TracMyAir: Smartphone-enabled spatiotemporal estimates for inhaled doses of particulate matter and ozone to personalize health outcomes

The study demonstrates the feasibility of the TracMyAir smartphone platform in generating personalized, hyperlocal estimates of inhaled particulate matter and ozone doses by integrating geolocation, diverse air quality data, and physiological metrics, thereby addressing key limitations of traditional fixed-site monitoring for environmental health research.

The Gist

Imagine trying to figure out how much "dirty air" you breathe in a day. For years, scientists have relied on big, stationary weather stations sitting on rooftops in city centers. It's like trying to guess how much rain you got wet by looking at a single rain gauge in the middle of a park, while you were actually running through a forest, hiding under an umbrella, and then sitting in a dry living room. That fixed gauge misses your movement, your shelter, and your specific activity.

TracMyAir is a new smartphone app designed to fix this problem. Think of it as a "personal air pollution detective" that rides in your pocket.

Here is how it works, broken down into simple concepts:

1. The Detective's Toolkit

Instead of just looking at one big map, the app combines several clues to build a picture of your personal air history:

• Your Location: It uses your phone's GPS to know exactly where you are (like a taxi driver tracking your route).
• The Air Report: It pulls data from official government sensors (the "big rain gauges") and a network of cheap, community sensors (like neighbors holding their own rain gauges).
• Your Shelter: It estimates how much pollution gets inside your house or office, kind of like calculating how much rain seeps through a roof.
• Your Movement: It uses your smartwatch or phone to see if you are walking, running, or sitting still.

2. The "Inhaled Dose" Calculation

The app doesn't just tell you the air quality outside; it calculates your "Inhaled Dose."

• The Analogy: Imagine pollution is like dust in a room. If you are sitting still, you breathe in a little dust. If you are sprinting, you are gulping down air (and dust) much faster.
• TracMyAir combines the amount of pollution in the air with how hard you are breathing. It creates a score that says, "Because you ran for 20 minutes in this specific neighborhood, you actually breathed in this much more pollution than someone sitting on a bench nearby."

3. What They Discovered

The researchers tested this on 18 people for over 1,500 hours. Here is what they found:

• Step Counting Works Best: When the app used your step count to guess how hard you were breathing, it was very accurate (like a perfect match). When it tried to guess based on your heart rate, it was a bit more wobbly, sometimes overestimating the dose.
• It's Personal: Two people living on the same street might breathe in very different amounts of pollution because one is jogging in the park and the other is cooking dinner indoors. The app captures these tiny, personal differences.
• Cheap Sensors Help: The app worked just as well using data from low-cost community sensors as it did with expensive government ones, especially when those sensors were close to where the people actually were.
• No Rich vs. Poor Pattern: Surprisingly, in this specific group, the level of pollution outside didn't strictly depend on whether the neighborhood was rich or poor. Pollution is everywhere, but how much you breathe depends on what you are doing.

The Big Picture

TracMyAir proves that we don't need expensive, lab-based equipment to track our personal exposure to bad air. By using the smartphone in your pocket and the smartwatch on your wrist, we can finally move from "guessing the average air quality for a city" to "knowing the exact air quality burden for you."

This is a game-changer for doctors and researchers. Instead of treating everyone the same, they can eventually say, "Because you breathe in this specific amount of pollution every day, you are at higher risk for this specific health issue," leading to truly personalized medicine.

Technical Summary

1. Problem Statement

The core challenge addressed in this research is the inaccuracy of current individual exposure assessments in environmental health. Traditional methods rely on fixed-site monitoring stations, which fail to capture critical variables affecting personal pollutant burden, including:

• Human Mobility: Individuals move through varying microenvironments throughout the day.
• Indoor Environments: Fixed outdoor sensors do not account for the filtration or accumulation of pollutants indoors.
• Physiological Variability: Exposure alone does not equal dose; the amount of pollutant actually inhaled depends on an individual's activity level and breathing rate.

Consequently, there is a lack of scalable, real-world tools capable of generating time-resolved, personalized estimates of inhaled dose (the actual amount of pollutant entering the body) rather than just ambient concentration.

2. Methodology

The authors developed and deployed TracMyAir, a smartphone-based digital health platform designed to integrate multiple data streams to model exposure and dose.

• Study Design: An exploratory study involving 18 adults contributing over 1,500 participant-hours of data under real-world conditions.
• Data Integration: The platform synthesized data from four primary sources:
  1. Geolocation: Smartphone GPS to track movement and location.
  2. Air Quality Data: A hybrid approach combining regulatory data (AirNow) and community-based low-cost sensor data (PurpleAir).
  3. Microenvironment Modeling: Algorithms to classify indoor vs. outdoor settings and apply building infiltration modeling to estimate how outdoor pollutants penetrate indoor spaces.
  4. Physiological Metrics: Wearable-derived data, specifically step counts and heart rate, to estimate ventilation rates.
• Computational Framework: The system computes eight tiers of hourly exposure estimates, progressing from raw ambient data to refined, individualized inhaled dose calculations for PM2.5 and Ozone (O3).

3. Key Contributions

• Scalable Digital Health Platform: Demonstrated the feasibility of using ubiquitous consumer devices (smartphones and smartwatches) to replace or augment expensive, stationary monitoring networks.
• Multi-Tiered Dose Modeling: Introduced a sophisticated framework that moves beyond simple exposure concentration to calculate inhaled dose, accounting for the "exposure-dose" gap caused by activity levels and indoor infiltration.
• Hybrid Sensor Integration: Validated the integration of dense, low-cost community sensor networks (PurpleAir) with regulatory data to improve spatial resolution.
• Physiological Personalization: Showed how wearable data can be used to dynamically adjust dose estimates based on real-time physical activity.

4. Key Results

• Dose Estimation Concordance:
  • Hourly dose estimates derived from step counts (smartphone/smartwatch) showed very high concordance (Spearman correlation p = 0.97–0.98).
  • Estimates based on heart rate showed greater variability and higher mean values, with slightly lower concordance (p = 0.82–0.92).
• Variance Explanation:
  • Exposure metrics explained 51–73% of the variance in PM2.5 inhaled dose.
  • Exposure metrics explained 68–84% of the variance in Ozone inhaled dose.
  • This indicates that physiological modeling significantly improves the accuracy of personal pollutant burden estimates.
• Sensor Network Correlation: Modeled doses based on regulatory networks and community sensors were strongly correlated (R = 0.84). Notably, community sensors were often located closer to participants, supporting their utility for hyperlocal assessment.
• Variability: The study observed substantial inter-individual (between people) and intra-individual (within the same person over time) variability, driven by dynamic transitions between microenvironments and changing activity patterns.
• Socioeconomic Analysis: No consistent association was found between outdoor pollutant levels and neighborhood socioeconomic status within this specific cohort.

5. Significance and Implications

The TracMyAir study demonstrates a paradigm shift in environmental epidemiology and precision medicine:

• Precision Environmental Medicine: By quantifying the actual inhaled dose rather than just ambient concentration, the platform enables more accurate links between environmental exposure and specific health outcomes.
• Epidemiological Utility: The ability to capture dynamic, hyperlocal exposure data allows for better study designs in population health, moving away from ecological fallacies associated with fixed-site monitoring.
• Scalability: The reliance on existing consumer hardware and open data networks suggests a cost-effective, scalable solution for monitoring personal air quality globally.
• Future Applications: The framework lays the groundwork for personalized health interventions, such as real-time alerts for individuals with respiratory conditions to avoid high-dose periods or microenvironments.