Investigating Application of a Clinical Algorithm on Real-World Electronic Medical Record Data for Earlier Detection of Amyotrophic Lateral Sclerosis

Topic:

Other

Poster Number: P287

Author(s):

Amer Ghavanini, MD, PhD, FRCP(C), CSCN(EMG), Trillium Health Partners, Toronto, ON, Canada; University of Toronto, Toronto, ON, Canada, Amanda Fiander, MD, FRCP(C), CSCN(EMG), Maritime Neurology, Halifax, NS, Canada, Dung Pham, PhD, Mitsubishi Tanabe Pharma Canada, Inc., Toronto, ON, Canada, Pinay Kainth, PhD, Mitsubishi Tanabe Pharma Canada, Inc., Toronto, ON, Canada, Angela Genge, MD, Clinical Research Unit, The Montreal Neurological Institute, Montreal, QC, Canada

Background: Amyotrophic lateral sclerosis (ALS) is a rare and progressive neurodegenerative disease that is difficult to diagnose. An estimated 3000 Canadians live with ALS and approximately 1000 are diagnosed annually. Delayed ALS diagnosis, reported to range from 9.1 to 27 months, can lead to mismanagement and deterioration of patient outcomes. To date, no single test definitively confirms ALS. Algorithm-based tools may be used to help physicians improve the timeliness and accuracy of ALS diagnosis.

Objective: To describe the study design of a clinical algorithm applied to neurologists’ electronic medical record (EMR) data to reduce the delay in time to diagnosis and treatment for patients with ALS.

Results: A clinical algorithm applied to neurologists’ EMR data categorizes patients into risk groupings based on evidence of upper and lower motor neuron abnormalities and spinal region involvement. EMR records of participating clinics are scanned, and the likelihood of ALS is estimated for patients whose records include a recent electromyography test. A report is sent to the clinic for each patient flagged for follow up. In an initial study, the algorithm demonstrated improved sensitivity and specificity of 93.9% and 98.0%, respectively, when compared to an original proof-of-concept study. This software has been registered as a class 1 medical device (Health Canada). An ongoing, 12-month pilot study aims to deploy this algorithm within 20 community clinics.

Conclusions: Clinical algorithms may aid in the diagnosis of ALS when applied to real-world EMR data in neurological community practice. The algorithm in this study has been proven to identify patients with an elevated risk of ALS to support expedited follow up, diagnosis, and treatment. This study aims to investigate additional details to improve the results of earlier algorithm testing and increase exposure to the clinical community.

Sponsorship: This study was sponsored by Mitsubishi Tanabe Pharma Canada, Inc.

Acknowledgments: The authors thank Christofer Baldwin, PhD, and Irene Brody, VMD, PhD, of p-value communications, Cedar Knolls, NJ, USA, for providing medical writing support. Editorial support was also provided by p-value communications. This support was funded by Mitsubishi Tanabe Pharma Canada, Inc., Toronto, ON, CA, in accordance with Good Publication Practice Guidelines 2022.