NLP Unlocks Patterns In Myasthenia Gravis Progression

Identification Of Myasthenia Gravis Exacerbations, Crises & Symptom Burden Using Rules-Based Natural Language Processing Applied To Neurologist Clinical Notes 

This poster was originally presented at Myasthenia Gravis Foundation of America (MGFA) American Association of Neuromuscular & Electrodiagnostic Medicine (AANEM) 2023, November 1-4, in Phoenix, AZ. 

Authors: Jonathan Darer, MD, MPH; Jacqueline Pesa, PhD, MPH; Zia Choudhry, MD; Alberto Batista, PharmD, MS; Purva Parab, PhD; Xiaoyun Yang, MS; Raghav Govindarajan, MD 

Affiliations: Health Analytics; Janssen Scientific Affairs, LLC, Titusville, NJ, USA; Janssen Pharmaceutical Companies of Johnson & Johnson; Hospital Sisters Health Systems Medical Group 

Introduction 

• Myasthenia gravis (MG) is a rare, chronic, autoantibody-mediated neuromuscular disorder affecting approximately 60,000 individuals in the United States. The goal of this study is to characterize the identification, clinical presentation, and management challenges related to individuals with MG crisis based upon chart review of neurologist clinical progress notes. 

Methods  

• Medical transcriptions from AI & NLP-powered Amplity AnswerY™ (formerly Amplity Insights™) were used, including the full text from dictated HIPAA-compliant clinical notes covering 150,000 physicians from over 40,000 clinics and hospitals in the United States between 2017 and 2022.  

• Notes from non-neurologists and notes with non-standard formatting were excluded from the analysis. 

Results 

• As shown in Table 1, after applying rules to the medical transcription data, 5183 notes were available for analysis. 

  According to Table 2: 

• Ptosis (34.7%), diplopia (30.0%), dysphagia (23.0%), and fatigue (22.5%) were the most frequently reported MG symptoms. 

• A total of 3085 patients with MG were identified, with 12.7% experiencing clinical deterioration events (e.g., exacerbations) and 4.3% experiencing MG crises. 

• MG clinical deterioration was defined as exacerbations, worsening, or crises, and was documented in 12.7%, 4.5%, and 4.3% of patients, respectively. 

• Table 3 displays the five categories of clinical deterioration, with examples of neurologist text extracted from the AnswerY database. 

Conclusion 

• Documentation of worsening symptomatology is common in MG, highlighting the fluctuating nature of the disease. 

• Prevalence of MG symptoms in this real-world, cross-sectional analysis of physician notes is lower than in longitudinal studies based upon patient questionnaires. Clinical documentation may focus upon symptoms of greatest importance to the practicing physician at the time of the encounter. 

• Neurologist documentation of MG deterioration includes a spectrum of disease severity. Qualitative expert review of real-world neurologist documentation yielded five categories of clinical deterioration ranging from mild daily fluctuation to deterioration requiring hospitalization or rescue therapies, to MG crisis requiring mechanical ventilation. 

DOWNLOAD POSTER 

AnswerY adds powerful context based on HIPAA-compliant patient-provider conversations. See how in other research topics, including disparities in healthcare access, GLP-1s, and biomarker testing in NSCLC.