Comparison of next‐generation portable pollution monitors to measure exposure to PM 2.5 from household air pollution in Puno, Peru

Vanessa Burrowes, Ricardo Piedrahita, Ajay Pillarisetti, Lindsay J. Underhill, Magdalena Fandiño‐Del‐Rio, Michael Johnson, Josiah L. Kephart, Stella M. Hartinger, Kyle Steenland, Luke P. Naeher, Katherine Kearns, Jennifer L. Peel, Maggie L. Clark, William Checkley, Abidan Nambajimana, Amit Verma, Amy E. Lovvorn, Anaité Díaz-Artiga, Aris T. Papageorghiou, Ashley Toenjes, Ashlinn Quinn, Azhar Nizam, Barry Ryan, Bonnie N. Young, Dana Boyd Barr, Dina Goodman, Eduardo Canúz, Elisa Puzzolo, Eric D. McCollum, Erick Mollinedo, Fiona Majorin, Florien Ndagijimana, Ghislaine Rosa, Gurusamy Thangavel, Howard H. Chang, Irma Sayury Pineda Fuentes, J. Jaime Miranda, Jean de Dieu Ntivuguruzwa, Jean Damascene Uwizeyimana, Jeremy A. Sarnat, Jiawen Liao, John P. McCracken, Joshua P. Rosenthal, Juan Gabriel Espinoza, Julia McPeek Campbell, Kalpana Balakrishnan, Kendra N. Williams, Kirk R. Smith, Krishnendu Mukhopadhyay, Lance A. Waller, Lawrence H. Moulton, Lindsay M. Jaacks, Lisa Elon, Lisa M. Thompson, Margaret A. Laws, Marilú Chiang, Marjorie Howard, Mary E. Crocker, Miles A. Kirby, Naveen Puttaswamy, Oscar de León, Phabiola Herrera, Rachel Craik, Rachel Merrick, Sankar Sambandam, Sarada S. Garg, Sarah Rajkumar, Savannah Gupton, Shakir Hossen, Sheela Sinharoy, Shirin Jabbarzadeh, Steven A. Harvey, Suzanne Simkovich, Thomas Clasen, Usha Ramakrishnan, Víctor G. Dávila‐Román, Vigneswari Aravindalochanan, Yunyun Chen, Zoë SakasIndoor 2019 · DOI

TL;DR

This study validates two new, portable PM2.5 monitors (ECM and UPAS) against a traditional gravimetric method in Puno, Peru. Both new devices showed comparable performance, suggesting they are viable alternatives for assessing household air pollution exposure in low-resource settings, which is crucial for understanding health impacts and evaluating interventions.

What Problem Does This Paper Address?

Household air pollution (HAP), primarily from burning solid fuels for cooking, is a major global health threat, contributing to millions of premature deaths annually, especially from respiratory illnesses and pneumonia in children. Accurately measuring personal exposure to fine particulate matter (PM2.5) is essential for understanding the true health burden of HAP, establishing dose-response relationships, and effectively evaluating the impact of interventions like cleaner cookstoves. However, traditional methods for measuring personal PM2.5 exposure are often bulky, expensive, and require specialized training, limiting their widespread use, particularly in low- and middle-income countries. This paper addresses the need for more accessible, user-friendly, and cost-effective portable monitors that can provide reliable PM2.5 measurements in real-world field settings.

Methods

The researchers conducted a pilot study in Puno, Peru, comparing two next-generation portable PM2.5 monitors – the Enhanced Children MicroPEM (ECM) and the Ultrasonic Personal Air Sampler (UPAS) – against a traditional gravimetric method using a Triplex Cyclone and SKC Air Pump. They performed co-location measurements, meaning all three devices were placed together to capture the same air samples. This was done in three scenarios:

  • Low Exposure: Field workers wore the devices to measure personal exposure in homes using gas or traditional stoves.
  • Medium Exposure: Study participants (primary cooks) wore the devices for 24 hours in their homes.
  • High Exposure: The devices were placed in kitchens at a set distance from the cookstove to measure area concentrations during cooking events.

The study focused on gravimetric measurements, where PM2.5 is collected on a filter and then weighed. They analyzed the data using Spearman correlation to assess the strength of the linear relationship between the devices and the Bland-Altman method to evaluate agreement and identify any systematic bias. Sensitivity analyses were also performed to assess the impact of outliers and data imputation.

Datasets

The study utilized data collected from 72 ECM and 71 UPAS co-locations with the traditional cyclone/pump sampler. A total of 215 filters were analyzed. The data was collected in Puno, Peru, a high-altitude region with specific cultural and environmental characteristics relevant to household air pollution. No new benchmark datasets were introduced; rather, the study aimed to validate existing instruments against a recognized standard.

Key Insights and Ideas

  • Accessibility is Key: The drive for new monitors stems from the limitations of traditional equipment. The ECM and UPAS are highlighted as being lighter, smaller, and easier to use, which is crucial for widespread adoption in resource-limited settings.
  • Real-World Validation Matters: Moving beyond lab-based testing, this study evaluated the monitors in a real-world, high-altitude environment with non-technical field staff, providing a more practical assessment of their performance.
  • Bias is Inherent, But Manageable: While no instrument is perfect, understanding the bias (systematic difference) between new monitors and the gold standard is vital. The study quantifies this bias for both ECM and UPAS, finding it within acceptable ranges for air sampling instruments.
  • Context is Crucial for Interpretation: The study notes that differences in findings between their main analysis and sensitivity analyses might be due to factors like filter clogging, duty cycling differences, and the specific environmental conditions of Puno, emphasizing the importance of considering these when interpreting results.

Key Results

  • Comparable Performance: Both the ECM and UPAS demonstrated similar correlation and agreement with the traditional gravimetric cyclone/pump method across all sample types (overall, area, and personal).
  • Correlation:
    • Overall samples: ECM (ρ = 0.91) and UPAS (ρ = 0.88) showed strong correlations with the gravimetric method.
  • Area samples: UPAS (ρ = 0.91) showed a stronger correlation than ECM (ρ = 0.81).
  • Personal samples: ECM (ρ = 0.74) showed a slightly better correlation than UPAS (ρ = 0.68).
  • Agreement (Bias):
    • Overall samples: UPAS had a lower mean difference (93.9 µg/m³) compared to ECM (121.7 µg/m³), indicating slightly better agreement.
  • Area samples: UPAS (mean difference of 251.1 µg/m³) showed better agreement than ECM (299.8 µg/m³).
  • Personal samples: ECM (mean difference of -14.0 µg/m³) showed better agreement than UPAS (-24.0 µg/m³).
  • Limit of Detection (LOD) Adjustment: When adjusted for the limit of detection, the agreement between both new devices and the cyclone/pump remained similar.
  • Real-time Data: The ECM's ability to provide real-time data allowed for visualization of temporal cooking patterns and peak exposure times, a valuable feature for understanding exposure dynamics.

Limitations and Future Work

The authors acknowledge several limitations:

  • Sample Size: The pilot study had a relatively small sample size (n=82), particularly for area samples (n=23), which limits the generalizability of the findings and the ability to definitively assess performance at very low concentrations. A larger sample size would strengthen the conclusions.
  • Outliers and Sensitivity Analysis: The presence of outliers, especially in area samples, influenced the Bland-Altman results. While sensitivity analyses helped explore this, the interpretation of area high-exposure samples was noted as potentially less relevant to the primary goal of assessing low-exposure personal exposures.
  • Duty Cycling: The use of different duty cycles for the ECM (to prevent filter clogging) and the UPAS/cyclone/pump (100% duty cycle) might have influenced mass loading and comparability, especially in highly polluted environments.
  • Field Staff Experience: The field staff had more prior experience with the ECM than the UPAS, which could have introduced subtle biases in handling and processing.
  • Geographic Specificity: The study was conducted in Puno, Peru. Further testing in other low- and middle-income countries with different environmental and fuel-use characteristics is needed.

Future work should focus on larger-scale validation studies with more diverse populations and environments. Further investigation into the impact of different duty cycles and filter sizes on measurement accuracy in highly polluted settings is also recommended. The potential of the ECM's real-time data for capturing temporal exposure patterns warrants further exploration.

Summary updated

Abstract

Assessment of personal exposure to PM<sub>2.5</sub> is critical for understanding intervention effectiveness and exposure-response relationships in household air pollution studies. In this pilot study, we compared PM<sub>2.5</sub> concentrations obtained from two next-generation personal exposure monitors (the Enhanced Children MicroPEM or ECM; and the Ultrasonic Personal Air Sampler or UPAS) to those obtained with a traditional Triplex Cyclone and SKC Air Pump (a gravimetric cyclone/pump sampler). We co-located cyclone/pumps with an ECM and UPAS to obtain 24-hour kitchen concentrations and personal exposure measurements. We measured Spearmen correlations and evaluated agreement using the Bland-Altman method. We obtained 215 filters from 72 ECM and 71 UPAS co-locations. Overall, the ECM and the UPAS had similar correlation (ECM ρ = 0.91 vs UPAS ρ = 0.88) and agreement (ECM mean difference of 121.7 µg/m<sup>3</sup> vs UPAS mean difference of 93.9 µg/m<sup>3</sup> ) with overlapping confidence intervals when compared against the cyclone/pump. When adjusted for the limit of detection, agreement between the devices and the cyclone/pump was also similar for all samples (ECM mean difference of 68.8 µg/m<sup>3</sup> vs UPAS mean difference of 65.4 µg/m<sup>3</sup> ) and personal exposure samples (ECM mean difference of -3.8 µg/m<sup>3</sup> vs UPAS mean difference of -12.9 µg/m<sup>3</sup> ). Both the ECM and UPAS produced comparable measurements when compared against a cyclone/pump setup.
BibTeX
@inproceedings{burrowes2019comparison,
  author = {Vanessa Burrowes and Ricardo Piedrahita and Ajay Pillarisetti and Lindsay J. Underhill and Magdalena Fandiño‐Del‐Rio and Michael Johnson and Josiah L. Kephart and Stella M. Hartinger and Kyle Steenland and Luke P. Naeher and Katherine Kearns and Jennifer L. Peel and Maggie L. Clark and William Checkley and Abidan Nambajimana and Amit Verma and Amy E. Lovvorn and Anaité Díaz-Artiga and Aris T. Papageorghiou and Ashley Toenjes and Ashlinn Quinn and Azhar Nizam and Barry Ryan and Bonnie N. Young and Dana Boyd Barr and Dina Goodman and Eduardo Canúz and Elisa Puzzolo and Eric D. McCollum and Erick Mollinedo and Fiona Majorin and Florien Ndagijimana and Ghislaine Rosa and Gurusamy Thangavel and Howard H. Chang and Irma Sayury Pineda Fuentes and J. Jaime Miranda and Jean de Dieu Ntivuguruzwa and Jean Damascene Uwizeyimana and Jennifer L. Peel and Jeremy A. Sarnat and Jiawen Liao and John P. McCracken and Joshua P. Rosenthal and Juan Gabriel Espinoza and Julia McPeek Campbell and Kalpana Balakrishnan and Kendra N. Williams and Kirk R. Smith and Krishnendu Mukhopadhyay and Lance A. Waller and Lawrence H. Moulton and Lindsay M. Jaacks and Lisa Elon and Lisa M. Thompson and Margaret A. Laws and Marilú Chiang and Marjorie Howard and Mary E. Crocker and Miles A. Kirby and Naveen Puttaswamy and Oscar de León and Phabiola Herrera and Rachel Craik and Rachel Merrick and Sankar Sambandam and Sarada S. Garg and Sarah Rajkumar and Savannah Gupton and Shakir Hossen and Sheela Sinharoy and Shirin Jabbarzadeh and Steven A. Harvey and Suzanne Simkovich and Thomas Clasen and Usha Ramakrishnan and Víctor G. Dávila‐Román and Vigneswari Aravindalochanan and Yunyun Chen and Zoë Sakas and Zoe Sakas},
  title = {{Comparison of next‐generation portable pollution monitors to measure exposure to PM <sub>2.5</sub> from household air pollution in Puno, Peru}},
  booktitle = {Indoor},
  year = {2019},
  doi = {10.1111/ina.12638},
}
Figure