Citizen Science, Innovative Technology: What’s Next?
Oral Presentation
Prepared by M. Barrett
Propeller Health, 634 West Main Street, Suite 102 , Madison, WI, 53703, United States
Contact Information: meredith.barrett@propellerhealth.com; 415-409-9258
ABSTRACT
Application of wireless inhaler sensors to enhance asthma surveillance and inform municipal interventions
Background
Louisville, Kentucky, ranks in the top 20 most challenging places to live with asthma in the US. Local asthma surveillance activities, which rely upon hospitalization reports and national survey prevalence data, do not provide real-time, spatially-explicit information that city leaders need to target the most effective interventions. The AIR Louisville programs aims to bring diverse local stakeholders, including citizens, government agencies, health systems, employers and community groups to address asthma in Jefferson County. The program works to 1) improve asthma outcomes for local citizens; 2) identify hotspots of asthma symptoms; 3) evaluate associations between asthma symptoms and environmental covariates in real-time and space; and ultimately inform municipal intervention scenarios.
Methods
Citizens (n=140) were enrolled through community locations and tracked their rescue inhaler use for 13 months with a GPS-enabled sensor, which passively collected the date, time and GPS location of each inhaler use. Using GIS and spatial tools, we mapped population-level asthma hotspots and extracted georeferenced environmental, climatic and socioeconomic data for each inhaler event. We modeled daily rescue inhaler use count using space-time resolved covariates through a negative binomial (NB) modeling technique. We developed full negative binomial models that controlled for confounding variables.
Results
Sensors collected 5,430 inhaler events, which were merged with 40 socioeconomic and environmental layers. Daily average rescue inhaler use among citizens decreased significantly over the program after enrollment (from an average 1.2 uses/day in the first week to 0.14 in the 25th week; 88% decrease; p<0.001), and asthma control improved. In the multivariate NB models, significant positive associations with asthma inhaler use were identified with PM10, Air Quality Index, grass pollen, weed pollen, mold, and crime (p< 0.01). Significant negative associations were identified with tree pollen, neighborhood income and property value (p< 0.01). These models will be used to test intervention scenarios by predicting their potential impact on asthma activity.
Discussion/conclusion
Utilizing sensors to capture the spatiotemporal signal of asthma can complement existing surveillance, improve understanding of environmental drivers, and help cities target interventions in neighborhoods where they will have the most impact. This pilot study has expanded and enrolled over 1000 participants, generating the largest citizen science asthma dataset ever collected.
Abstract for Overall Plenary Session:
There are budding scientists around us no matter where we might be every day. Technology is allowing us to do things once unimaginable. For the older generation, the information available at their fingertips is overwhelming. For the younger generation, science is part of their everyday lives – not just an area of interest in which you might chose to have a career. The larger population of individuals considered ‘in the middle’ ask more questions, dig deeper than most once did, and expect more information about their personal environment and health.
Citizen science and crowdsourcing are approaches that educate, engage, and empower the public to apply their curiosity and talents to a wide range of real-world problems with programs popping-up all over the country to raise greater awareness. We, as scientists, should ensure the general public is aware of tools and programs and encourage them to take advantage of using them. The vast amounts of data being generated will be in our hands to appropriately find further use to advance science, and protect the environment and public health.
The first two presentations will provide a high-level, overarching prospective on citizen science, technology, and crowdsourcing communities. Afterward, there will be a three person panel of experts working on crowdsourcing, citizen communities, and health initiatives that will be presented as examples with a short time for questions and discussion to follow.
Overarching (High Level) Citizen Science, Technology, and Crowdsourcing Communities
8:15 Lea Shanley, University of North Carolina - Chapel Hill
9:00 Waleed Abdalati, University of Colorado
Panel on Crowdsourcing/Citizen Communities and Health
10:15 Kathleen Weathers, Cary Institute of Ecosystem Studies
10:45 Meredith Barrett, Propeller Health
11:15 Sally Okun, Patients Like Me
11:45 General Discussion/ Questions
Oral Presentation
Prepared by M. Barrett
Propeller Health, 634 West Main Street, Suite 102 , Madison, WI, 53703, United States
Contact Information: meredith.barrett@propellerhealth.com; 415-409-9258
ABSTRACT
Application of wireless inhaler sensors to enhance asthma surveillance and inform municipal interventions
Background
Louisville, Kentucky, ranks in the top 20 most challenging places to live with asthma in the US. Local asthma surveillance activities, which rely upon hospitalization reports and national survey prevalence data, do not provide real-time, spatially-explicit information that city leaders need to target the most effective interventions. The AIR Louisville programs aims to bring diverse local stakeholders, including citizens, government agencies, health systems, employers and community groups to address asthma in Jefferson County. The program works to 1) improve asthma outcomes for local citizens; 2) identify hotspots of asthma symptoms; 3) evaluate associations between asthma symptoms and environmental covariates in real-time and space; and ultimately inform municipal intervention scenarios.
Methods
Citizens (n=140) were enrolled through community locations and tracked their rescue inhaler use for 13 months with a GPS-enabled sensor, which passively collected the date, time and GPS location of each inhaler use. Using GIS and spatial tools, we mapped population-level asthma hotspots and extracted georeferenced environmental, climatic and socioeconomic data for each inhaler event. We modeled daily rescue inhaler use count using space-time resolved covariates through a negative binomial (NB) modeling technique. We developed full negative binomial models that controlled for confounding variables.
Results
Sensors collected 5,430 inhaler events, which were merged with 40 socioeconomic and environmental layers. Daily average rescue inhaler use among citizens decreased significantly over the program after enrollment (from an average 1.2 uses/day in the first week to 0.14 in the 25th week; 88% decrease; p<0.001), and asthma control improved. In the multivariate NB models, significant positive associations with asthma inhaler use were identified with PM10, Air Quality Index, grass pollen, weed pollen, mold, and crime (p< 0.01). Significant negative associations were identified with tree pollen, neighborhood income and property value (p< 0.01). These models will be used to test intervention scenarios by predicting their potential impact on asthma activity.
Discussion/conclusion
Utilizing sensors to capture the spatiotemporal signal of asthma can complement existing surveillance, improve understanding of environmental drivers, and help cities target interventions in neighborhoods where they will have the most impact. This pilot study has expanded and enrolled over 1000 participants, generating the largest citizen science asthma dataset ever collected.
Abstract for Overall Plenary Session:
There are budding scientists around us no matter where we might be every day. Technology is allowing us to do things once unimaginable. For the older generation, the information available at their fingertips is overwhelming. For the younger generation, science is part of their everyday lives – not just an area of interest in which you might chose to have a career. The larger population of individuals considered ‘in the middle’ ask more questions, dig deeper than most once did, and expect more information about their personal environment and health.
Citizen science and crowdsourcing are approaches that educate, engage, and empower the public to apply their curiosity and talents to a wide range of real-world problems with programs popping-up all over the country to raise greater awareness. We, as scientists, should ensure the general public is aware of tools and programs and encourage them to take advantage of using them. The vast amounts of data being generated will be in our hands to appropriately find further use to advance science, and protect the environment and public health.
The first two presentations will provide a high-level, overarching prospective on citizen science, technology, and crowdsourcing communities. Afterward, there will be a three person panel of experts working on crowdsourcing, citizen communities, and health initiatives that will be presented as examples with a short time for questions and discussion to follow.
Overarching (High Level) Citizen Science, Technology, and Crowdsourcing Communities
8:15 Lea Shanley, University of North Carolina - Chapel Hill
9:00 Waleed Abdalati, University of Colorado
Panel on Crowdsourcing/Citizen Communities and Health
10:15 Kathleen Weathers, Cary Institute of Ecosystem Studies
10:45 Meredith Barrett, Propeller Health
11:15 Sally Okun, Patients Like Me
11:45 General Discussion/ Questions
