TL;DR
This study quantifies the health risks associated with combined exposure to PM2.5, NO2, and O3 across various urban microenvironments, revealing significant inter- and intra-individual variability. The findings highlight the limitations of using fixed-site monitors as exposure surrogates and emphasize the importance of personal exposure measurements for accurate health risk assessment and targeted intervention strategies.
What Problem Does This Paper Address?
Estimating the health impacts of air pollution typically relies on data from fixed-site monitors (FSMs) or residential locations. However, this approach overlooks the fact that individuals spend varying amounts of time in different microenvironments (home, office, outdoors, transit), each with distinct pollutant concentrations. This variability in exposure, coupled with differences in individual activity patterns and pollutant toxicity, leads to significant differences in personal health risks. The paper addresses the critical gap in understanding the combined health risks from multiple pollutants across these diverse microenvironments and quantifies the inter- and intra-individual variability in these risks, which is crucial for developing effective public health interventions.
Methods
The researchers conducted a longitudinal study involving 21 participants who wore portable air quality monitoring devices for multiple consecutive days. These devices measured PM2.5, NO2, and O3 concentrations at a 1-minute resolution across various microenvironments. Participants also logged their time spent in each location using a smartphone app.
Key methodological aspects include:
- Personal Exposure Measurement: Direct measurement of pollutant concentrations experienced by individuals in real-time across different microenvironments.
- Time-Location Tracking: Detailed recording of participants' time spent in each microenvironment to link exposure with location.
- Added Health Risk (AR) Model: An epidemiological model was adapted to quantify the health risk associated with a mixture of PM2.5, NO2, and O3. This model accounts for the different toxicities of each pollutant.
- Time-Integrated Health Risk (TIAR): The AR was integrated over time spent in each microenvironment to represent the cumulative daily health risk.
- Exposure Error Quantification: The difference between personal exposure concentrations and ambient concentrations (from FSMs and residence locations) was calculated to assess the accuracy of surrogate exposure data.
- Variance Component Analysis: Linear mixed-effects models were used to quantify the proportions of inter-individual (differences between people) and intra-individual (daily fluctuations within a person) variability in TIAR.
- Correlation Analysis: Pearson correlation was used to identify key contributors to both inter- and intra-individual variability in TIAR.
Datasets
The study utilized personal exposure data collected from 21 participants over 106 person-days. This data was supplemented by:
- Ambient Air Quality Data: Hourly average concentrations of PM2.5, NO2, and O3 from Hong Kong Environmental Protection Department (HKEPD) fixed-site monitors (FSMs).
- Residence Outdoor Air Quality Data: Measurements from a portable outdoor air quality system placed at the residences of seven participants.
- Participant Demographic and Residential Data: Collected via a web-based questionnaire.
Key Insights and Ideas
- Surrogates Fall Short: Using FSM or residence ambient air quality data as a proxy for personal exposure leads to substantial errors, particularly for NO2 and O3, due to their high spatial variability and complex atmospheric interactions (like O3 titration by NO).
- Microenvironments Matter: Different microenvironments contribute disproportionately to overall daily health risk. While participants spend most of their time indoors, "others indoor," "outdoor," and "transit" environments, despite shorter durations, can have a higher health risk intensity due to elevated pollutant concentrations.
- Inter- and Intra-Individual Variability are Crucial: Both differences between individuals and daily fluctuations within an individual's exposure are significant contributors to overall health risk variability. For PM2.5, intra-individual variability dominates, suggesting personal activity is key. For gaseous pollutants (NO2, O3), inter-individual variability is more prominent, highlighting the importance of location. For the combined risk, both are equally important.
- Home Indoor is a Major Contributor: Despite lower pollutant concentrations compared to some other microenvironments, the "home indoor" environment contributes significantly to the daily time-integrated health risk due to the substantial amount of time individuals spend there.
- Targeted Interventions: Identifying key contributors to variability (e.g., PM2.5, NO2, and O3 at home; O3 at the office; ambient PM2.5) allows for the development of more precise interventions to reduce health risks.
Key Results
- Exposure Error: Mean exposure error for PM2.5 was 39-45%, and for NO2 was 70-87%, when using FSM or residence data as surrogates.
- TIAR Variability: The daily time-integrated health risk for combined PM2.5, NO2, and O3 (TIARcombine) varied by a factor of 2.5 among participants and up to 2.5 for a given person across measured days.
- Variability Proportions: For PM2.5, intra-individual variability accounted for 78% of the total variance, while for NO2 and O3, inter- and intra-individual variabilities were more comparable (54% and 46% for NO2, 40% and 60% for O3, respectively). For TIARcombine, inter- and intra-individual variability were almost equally important (53% and 47%).
- Key Correlates of TIARcombine:
- Inter-individual: PM2.5, NO2, and O3 at 'home indoor'; O3 at 'office indoor'; and ambient PM2.5.
- Intra-individual: O3 at 'home indoor' showed the strongest correlation (r=0.76), followed by PM2.5 and NO2 at 'home indoor', NO2 in 'others indoor' and 'transit', and O3 at 'office indoor'.
Limitations and Future Work
The authors acknowledge several limitations:
- Small Sample Size: The study involved only 21 participants due to the cost of equipment and the burden of carrying the portable sensors. This limits the generalizability of the findings to broader populations, particularly those not affiliated with a university.
- Participant Demographics: The study population was limited to university-affiliated individuals, potentially excluding other demographic groups (e.g., housewives, unemployed, retired, or those in specific industrial settings).
Future work could focus on:
- Larger Scale Studies: Expanding the study to include a larger and more diverse participant pool to improve generalizability.
- Broader Microenvironment Coverage: Investigating additional microenvironments relevant to different occupational groups or lifestyles.
- Long-Term Health Outcomes: Linking personal exposure data and health risk assessments to actual long-term health outcomes.
- Intervention Effectiveness: Evaluating the effectiveness of specific interventions suggested by the study's findings.
- Methodological Refinement: Exploring more advanced statistical models to further disentangle the complex interplay of factors influencing personal exposure and health risk.
Summary updated