PreprintArticleVersion 1Preserved in Portico This version is not peer-reviewed
Digital Triage Tool Using Artificial Intelligence and Patient History for Detecting Selected Neurological Diseases and Sensing the Bottleneck between Symptoms, Diagnosis, and Therapy
Version 1
: Received: 12 March 2022 / Approved: 14 March 2022 / Online: 14 March 2022 (08:58:29 CET)
How to cite:
Grigull, L.; Lechner, W.; Klawonn, F. Digital Triage Tool Using Artificial Intelligence and Patient History for Detecting Selected Neurological Diseases and Sensing the Bottleneck between Symptoms, Diagnosis, and Therapy. Preprints.org2022, 2022030178. https://doi.org/10.20944/preprints202203.0178.v1.
Grigull, L.; Lechner, W.; Klawonn, F. Digital Triage Tool Using Artificial Intelligence and Patient History for Detecting Selected Neurological Diseases and Sensing the Bottleneck between Symptoms, Diagnosis, and Therapy. Preprints.org 2022, 2022030178. https://doi.org/10.20944/preprints202203.0178.v1.
Cite as:
Grigull, L.; Lechner, W.; Klawonn, F. Digital Triage Tool Using Artificial Intelligence and Patient History for Detecting Selected Neurological Diseases and Sensing the Bottleneck between Symptoms, Diagnosis, and Therapy. Preprints.org2022, 2022030178. https://doi.org/10.20944/preprints202203.0178.v1.
Grigull, L.; Lechner, W.; Klawonn, F. Digital Triage Tool Using Artificial Intelligence and Patient History for Detecting Selected Neurological Diseases and Sensing the Bottleneck between Symptoms, Diagnosis, and Therapy. Preprints.org 2022, 2022030178. https://doi.org/10.20944/preprints202203.0178.v1.
Abstract
During the COVID-19 pandemic, individuals with symptoms other than cough or fever have refrained from seeking medical advice. However, a delay in treatment might lead to serious consequences. At the same time, digital health initiatives have emerged to overcome this bottleneck of healthcare. Herein, we report the results of a multi-center initiative using a combination of patient history and artificial intelligence (AI) to identify individuals with rare neuromuscular diseases. First, a questionnaire with 46 items was developed by interviewing patients with muscular dystrophies, amyotrophic lateral sclerosis, Morbus Pompe, neuropathies, and myasthenia gravis. Second, patients with proven neurological diseases answered the questionnaire. Third, a combination of classifiers (artificial neural network, support vector, and random forest) was trained and, finally, the system was challenged with new questionnaires. Users with an abnormal questionnaire pattern received a unique code for data privacy and contact details for a neurologist for further advice. The neurologists confirmed or refuted the AI-based diagnosis. The questionnaire was accessed 3122 times, leading to 853 unique codes. Only for a few patients the computer-based diagnoses and the confirmed final diagnoses were reported to us. However, for these few patients, the genetic testing and high CK levels finally ended their long-lasting diagnostic odyssey.
Keywords
artificial intelligence; data mining; diagnostic decision support; rare diseases; questionnaire anamnesis; neuromuscular diseases; high latencies
Subject
Computer Science and Mathematics, Computational Mathematics
Copyright:
This is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
