Online crowd surveillance has been recently used as a means of tracking emergent risk to public health. Most commonly, these efforts involve the collection of online search queries and social media interactions to document acute changes in incidence or symptoms occurrence to primarily infectious disease agents, such as influenza. These projects have provided public health and medical professionals with insights into the spatial and temporal dynamics associated with these disease processes. Urban air pollution provides a key test case for the evaluation of online surveillance approaches for non-infectious environmental risks. Recent estimates attribute over 3 million excess mortalities to ambient particulate matter and ozone, collectively placing these pollutants among the most substantial contributors to global disease burden. Despite this, it is likely that traditional air pollution exposure indicators, namely ambient monitoring site measurements, at best, provide adequate estimates of biologically-relevant exposure, limited in their ability to reflect exposures for individuals not living near pollutant monitoring sites. To address the shortcomings of existing methods, we propose to develop and validate a Crowd-Generated indicator of Air Pollution exposures and health response, or ?CGAP? (Aim1a). The specific CGAP indicator for our case study, ozone (O3)(CGAP-O3) will be constructed using trends in online search behavior in Google Search terms and Social Media (Twitter) postings, validated and calibrated against O3 ground monitoring sites in Atlanta, GA (Aim 1b). We will also construct an online visualization front- end to enable interactive exploration and analysis of daily CGAP-O3 intensities, to enhance online accessibility for individuals interested in the risks associated with daily exposure to this pollutant. Finally, we will compare results from epidemiologic models examining time series associations between traditionally-monitored O3 and daily counts of asthma-related emergency department (ED) visits with those from models using the crowd-generated ozone exposure metric (CGAP-O3) (Aim 2). To establish the initial proof-of-principle, this analysis focuses on the geographic domain of metropolitan Atlanta. Epidemiologic associations between O3 and asthma ED visits have been previously shown to be consistently positive and significant associations in our Atlanta studies. The key benefits from this initiative include the development of a more temporally-sensitive, individual-level indicator of acute response associated with O3 exposures; with the potential of reflecting sub-clinical responses not captured in traditional acute health indicators, such as emergency department visits. Moreover, we anticipate that the proposed CGAP metric will provide greater spatial coverage for locales without traditional stationary air pollution monitoring.
We propose to develop innovative informatics techniques for effective tracking of exposure to urban air pollution by adapting and extending the state-of-the-art in computational text understanding and social data mining methods. The resulting techniques have the potential to improve scientific and public understandings of air pollution health effects, while enhancing engaged decision-making to better manage societal and individual risks associated with this environmental driver of global disease burden.
