A101-23 Predictive Modeling of Respiratory Exacerbations Using Remote Patient Monitoring Data

Rationale Early detection of respiratory exacerbations in patients with chronic respiratory diseases remains a critical unmet need. We hypothesized that integrating remote patient monitoring (RPM) data models with clinical and demographic factors could improve early risk stratification for exacerbations. Methods RPM data were collected from 50 patients (age range 13-86 yrs; mean 59.6 yrs; 54 % female) over three months, generating >1,000 daily measurements per patient. Variables included spirometry (FEV1, FVC, peak flow), oxygen saturation, demographic data, biologics use, and insurance type. Exacerbations were defined by clinically significant drops in FEV1, FVC, or SpO2, corroborated by alert counts. Logistic regression and k-nearest neighbors (k-NN) models were developed to predict high-risk events. Model performance was evaluated via accuracy, sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV). Results The predictive model achieved 81 % overall accuracy, sensitivity 0.78, specificity 0.83, PPV 0.79, and NPV 0.82. Risk scores ranged from 0.16 to 0.91, and patients with ≥1 red alert had a 2.1-fold higher likelihood of adverse outcomes. Biologic use was associated with reduced exacerbation risk, while insurance type had no significant predictive impact. Conclusions AI-driven RPM models can effectively stratify respiratory patients by exacerbation risk, providing actionable insights for early clinical intervention. This approach has potential to reduce hospitalizations, optimize resource allocation, and personalize respiratory care. Future work will expand to larger cohorts and integrate additional biomarkers to enhance predictive accuracy. Keywords: Remote Patient Monitoring; Chronic Respiratory Disease; Respiratory Exacerbations; Machine Learning; Predictive Modeling; Risk Stratification; Biologic Therapy References: 1.Association, A. L. (n.d.). Our impact. American Lung Association Impact | American Lung Association. https://www.lung.org/about-us/our-impact 2.DeVrieze, B. W., Goldin, J., & Giwa, A. O. (2024, October 6). Peak Flow Rate Measurement. Nih.gov; StatPearls Publishing. https://www.ncbi.nlm.nih.gov/books/NBK459325/ 3.Kitch, B. T., A. David Paltiel, Kuntz, K. M., Dockery, D. W., Schouten, J. P., Weiss, S. T., & Fuhlbrigge, A. L. (2004). A Single Measure of FEV 1 Is Associated With Risk of Asthma Attacks in Long-term Follow-up. CHEST Journal, 126(6), 1875-1882. https://doi.org/10.1378/chest.126.6.1875 4.Nurmagambetov, T., Kuwahara, R., & Garbe, P. (2018). The Economic Burden of Asthma in the United States, 2008-2013. Annals of the American Thoracic Society, 15(3), 348-356. https://doi.org/10.1513/annalsats.201703-259oc This abstract is funded by: self funded