Code-Based Algorithms for Identifying Dementia in Electronic Health Records: Bridging the Gap Between Theory and Practice

Article type: Article Commentary

Authors: Chen, Shanquan^{a; *} | Wang, Yuqi^b | Mueller, Christoph^{c; d}

Affiliations: [a] International Centre for Evidence in Disability, London School of Hygiene & Tropical Medicine, London, UK | [b] Department of Computer Science, University College London, London, UK | [c] King’s College London, London, UK | [d] South London and Maudsley NHS Foundation Trust, London, UK

Correspondence: [*] Correspondence to: Shanquan Chen, PhD, Assistant Professor, International Centre for Evidence in Disability, London School of Hygiene & Tropical Medicine, London, WC1E 7HT, UK. E-mail: shanquan.chen@lshtm.ac.uk.

Abstract: Code-based algorithms are crucial tools in the detection of dementia using electronic health record data, with broad applications in medical research and healthcare. Vassilaki et al.’s study explores the efficacy of code-based algorithms in dementia detection using electronic health record data, achieving approximately 70% sensitivity and positive predictive value. Despite the promising results, the algorithms fail to detect around 30% of dementia cases, highlighting challenges in distinguishing cognitive decline factors. The study emphasizes the need for algorithmic improvements and further exploration across diverse healthcare systems and populations, serving as a critical step toward bridging gaps in dementia care and understanding.

Keywords: Alzheimer’s disease, code-based algorithm, dementia, electronic health record

DOI: 10.3233/JAD-230887

Journal: Journal of Alzheimer's Disease, vol. 95, no. 3, pp. 941-943, 2023