classification. We explore the difficulties we encountered applying ClinicalBioBERT to hospital text. Further analysis suggests that more human prior knowledge should be considered.