055 The digital doctor: a fully automated stratification and monitoring system for patients with memory complaints. Issue 12 (14th November 2019)

• Record Type: Journal Article
• Authors: O'Malley, Ronan; Mirhedari, Bahman; Harkness, Kirsty; Reuber, Markus; Venneri, Annalena; Christensen, Heidi; Blackburn, Daniel
• Is Part Of: Journal of neurology, neurosurgery and psychiatry. Volume 90:Issue 12(2019)
• Page Start: A23
• Page End: A23
• Publication Date: 2019-11-14
• Subjects: Neurology -- Periodicals; Nervous system -- Surgery -- Periodicals; Psychiatry -- Periodicals; 616.8
• DOI: 10.1136/jnnp-2019-ABN-2.76
• Languages: English
• ISSNs: 0022-3050

Abstract

Abstract : Introduction: Referrals to specialist memory clinics have increased out of proportion to the incidence of dementia. Time and financial pressures are consequently exerted on a service striving to deliver high quality care. We have developed a fully automated 'Digital Doctor' with the aim of providing pre-clinic risk stratification and ongoing monitoring for patients with memory concerns. Methods: We recruited 15 participants with Functional Memory Disorder (FMD), Mild Cognitive Impairment (MCI) and Alzheimer's disease each as well as 15 healthy controls. Participants answered 12 questions posed by the 'Digital Doctor'. Audio and visual data is analysed using diarization and automatic speech recognition tools and machine learning classifiers. Results: The 'Digital Doctor' can distinguish between neuro-degenerative dementia and FMD with an accuracy of 95%. We will have results of a 4-way classification accuracy (HC, FMD, MCI & AD) at time of conference. Discussion: We demonstrate the potential value of the 'Digital Doctor' as a stratification and triage tool. Accuracy will be improved with greater number of users and inclusion of fluency and picture description data. Patients at low risk could avoid the burden of a clinic appointment, whilst patients with higher risk could benefit from a more streamlined service.