Daily digital biomarkers in the follow-up and clustering of patients with asthma

Background and Research question We aimed to assess whether levels of digital biomarkers can reflect monthly patterns of asthma controlStudy design and methods We performed a longitudinal study on patients with asthma and comorbid rhinitis who filled ≥26 days of data in a month in the MASK-air® app and who reported at least 1 day of treatment with an inhaled corticosteroid with or without a long-acting β2-agonist (ICS ± LABA). We applied k-means cluster analysis to define clusters of months according to daily asthma control and medication use. Clusters were compared using digital biomarkers (visual analogue scale [VAS] on asthma symptoms and electronic daily asthma control score [e-DASTHMA]). We compared patients who did not switch with patients who switched their ICS ± LABA.Results We assessed 243 patients and 1358 months. We identified three clusters of poor asthma control despite high ICS ± LABA adherence, one cluster of poor asthma control and poor ICS ± LABA adherence, one cluster of good asthma control and high ICS ± LABA adherence and one cluster of good asthma control despite poor ICS ± LABA adherence. These clusters displayed relevant differences in VAS asthma and e-DASTHMA levels. Similar clusters were found in ‘non-switchers’ versus ‘switchers’.Conclusion Levels of digital biomarkers reflect asthma control patterns and might be used to monitor patients with asthma.