Dataset Information

Automated classification of time-activity-location patterns for improved estimation of personal exposure to air pollution.

ABSTRACT:

Background

Air pollution epidemiology has primarily relied on measurements from fixed outdoor air quality monitoring stations to derive population-scale exposure. Characterisation of individual time-activity-location patterns is critical for accurate estimations of personal exposure and dose because pollutant concentrations and inhalation rates vary significantly by location and activity.

Methods

We developed and evaluated an automated model to classify major exposure-related microenvironments (home, work, other static, in-transit) and separated them into indoor and outdoor locations, sleeping activity and five modes of transport (walking, cycling, car, bus, metro/train) with multidisciplinary methods from the fields of movement ecology and artificial intelligence. As input parameters, we used GPS coordinates, accelerometry, and noise, collected at 1 min intervals with a validated Personal Air quality Monitor (PAM) carried by 35 volunteers for one week each. The model classifications were then evaluated against manual time-activity logs kept by participants.

Results

Overall, the model performed reliably in classifying home, work, and other indoor microenvironments (F1-score>0.70) but only moderately well for sleeping and visits to outdoor microenvironments (F1-score=0.57 and 0.3 respectively). Random forest approaches performed very well in classifying modes of transport (F1-score>0.91). We found that the performance of the automated methods significantly surpassed those of manual logs.

Conclusions

Automated models for time-activity classification can markedly improve exposure metrics. Such models can be developed in many programming languages, and if well formulated can have general applicability in large-scale health studies, providing a comprehensive picture of environmental health risks during daily life with readily gathered parameters from smartphone technologies.

SUBMITTER: Chatzidiakou L

PROVIDER: S-EPMC9733291 | biostudies-literature | 2022 Dec

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Automated classification of time-activity-location patterns for improved estimation of personal exposure to air pollution.

Chatzidiakou Lia L Krause Anika A Kellaway Mike M Han Yiqun Y Li Yilin Y Martin Elizabeth E Kelly Frank J FJ Zhu Tong T Barratt Benjamin B Jones Roderic L RL

Environmental health : a global access science source 20221209 1

<h4>Background</h4>Air pollution epidemiology has primarily relied on measurements from fixed outdoor air quality monitoring stations to derive population-scale exposure. Characterisation of individual time-activity-location patterns is critical for accurate estimations of personal exposure and dose because pollutant concentrations and inhalation rates vary significantly by location and activity.<h4>Methods</h4>We developed and evaluated an automated model to classify major exposure-related micr ...[more]

PMID: 36482402

Similar Datasets

Project description:The era of big data has enabled sophisticated models to predict air pollution concentrations over space and time. Historically these models have been evaluated using overall metrics that measure how close predictions are to monitoring data. However, overall methods are not designed to distinguish error at timescales most relevant for epidemiologic studies, such as day-to-day errors that impact studies of short-term health associations. We introduce frequency band model performance, which quantifies health estimation capacity of air quality prediction models for time series studies of air pollution and health. Frequency band model performance uses a discrete Fourier transform to evaluate prediction models at timescales of interest. We simulated fine particulate matter (PM2.5), with errors at timescales varying from acute to seasonal, and health time series data. To compare evaluation approaches, we use correlations and root mean squared error (RMSE). Additionally, we assess health estimation capacity through bias and RMSE in estimated health associations. We apply frequency band model performance to PM2.5 predictions at 17 monitors in 8 US cities. In simulations, frequency band model performance rates predictions better (lower RMSE, higher correlation) when there is no error at a particular timescale (e.g., acute) and worse when error is added to that timescale, compared to overall approaches. Further, frequency band model performance is more strongly associated (R2 = 0.95) with health association bias compared to overall approaches (R2 = 0.57). For PM2.5 predictions in Salt Lake City, UT, frequency band model performance better identifies acute error that may impact estimated short-term health associations. For epidemiologic studies, frequency band model performance provides an improvement over existing approaches because it evaluates models at the timescale of interest and is more strongly associated with bias in estimated health associations. Evaluating prediction models at timescales relevant for health studies is critical to determining whether model error will impact estimated health associations.

Project description:ObjectiveAir pollution is a leading preventable risk factor for cardiovascular diseases. Previous studies mostly relied on concentrations at residence, which might not represent personal exposure. Personal air pollution exposure has a greater variability compared with levels of ambient air pollution, facilitating evaluation of exposure-response functions and vascular pathophysiology. We aimed to evaluate the association between predicted annual personal exposure to PM2.5 and black carbon (BC) and three vascular damage markers in peri-urban South India.MethodsWe analyzed the third wave of the APCAPS cohort (2010-2012), which recruited participants from 28 villages. We used predicted personal exposure to PM2.5 and BC derived from 610 participant-days of 24 h average gravimetric PM2.5 and BC measurements and predictors related to usual time-activity. Outcomes included carotid intima-media thickness (CIMT), carotid-femoral pulse wave velocity (cf-PWV) and augmentation index (AIx). We fit linear mixed models, adjusting for potential confounders and accounting for the clustered data structure. We evaluated nonlinear associations using generalized additive mixed models.ResultsOf the 3017 participants (mean age 38 years), 1453 (48%) were women. The average PM2.5 exposure was 51 µg/m3 (range 13-85) for men, and 61 µg/m3 (range 40-120) for women, while the average BC was 4 µg/m3 (range 3-7) for men and 8 µg/m3 (range 3-22) for women. A 10 μg/m3 increase of PM2.5 was positively associated with CIMT (0.026 mm, 95% CI 0.014, 0.037), cf-PWV (0.069 m/s, 95% CI 0.008, 0.131) and AIx (0.8%, 95% CI 0.3, 1.3) among men. The exposure-response function for PM2.5 and AIx among men showed non-linearity, particularly within the exposure range dominated by tobacco smoking and occupational exposures. Both PM2.5 and BC were positively associated with AIx among women (0.6%, 95% CI 0.2, 1.0, per 10 μg/m3 PM2.5; 0.5%, 95% CI 0.1, 0.8, per 2 μg/m3 BC).ConclusionsPersonal exposure to particulate matter was associated with vascular damage in a peri-urban population in South India. Personal exposure to particulate matter appears to have gender-specific effects on the type of vascular damage, potentially reflecting differences in sources of personal exposure by gender.

Project description:Mobile phones have a large spectrum of applications, aiding in risk prevention and improving health and wellbeing of their owners. So far, however, they have not been used for direct assessment of personal exposure to air pollution. In this study, we comprehensively evaluated the first, and the only available, mobile phone-BROAD Life-equipped with air pollution sensors (PM2.5 and VOC), to answer the question whether this technology is a viable option in the quest of reducing the burden of disease to air pollution. We tested its performance, applicability and suitability for the purpose by subjecting it to varied concentrations of different types of aerosol particles (cigarette smoke, petrol exhaust and concrete dust) and formaldehyde under controlled laboratory conditions, as well as to ambient particles during field measurements. Six reference instruments were used in the study: AEROTRAK Optical Particle Counter (OPC model number 9306), DustTrak, Aerodynamic Particle Counter (APS), Scanning Mobility Particle Sizer (SMPS), Tapered Element Oscillating Microbalance (TEOM) and Formaldehyde Analyser. Overall, we found that the phone's response was linear at higher particle number concentrations in the chamber, above 5 and 10 μg m-3, for combustion and concrete dust particles, respectively, and for higher formaldehyde concentrations, making it potentially suitable for applications in polluted environments. At lower ambient concentrations of particles around 10 ug m-3 and 20 μg m-3 for PM2.5 and PM10, respectively, the phone's response was below its noise level, suggesting that it is not suitable for ambient monitoring under relatively clean urban conditions. This mobile phone has a number of limitations that may hinder its use in personal exposure and for continuous monitoring. Despite these limitations, it may be used for comparative assessments, for example when comparing outcomes of intervention measures or local impacts of air pollution sources. It should be kept in mind, however, that a mobile phone measuring air quality alone cannot as such 'reduce the burden of disease to air pollution, as knowing ambient concentrations is only one of the building block in this quest. As long as individuals cannot avoid exposure e.g. in urban areas, knowing concentrations is not sufficient to reduce potential adverse effects. Yet, there are many situations and microenvironments, which individuals could avoid knowing the concentrations and also being aware of the risk caused by exposure to them. This includes for example to proximity to vehicle emissions, either for social purposes (e.g. street cafes) or exercising (e.g. walking or jogging along busy roads)or indoor environments affected by combustion emissions (smoking, candle burning, open fire).

Project description:BackgroundExposure to fine particulate matter (PM2.5) increases the risk of asthma exacerbations, and thus, monitoring personal exposure to PM2.5 may aid in disease self-management. Low-cost, portable air pollution sensors offer a convenient way to measure personal pollution exposure directly and may improve personalized monitoring compared with traditional methods that rely on stationary monitoring stations. We aimed to understand whether adults with asthma would be willing to use personal sensors to monitor their exposure to air pollution and to assess the feasibility of using sensors to measure real-time PM2.5 exposure.MethodsWe conducted semi-structured interviews with 15 adults with asthma to understand their willingness to use a personal pollution sensor and their privacy preferences with regard to sensor data. Student research assistants used HabitatMap AirBeam devices to take PM2.5 measurements at 1-s intervals while walking in Philadelphia neighborhoods in May-August 2018. AirBeam PM2.5 measurements were compared to concurrent measurements taken by three nearby regulatory monitors.ResultsAll interview participants stated that they would use a personal air pollution sensor, though the consensus was that devices should be small (watch- or palm-sized) and light. Patients were generally unconcerned about privacy or sharing their GPS location, with only two stating they would not share their GPS location under any circumstances. PM2.5 measurements were taken using AirBeam sensors on 34 walks that extended through five Philadelphia neighborhoods. The range of sensor PM2.5 measurements was 0.6-97.6 μg/mL (mean 6.8 μg/mL), compared to 0-22.6 μg/mL (mean 9.0 μg/mL) measured by nearby regulatory monitors. Compared to stationary measurements, which were only available as 1-h integrated averages at discrete monitoring sites, sensor measurements permitted characterization of fine-scale fluctuations in PM2.5 levels over time and space.ConclusionsPatients were generally interested in using sensors to monitor their personal exposure to PM2.5 and willing to share personal sensor data with health care providers and researchers. Compared to traditional methods of personal exposure assessment, sensors captured personalized air quality information at higher spatiotemporal resolution. Improvements to currently available sensors, including more reliable Bluetooth connectivity, increased portability, and longer battery life would facilitate their use in a general patient population.

Dataset Information

Automated classification of time-activity-location patterns for improved estimation of personal exposure to air pollution.

Background

Methods

Results

Conclusions

Publications

Automated classification of time-activity-location patterns for improved estimation of personal exposure to air pollution.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets