This New Investigator Award study uses portable, real-time sensors to assess misclassification associated with surrogate measures of exposure in a panel of participants that live close to, or far from, traffic and have electric or gas cook stoves. The focus is on ultrafine particles, PM2.5, black carbon, and NO2.
Abstract for the 2015 HEI Annual Conference
Validation and Use of Novel Sensors to Assess Human Exposures to Airborne Pollutants
Juana Maria Delgado-Saborit, Adobi Okam, and Maryam Shehab
Division of Environmental Health & Risk Management, School of Geography, Earth & Environmental Sciences, and College of Life and Environmental Sciences, University of Birmingham, Edgbaston, United Kingdom
Background Epidemiologic research on air pollution traditionally uses data collected in central site monitors as their primary data source. Central site data are known to lead to exposure misclassification, affecting the sensitivity of epidemiologic studies on air pollution effects and seriously biasing estimates of the health effect of a pollutant. Advances in technology have made commercially available a range of sensors with in-built accelerometers and high temporal resolution that may allow more accurate measurement of personal exposures and permit the estimation of lung intake.
Objectives The overall aim of the project is to (a) characterize lung intake of ultrafine particles (UFP), particulate matter with aerodynamic diameter less than 2.5 μm (PM2.5), and black carbon (BC); (b) assess the degree of misclassification associated with using surrogate measures; and (c) identify key activities and sources contributing the most to lung intake of these pollutants.
Methods Prior to using the sensors with subjects for measuring exposures, these have been validated in a set of on-line, off-line and laboratory experiments. During the on-line experiments, all sensors have been collocated alongside recognised reference methods in five monitoring sites. Off-line validation tests compared the integrated PM2.5 sensor readings against the concentrations gravimetrically determined on the sensor’s own downstream filter and on a second filter collected with a separate pump. Off-line microaethalometer validation compared the integrated mass of BC measured by the microaethalometer with elemental carbon and black smoke (BS) concentrations. The laboratory validation of the NO2 sensor aimed at calibrating the sensor readings using known NO2 concentrations generated in a calibration system. The UFP sensors have been validated in the laboratory by comparing the UFP concentrations and average diameter measured by the UFP sensor with those measured by a SMPS and a CPC after being challenged with DEHS particles generated in the lab. Forty subjects are being recruited to assess their exposure to PM2.5, UFP and BC according to a combination of two key determinants: exposure to traffic by home location and use of gas cooking appliances. Real time measurements are been collected over 4 days.
Results On-line tests results show good agreement between the DISCmini UFP sensor and the CPC (R2 0.88-0.98); the microaethalometer AE-51 with the reference aethalometer AE-22 (R2 0.81- 0.92), and the microPEM PM2.5 sensor with the FDMS-TEOM (0.73-0.90). The Aeroqual NO2 sensor could not be validated since the agreement between the sensor and the chemiluminiscence analyser was not constant. Microaethalometer off-line tests showed good correlation between BC and BS concentrations (R2 0.8-0.9). The laboratory intercomparison of the UFP sensors and the SMPS/CPC also showed good agreement (R2 0.83-0.99). However, the number concentration is generally overestimated in the presence of large particles (Dp>300 nm) and polydisperse aerosols. Preliminary analysis of BC exposure data shows that personal exposures are higher than home concentrations by a factor of 2. Commuting and cooking are identified as relevant activities contributing to exposures.
Acknowledgment Dr Delgado-Saborit counts with the advice of an Advisory Board formed by Professors Roy M Harrison, Jon G Ayres, Ross Anderson, and Ben Armstrong.
Poster by Delgado-Saborit et al, 2016 HEI Annual Conference