Can smartphone data identify the local environmental drivers of respiratory disease?
Jones, PJ and Koolhof, IS and Wheeler, AJ and Williamson, GJ and Lucani, C and Campbell, SL and Bowman, DMJS and Johnston, FH, Can smartphone data identify the local environmental drivers of respiratory disease?, Environmental Research, 182 Article 109118. ISSN 0013-9351 (2020) [Refereed Article]
Asthma and allergic rhinitis (or hay fever) are ubiquitous, chronic health conditions that seasonally affect a
sizeable proportion of the population. Both are commonly triggered or exacerbated by environmental conditions
including aeroallergens, air quality and weather. Smartphone technology offers new opportunities to identify
environmental drivers by allowing large-scale, real-time collection of day-to-day symptoms. As yet, however,
few studies have explored the potential of this technology to provide useful epidemiological data on environment-
symptom relationships. Here, we use data from the smartphone app ‘AirRater’ to examine relationships
between asthma and allergic rhinitis symptoms and weather, air quality and pollen loads in Hobart, Tasmania,
Australia. We draw on symptom data logged by app users over a three-year period and use time-series analysis to
assess the relationship between symptoms and environmental co-variates. Symptoms are associated with particulate
matter (IRR 1.06, 95% CI: 1.04–1.08), maximum temperature (IRR 1.28, 95% CI: 1.13–1.44) and pollen
taxa including Betula (IRR 1.04, 95% CI: 1.02–1.07), Cupressaceae (IRR 1.02, 95% CI: 1.01–1.04), Myrtaceae
(IRR 1.06, 95% CI: 1.02–1.10) and Poaceae (IRR 1.05, 95% CI: 1.01–1.09). The importance of these pollen taxa
varies seasonally and more taxa are associated with allergic rhinitis (eye/nose) than asthma (lung) symptoms.
Our results are congruent with established epidemiological evidence, while providing important local insights
including the association between symptoms and Myrtaceae pollen. We conclude that smartphone-sourced data
can be a useful tool in environmental epidemiology.