View Abstract

A4982 - Revon Smart Symptom Tracker: A Mobile Diagnostic and Triage Application for Reducing COPD Exacerbations
Author Block: S. Swaminathan1, A. N. Gerber2, K. Qirko1, T. Smith1, E. Corcoran3, N. Wysham4; 1Revon Systems Inc, Crestwood, KY, United States, 2Medicine, National Jewish Health, Denver, CO, United States, 3Pulmonology, Kaiser Permanente, Portland, OR, United States, 4Pulmonary and Critical Care, The Vancouver Clinic, Vancouver, WA, United States.
RATIONALE
COPD (Chronic Obstructive Pulmonary Disease) imposes a significant burden on patients’ daily lives. Flare-ups of this condition (exacerbations) are a frequent trigger of physician and hospital visits, which are costly and distressing to patients, and are associated with long-term decline in respiratory function and health. The need for novel solutions that limit the impact of exacerbations on patient health is abundantly apparent. One emerging approach to addressing COPD exacerbation is early detection by way of mobile app technology. Many of these apps, however, employ rule-based decision frameworks, which struggle to capture the size and complexity of the variable space relevant to triage and diagnosis. In this study, we developed the Revon Smart Symptom Tracker (RSST), which is a machine learning-based mobile application for early-detection of COPD exacerbations and patient triage.
METHODS
The RSST is trained on the opinion of 6 pulmonologists each triaging 750 computer generated patient cases. Each case is simulated as a distinct combination of variables, optimized to cover a statistically comprehensive set of plausible and clinically relevant scenarios based on literature review and physician consultation. The algorithm outputs 1) an exacerbation diagnosis, and 2) a triage recommendation from four choices (no action, continue usual treatment and return in 1-2 days, call MD, and go to ER). Initial validation of the RSST’s accuracy and safety is done by comparison to a physician panel consensus. The RSST is enabled for clinical use with IOS and Android phones. Further outcome studies on patient exacerbations, anxiety, quality of life, and user experience are ongoing in a clinical trial at The Vancouver Clinic (TVC).
RESULTS
Algorithm performance was assessed on the extent of concordance with the majority opinion of a 9-pulmonologist panel triaging a 101 clinically representative validation cases. The RSST correctly assigned the exacerbation and triage classes with accuracies of 97% and 89% respectively, better than all 9 MDs participating in all classification categories. The RSST further demonstrated superior accuracy and sensitivity in identifying COPD scenarios requiring emergency care. Outcome data from TVC indicates a positive influence of the RSST on patient symptom escalations.
CONCLUSION
The RSST exhibits comparable or better performance than physicians in identifying exacerbations and making an associated recommendation. Preliminary data also suggests that the RSST could be therapeutically effective in relevant COPD populations. Future iterations will incorporate the user-generated data from the current clinical trials in further refining triage decisions.