Introduction
Wearable sensors that measure personal exposure to air pollution represent a paradigm shift in exposure assessment and epidemiologic research. In contrast to stationary air monitoring, these devices provide users their individual-level information regarding exposure and can identify specific locations and activities associated with increased exposures, including being indoors, in transit (car, bus, and walking), and other settings [Reference Johnston, Juarez, Navarro, Hernandez and Gutschow1,Reference Nieuwenhuijsen, Donaire-Gonzalez, Foraster, Martinez and Cisneros2]. Wearable air pollution monitors can also be linked to spatiotemporal geolocations and paired with other sensors to capture physiologic (e.g., heart rate variability and lung function) and physical activity (e.g., accelerometer) data, thereby connecting personal exposure data with individual health indicators [Reference Turner, Brokamp, Wolfe, Reponen and Ryan3]. As the reliability and accuracy of low-cost air monitoring devices continue to improve their use is expected to become widespread in research studies and by citizen-scientists who seek information regarding personal pollutant exposures [Reference Larkin and Hystad4].
While personal air monitors will provide new opportunities to researchers, the individual-level data they measure requires novel human subject considerations. In contrast to stationary air monitoring or statistical prediction models to estimate air pollution exposure, personal air monitors provide individual level and actionable information. Participants in studies that measure environmental exposures and chemicals in biospecimens report a desire to receive their results, and researchers agree that participants have a right to their individual results [Reference Ohayon, Cousins, Brown, Morello-Frosch and Brody5–Reference Wu, McClean, Brown, Aschengrau and Webster8]. Effective report back of biomonitoring results also increases participants’ engagement with research, and guidelines for providing individual results to participants in biomonitoring studies are now available [Reference Brody, Dunagan, Morello-Frosch, Brown, Patton and Rudel9–Reference Brody, Cirillo and Boronow11]. Biomonitoring studies have also demonstrated that engaging with communities and research participants in the development of report-back materials is critical to effectively develop report-back materials that are appropriate and specific to the study objectives and population [Reference Ohayon, Cousins, Brown, Morello-Frosch and Brody5,Reference Finn and O’Fallon12]. There are critical differences, however, between biomonitoring studies and personal air monitoring and, to our knowledge, studies involving children and adolescents using personal air monitors have not examined participants’ desire to receive their exposure results nor considered best practices to return these results using an understandable and actionable approach.
Therefore, the objective of this study was to engage with participants and caregivers enrolled in the Ecological Momentary Assessment and Personal Particle Exposure (EcoMAPPE), a panel study of adolescents’ personal exposure to ultrafine particles (UFPs) and respiratory health, to determine appropriate methods to provide personal air monitoring results to research study participants. In addition, we applied qualitative research methods to identify the perceived benefits, concerns, and desired content and format for reporting personal air monitoring results back to participants and their caregivers.
Materials and Methods
Overview of Report-Back Development
We employed an iterative, participant-engaged approach to develop personal air monitoring report-back materials that included focus groups of adolescent study participants who completed a study of personal air sampling and their caregivers. In addition, we sought input from an expert panel of researchers experienced in community-engagement, reporting study results, and human subjects research regulatory affairs. Following the initial development of materials, we reconvened our focus groups prior to making final changes to the materials.
Study Population
We enrolled a subset of adolescents (n = 8) who participated in the EcoMAPPE study and their caregivers (n = 8) to participate in separate focus groups for this study. Briefly, the EcoMAPPE study enrolled 118 adolescents aged 13–17 years to complete personal air monitoring and respiratory health assessments [Reference Turner, Brokamp, Wolfe, Reponen and Ryan13]. Personal exposure to UFPs (particulate matter < 100 nm in aerodynamic diameter) was measured during a one-week sampling campaign using the Personal Ultrafine Particle Counter (PUFP C200, EnMont LLC, Cincinnati, OH, USA), a wearable condensation particle counter with a built-in global positioning system (GPS) [Reference Ryan, Son, Wolfe, Lockey, Brokamp and LeMasters14]. The PUFP measures UFP concentrations per cm3 of air every 1 s over concentrations of 0 – 2x105 particles/cm3 (p/cc) with an accuracy of ± 10%. EcoMAPPE participants were instructed to wear the PUFP C200 for 3 h each day of sampling (the maximum battery life of the PUFP on a full charge). Additional details regarding the EcoMAPPE study and results have been described elsewhere [Reference Turner, Brokamp, Wolfe, Reponen and Ryan3,Reference Turner, Brokamp, Wolfe, Reponen and Ryan13].
We recruited adolescent focus group participants by stratifying the EcoMAPPE study population by self-reported race (White/non-White) and gender (male/female) and randomly contacting participants among each strata to ensure focus group participants were not the single representative of their race and gender. Recruitment was conducted by providing information about the study to the participant and caregivers by email with follow-up by phone. Caregivers of interested adolescents were also enrolled to form two focus groups of eight participants each. The size of each focus group was intentionally limited to eight to allow hands-on design and to encourage discussion. Ethical approval to conduct this research was obtained from the Cincinnati Children’s Hospital Institutional Review Board. Prior to participating in the focus groups, all caregivers and adolescents provided written informed consent and assent, respectively.
Adolescent and Caregiver Focus Groups
We employed a qualitative approach to facilitate interactive discussion and co-design of report-back materials, with particular attention to content, design features, and communication strategies. Separate (adolescent and caregiver) in-person focus groups were semi-structured and designed to elicit meaningful qualitative feedback regarding their interest in receiving study results, perceived benefits and concerns, content, format, and delivery methods for reporting results back to participants and caregivers following personal air pollution monitoring. Adolescent and caregiver focus groups were scheduled to last one hour and facilitated by one qualitative researcher (AP and EH, respectively) and one member of the research study team. Focus groups were audio-recorded and transcribed verbatim.
Focus group guides were designed prior to meeting by the qualitative methodologists with additional input from the study team. Participant and caregiver focus group guides were similar, though the adolescent focus group guide included additional time for rapport building and activities to encourage participation. In both focus groups, participants were asked to describe the EcoMAPPE study to gain a better understanding of the words that they used to explain the research. Participants in the adolescent focus group were asked to design their own report-back materials, including drawing out an example of their own personal air pollution monitoring periods with relevant information to explain this to someone who had not participated in the study. Participants then did a “gallery walk” of the examples and wrote down the elements that they liked from each example. Participants also discussed how they would change their own map based on what others had drawn. In each of the focus groups, participants were provided examples of report-back documents that contained information about UFPs and prototype interactive graphs where users could toggle between different sampling days and asked to provide feedback on the materials including questions they had and suggestions for improvement.
Expert Panel Review
We solicited feedback from a panel of four external experts in academia not involved with the EcoMAPPE study and with experience in community-engaged research, environmental health, personal exposure result reporting, and institutional review boards. Prior to a virtual meeting, the expert panel was provided with an electronic draft of the report-back materials developed following the initial focus groups of participants and caregivers. During a virtual semi-structured meeting, the expert panel reviewed all sections of the draft report-back material with the study team and provided suggestions for revisions that included changes to the language, figures, maps, colors, and information provided.
Repeat Focus Group
After revising the report-back document based on feedback from the expert panel, we reconvened the caregiver and adolescent focus groups in the summer of 2021 (Fig. 1). Due to the COVID-19 pandemic, the second focus groups were conducted virtually with separate sessions for adolescents and caregivers. In both focus groups, participants were provided a copy of the revised electronic report-back document followed by a guided discussion to elicit feedback on the presentation, content, readability, and usability of the report. Final revisions to the report-back document were made based on this feedback.
Fig. 1. Introduction to the report-back document.
Focus Group Analysis
We conducted a thematic analysis of the transcribed focus groups. Two members of the research study team and one data analyst external to the study made up the analysis team. All three analysts independently read the data line by line and recorded inductively derived open code (i.e., in vivo codes) for each relevant section of text in the data. The three data analysts reviewed codes together and came to consensus with the external team member providing feedback regarding assumptions and biases held by the research study team members. This process resulted in the development of a codebook that included examples of themes and subthemes. One research team member (AP) assigned the codes to themes and subthemes using inductive coding. The synthesized data were shared with the rest of the analysis team for feedback. Data and coding were organized using ATLAS.ti qualitative data analysis software.
Report-Back Document Development
Based on focus group input, we designed the report-back document to be electronic and interactive. We developed a digital HTML report using R (4.1.1) and R Markdown (2.11) which incorporated instructional images and interactive plots and maps. Participant UFP sampling data were aggregated to 5-second median values for visualization, and the R packages “highcharter” (0.9.4) and “mapdeck” (0.3.4) were used to produce interactive plots and maps [Reference Kunst15,Reference Cooley16]. Standalone, instructional images were also included in the interactive report to provide background information on the personal sampler, UFPs, pollution sources, and reduction strategies.
Results
Focus Group Results
We enrolled a total of eight adolescents who participated in the EcoMAPPE study and one corresponding caregiver to provide their input on the report-back documents during focus group sessions. Adolescent participants were, on average, 14.4 years of age (range: 13–16 years) and one-half (n = 4) were female. Participants self-reported race included Black/Bi-racial (n = 2), White (n = 4), and more than one race (n = 2) and 25% (n = 2) reported having current asthma. Enrolled caregivers were, on average, 49 years of age, primarily female (n = 6), and White (n = 6). Five themes were derived from the coded focus group data and include: 1) feasibility and acceptability of the study, 2) visualizing data, 3) what participants want to know, 4) problems with the “black box,” and 5) utilization of the app (i.e., electronic report-back document).
When the focus group participants were asked to draw an example of how they would visualize their personal air pollution time periods, the prompt provided was “Think of an example of one of the times you wore the sensor. Using the paper in front of you sketch out a map of the time-period.” They were then asked to explain their map to the person next to them and adjust based on that conversation – what did they need to add or change to make the map understandable? The drawings from the adolescent focus groups revealed their preferences for the information included in report-back materials and how they would communicate that information. All participants drew the places that they went while in the study and many had some indication of travelling to and from each place. All participants indicated the level of pollution of each location using either color or density of dots or a combination of both. Two participants provided a key that indicated how the colors corresponded with the level of pollution and one of these participants provided a graph that indicated change in pollution as they traveled to and from different locations.
Report-Back Document
Participants provided feedback on the visualization of the data on the mockup report-back materials. They shared which colors were better for visualizing the level of exposure to UFPs. One participant said, “when you’re in one spot and there are bigger [dots] and smaller [dots] that might be hard to see but with the color you can understand it a little better.” They also asked that that the report-back materials “map out the timeline” and the “direction we were traveling.” They also preferred illustrations to words, when possible and said things like, “it needs to be something that’s not childish but it still needs to be fun and catch attention.” Participants said they would “open it [the report-back document] up on my phone because I don’t actually have a computer with me and then go look at it on the computer because it’s easier.”
Based on this feedback from participants, caregivers, and experts, we produced a single web page to be provided electronically via email to the study participants and their caregivers. The HTML report-back document was composed of 1) an introduction to the materials including an overview of UFPs (Fig. 1), 2) a comparison of their median exposure compared to the entire study population (Fig. 2), 3) a guide to understanding the personal sensor and results (Fig. 3), 4) visualization of their personal sampling data using an interactive map and time-series of data corresponding to each day of personal sampling (Fig. 4), and 4) information on sources of UFPs, potential health risks, and strategies to reduce exposure (Fig. 5).
Fig. 2. Comparison of participant’s median ultrafine particle exposure to the overall study median exposure.
Fig. 3. Guide to understanding the personal air sensor and results.
Fig. 4. Interactive map and time series of personal air monitoring results for each day of personal sampling.
Fig. 5. Additional information regarding sources of ultrafine particles and strategies to reduce personal exposure.
A key component to the report-back document included a statement that personal exposures change frequently based on many factors (Fig. 1). This point was emphasized by the expert panel as critical for the participants to recognize. Following a reminder of their dates of participation, the report-back document provides background information on UFPs, routes of exposure, and potential health risks. Language included in this section was informed by the focus group participants and the expert panel feedback from the expert panel. The first data provided to the participants are their overall median exposure levels recorded during the week of personal sampling (Fig. 2). Following discussion in the focus groups and expert panel, we determined a bar chart was the most informative approach to present this data. Prior to providing their overall median exposures, a brief definition of median is provided. Importantly, caregivers expressed a desire to place the overall levels of exposure in the context of other studies. Therefore, a general statement regarding outdoor exposure levels in other studies is also provided.
Prior to reporting the results of individual real-time sampling data, a summary of the PUFP sensor is provided (Fig. 3). Then, an informative guide to understand the maps and time-series data is presented (Fig. 3) so that participants can interpret their personal sampling data. This includes a color scale, which represent UFP exposure categories ordered from low (green; < 25K p/cc) to high (red; > 100K p/cc). As shown in Fig. 4, instructions for how to interact with their personal particle data is provided. Participants can explore the map of their personal exposures by zooming in and out and panning across the locations of each measurement. Participants can also rotate the map to visualize the locations in 3D. Maps and corresponding time-series plots also display additional information when the user places their cursor or taps on a data point. For the maps, the UFP concentration is displayed. Likewise, the UFP concentration, date and time of the measurement, and a vertical bar highlighting the data point are displayed while hovering on the time-series plot. Separate tabs corresponding to each sampling period are available so that participants can visualize the results of their sampling by day of participation. The report-back document concludes with additional information regarding indoor and outdoor sources of UFPs and potential strategies to reduce exposures (Fig. 5).
Discussion
To our knowledge, previous studies of personal air pollution exposure have not examined participants’ desire to receive their results nor considered best practices to return these results using an understandable and actionable approach. In this study, we engaged adolescents who completed personal air sampling for UFPs, their caregivers, and academic experts to co-design report-back materials that is appropriate and specific to the study population – a critical component to developing report-back documents [Reference Ohayon, Cousins, Brown, Morello-Frosch and Brody5,Reference Finn and O’Fallon12]. Using this approach, we identified key components to the report-back document including background information about UFPs and the device that was used to measure their exposure, providing a summary of individual results compared to others, an interactive visualization of their personal monitoring results, and information on sources of exposure, reduction strategies, and health risks.
Importantly, our report-back document was created as an HTML document based on our participants desire to receive their data in electronic format. In addition to being the preferred method of adolescents to receive their study results, an electronic document provides some additional benefits compared to a written or printed report. For example, we were able to embed interactive maps and figures into the electronic HTML document which allows the viewer to zoom in to specific locations and times of interest. In addition, electronic documents may enhance readability for adolescents as information can be tabulated of separate pages or links to decrease the overall length of the document. Finally, electronic documents may be generated and distributed in shorter time frames allowing for results to be provided with smaller delays from the time of data collection to dissemination.
In response to participants in biomonitoring studies desire to receive their results, guidelines for providing biomonitoring results back to study participants are now available [Reference Brody, Dunagan, Morello-Frosch, Brown, Patton and Rudel9,Reference Dunagan, Brody and Morello-Frosch10]. There are, however, important differences between biomonitoring studies and data collected by personal air monitors at a high resolution over time that necessitated this study. For example, biomonitoring studies typically collect a limited number of biospecimens (e.g., blood, urine) to assess the presence of chemicals with varying half-lives. Frequently, months or years may pass from the time of sample collection to return of biomonitoring results due to the time required to perform laboratory assays. In contrast, personal air monitors measure individuals’ exposure with high temporal resolution (e.g., every 1 s) and can be combined with Bluetooth or other technology to facilitate immediate data transfer to the study participant or researcher. This temporal resolution, when combined with GPS-captured location, provides a unique opportunity to identify places, activities, and times with elevated personal exposures. Thus, personal air monitoring provides the user actionable information to raise awareness and guide behavioral changes to reduce exposure [Reference Steinle, Reis and Sabel17,Reference Steinle, Reis and Sabel18]. These behavioral changes offer the potential to significantly reduce overall exposure to air pollution given that short term but high air pollution events contribute a disproportionate percentage of individuals’ total daily exposure [Reference Ryan, Son, Wolfe, Lockey, Brokamp and LeMasters14,Reference Van Ryswyk, Wheeler and Wallace19–Reference Buonanno, Stabile, Morawska and Russi24]. Furthermore, behavioral modifications to reduce exposures, including modifying cycling and pedestrian routes, can significantly reduce personal UFP, black carbon, and particulate matter (PM) exposures [Reference Cole-Hunter, Jayaratne, Stewart, Hadaway, Morawska and Solomon25,Reference Hankey, Lindsey and Marshall26].
Providing study participants their individual results may also lead to increased trust in science, environmental health literacy, individual empowerment, and motivation to reduce environmental exposures [Reference Brody, Dunagan, Morello-Frosch, Brown, Patton and Rudel9]. Engaging participants by providing them the results of their research provides opportunity to progress from recognizing the relationship between exposure and health, to actively reducing exposure and improving health [Reference Finn and O’Fallon12]. Reporting results to participants also facilitates a collaborative research approach where participants not only provide data but also receive feedback. This engagement can enhance recruitment and retention, detect novel sources of exposure, and identify new research topics [Reference Ramirez-Andreotta, Brody, Lothrop, Loh, Beamer and Brown6,Reference Ramirez-Andreotta, Brody, Lothrop, Loh, Beamer and Brown27].
In addition to engaging with participants through focus groups, our study had several additional strengths including the input of experts experienced in community-engaged research, human subjects’ research, and personal air sampling but not associated with the EcoMAPPE study. These researchers provided external perspective and suggestions on our report-back materials prior to our second focus groups. Our focus on UFPs, an air pollutant hypothesized to be more toxic than larger particles but whose human health effects are understudied [Reference Elder and Oberdörster28–32], has also allowed us to develop reports of exposure to a pollutant with unknown health effects. Finally, we provided the participants with interactive maps and corresponding time-series data enabling them to identify specific times and locations where their personal exposures were elevated or decreased. Future research will examine whether providing participants their individual results increase their knowledge and awareness of environmental health, air pollution, and motivates changes in behavior to reduce air pollution exposures.
Acknowledgments
This study was funded by the National Institute of Environmental Health Sciences grant R21ES030092 (PR, EH, CB, AP, CW). The authors would like to thank all the study participants and their caregivers who contributed to this study and Elise Wright for her assistance in generating figures for the report-back document. The authors would also like to thank Drs. Katrina Korfmacher, Micheal Linke, Ellen Hahn, and Nick Newman for their participation in the expert panel and input regarding the design and content of the report-back materials.
Disclosures
The authors declare that they have no actual or potential conflicts of interest.
Ethics approval and consent to participate
The Cincinnati Children’s Hospital Medical Center Institutional Review Board approved this study. All participants provided consent and assent prior to participation. All authors have reviewed and approved the manuscript for publication. Examples of report-back materials and code are available upon request.
Open access
