Home-based Digital Assessments with Applied Sentiment & Emotion AI Capture Improved Quality-of-life in Asthma Patients

With the rise of digital transformation in the pharmaceutical industry, digital therapeutics are being integrated in drug development clinical trials. In the TWINKLE study, information about asthmatic patients’ disease control and quality-of-life (QoL) was measured by daily video recording, in conjunction with daily electronic questionnaires and home-based spirometry. From the video messages, sentiment and emotion AI was applied to detect subtle QoL changes in asthmatic patients after receiving treatments. Sentiment scores, derived from patients’ daily messages via natural language processing, correlated strongly with metrics of lung functions and outcomes of electronic questionnaires. However, video-derived emotional analysis exhibited strong interpersonal variations and systematic biases, yet still showed utility in detecting QoL changes after personalized calibration and signal aggregation. Compared to traditional patient-reported outcomes, all three categories of digital measurements were able to detect significantly improved asthma control from patients who responded to treatments. The result provides insights into developing novel digital outcomes through the application of connected digital devices and advanced AI tailored to clinical settings.Clinical relevance— Digital outcomes involving connected digital devices and AI for sentiment/emotion analysis could capture subtle QoL changes reliably and earlier than hospital visits, reducing burden and improving disease management. Integrating digital therapeutics in asthma drug development trials may prove to be feasible and valuable.