Journal of Exposure Science and Environmental Epidemiology (2015), 1–7 © 2015 Nature America, Inc. All rights reserved 1559-0631/15
www.nature.com/jes
ORIGINAL ARTICLE
Time–location patterns of a diverse population of older adults: the Multi-Ethnic Study of Atherosclerosis and Air Pollution (MESA Air) Elizabeth W. Spalt1, Cynthia L. Curl1, Ryan W. Allen2, Martin Cohen1, Sara D. Adar3, Karen H. Stukovsky4, Ed Avol5, Cecilia Castro-Diehl6, Cathy Nunn7, Karen Mancera-Cuevas8 and Joel D. Kaufman1,9,10 The primary aim of this analysis was to present and describe questionnaire data characterizing time–location patterns of an older, multiethnic population from six American cities. We evaluated the consistency of results from repeated administration of this questionnaire and between this questionnaire and other questionnaires collected from participants of the Multi-Ethnic Study of Atherosclerosis and Air Pollution (MESA Air). Participants reported spending most of their time inside their homes (average: 121 h/ week or 72%). More than 50% of the participants reported spending no time in several of the location options, including at home outdoors, at work/volunteer/school locations indoors or outdoors, or in “other” locations outdoors. We observed consistency between self-reported time–location patterns from repeated administration of the time–location questionnaire and compared with other survey instruments. Comparisons with national cohorts demonstrated the differences in time–location patterns in the MESA Air cohort due to differences in demographics, but the data showed similar trends in patterns by age, gender, season, and employment status. This study was the first to explicitly examine the time–location patterns in an older, multiethnic population and the first to add data on Chinese participants. These data can be used to inform future epidemiological research of MESA Air and other studies that include diverse populations. Journal of Exposure Science and Environmental Epidemiology advance online publication, 29 April 2015; doi:10.1038/jes.2015.29 Keywords: epidemiology; personal exposure; population-based studies
INTRODUCTION A standard method for assessing the individual-level exposure to environmental pollutants is to calculate the “time-weighted” average of microenvironmental concentrations. This approach can estimate exposure to many different types of pollutants, including air pollutants.1,2 The more variable the pollutant concentrations are across microenvironments, the greater the need to accurately characterize both the pollutant concentrations and the time spent in each location. Air pollutant concentrations can be quite variable across proximal microenvironments. Ambient-generated particulate matter concentrations can decrease 50% or more going from outside to inside a home.3–5 Concentrations of particles, volatile organic compounds, nitric oxide, and nitrogen dioxide inside vehicles traveling on roadways have also been found to be significantly higher than ambient levels.6–13 Despite this knowledge, most air pollution epidemiology relies on the assumption that residential outdoor concentrations of air pollutants are good proxies for individual exposure.14–21 This assumption is based on data from studies suggesting that
individuals spend the majority of time at home.22–25 However, failing to characterize time spent in other microenvironments, especially the indoor environment at home, will likely introduce measurement error. Setton et al.26 found that residence-only estimates of air pollution exposure have considerable bias compared with estimates that take into account the time spent working or traveling. State-of-the-art air pollution epidemiology is now moving from simply using outdoor concentration data to incorporating information on person-specific characteristics to more accurately estimate individual exposures.27 However, this approach requires more accurate microenvironment time–location data. Several large-scale studies of time–location patterns provide time–location data for use in exposure assessments. The National Human Activity Pattern Survey (NHAPS) includes 24-h diary data in 82 different locations for 9,386 participants in 48 states.24 The Consolidated Human Activity Database (CHAD)22 includes NHAPS, along with data from other national and regional studies. Data from these types of large-scale studies or databases can be a useful alternative for collecting study-specific data. Moreover, these data are used in national exposure models including USEPA air pollution exposure models: Air Pollution Exposure Model28
1 Department of Environmental and Occupational Health Sciences, University of Washington, Seattle, Washington, USA; 2Faculty of Health Sciences, Simon Fraser University, Burnaby, British Columbia, Canada; 3Department of Epidemiology, University of Michigan, Ann Arbor, Michigan, USA; 4Department of Biostatistics, University of Washington, Seattle, Washington, USA; 5Department of Preventive Medicine, University of Southern California, Los Angeles, California, USA; 6Department of Medicine, Columbia University, New York, New York, USA; 7Department of Public Health Sciences, Wake Forest University School of Medicine, Winston-Salem, North Carolina, USA; 8Division of Rheumatology, Feinberg School of Medicine, Northwestern University, Chicago, Illinois, USA; 9Department of Epidemiology, University of Washington, Seattle, Washington, USA and 10 Department of Medicine, University of Washington, Seattle, Washington, USA. Correspondence: Elizabeth W. Spalt, Department of Environmental and Occupational Health Sciences, University of Washington, Campus Box 354695, 4225 Roosevelt Way NE, Suite 302, Seattle WA 98105, USA. Tel.: 206-897-1436. Fax: 206-897-1991. E-mail: [email protected] Received 6 May 2014; revised 25 February 2015; accepted 26 February 2015
Time–location patterns: MESA Air Spalt et al
2 and Stochastic Human Exposure and Dose Simulation model (SHEDS).29 However, existing information may not be representative of unique populations. Studies of very specific cohorts (e.g., pregnant women;30 minorities, and populations living below the poverty line31) have found it necessary to collect data on patterns of time–location specific to their populations. These data not only are valuable for the study in which they were collected, but also broaden the population for which the time–location patterns are understood. USEPA has recently made specific recommendations for the collection of additional time–location data for older populations.32 The primary aim of this analysis is to present the time–location questionnaire data of an older, multiethnic cohort of American adults in the Multi-Ethnic Study of Atherosclerosis and Air Pollution (MESA Air).27 We also evaluate the reliability of time– location data by comparison of repeated administration of the questionnaire and across several survey instruments. Finally, we compare the time–location patterns of MESA Air participants to those in national cohorts. This work adds to the literature as the MESA Air cohort is older and more racially/ethnically diverse than other cohorts in which the time–location patterns have previously been described.
was not administered at the same time as the MAQ, results from the previous clinic exam (230-1,122 days prior to the MAQ) were used. Each of these questionnaires is described in more detail below.
MESA Air Questionnaire The MAQ was the primary data collection tool for gathering information on home characteristics relevant to pollutant infiltration efficiencies and about behaviors related to individual exposures. Participants were asked specific questions about their typical time–location patterns in winter and summer and could designate if their patterns were the same in both seasons. For each day of the week, the MAQ included questions documenting time spent (rounded to the nearest hour) in each of the seven locations — home indoors, home outdoors, work/volunteer/school indoors, work/ volunteer/school outdoors, in transit (e.g., car, bike), other indoors, and other outdoors. Participants also designated which days of the week they considered weekends and weekdays. The amount of time by transit mode (e.g., walking/biking, car/taxi, bus, train/subway), road types traveled (e.g., freeways, residential streets), and traffic conditions experienced (e.g., light traffic/moving at the speed limit) were also documented. Additional questions asked about home characteristics related to building type, building age, the presence of an attached garage, and other factors relevant to infiltration. Methods for addressing missing MAQ data are presented in the Supplementary Materials.
Time–Location Diary METHODS Study Population MESA Air is an ancillary study to the Multi-Ethnic Study of Atherosclerosis (MESA33), a long-term study of the progression of cardiovascular disease (CVD) in adults. MESA included 6814 participants from six US communities: Baltimore, MD; Chicago, IL; Forsyth County, NC; Los Angeles, CA; New York, NY; and St. Paul, MN. MESA participants were aged 45–84 years at enrollment between 2000 and 2002, with an approximately equal gender ratio, and were free of recognized CVD at baseline. Four ethnic/racial groups were targeted for inclusion: white, black, Hispanic, and Chinese, but recruitment of racial/ethnic groups varied by study site. All MESA Air participants provided informed consent before participation. The primary aim of MESA Air is to understand the relationship between the individual-level ambient-derived air pollution exposure and the progression of subclinical CVD and the incidence of clinical disease events.27 MESA Air includes 6424 participants, primarily recruited between 2005 and 2007 from the MESA cohort. Additional participants were recruited from a second ancillary study to MESA, the MESA Family Study (n = 490), and directly for MESA Air (n = 257) in three additional areas near existing MESA communities, two in the Los Angeles basin and one near New York City.27 An additional 1127 MESA participants were included in MESA Air analyses but did not complete all aspects of the MESA Air study and were not included in this analysis.
Overview of Participant Questionnaires Every MESA Air participant was asked to complete a comprehensive MESA Air Questionnaire (MAQ, see Supplementary Materials) at recruitment (hereafter referred to as the baseline MAQ) to describe typical activity patterns in the winter and summer. These questions were asked during a face-to-face interview during a MESA clinic exam between 2005 and 2007. The MAQ was repeated up to six times during follow-up phone calls and a later clinic exam; repeated administration was triggered for participants who indicated a major change in lifestyle (change in residence, work or school status, caregiver status, or in household members). At a subsequent exam, the MAQ was also re-administered to a small subset of participants (n = 10%) who did not report a major lifestyle change to assess changes in responses over time. For participants with multiple MAQs, the time between administration ranged from o3 months to 6.7 years. Detailed daily time–location diaries (TLDs, see Supplementary Materials) were also completed by a small subset of participants (n = 89) who took part in a personal air pollution exposure study conducted between October 2006 and July 2008.34 In addition, MESA gathered information on various participant activities including employment and physical activity from all participants via technician-administered Personal History and Physical Activity Questionnaires. Personal History Questionnaires were administered at all clinic exams, whereas Physical Activity Questionnaires were administered at all but one exam. When the Physical Activity Questionnaire Journal of Exposure Science and Environmental Epidemiology (2015), 1 – 7
In the TLDs, the subset of participants undergoing personal sampling recorded hourly time spent in specified locations for each day during two 2-week sampling periods using the following classifications — home indoors, home outdoors, motorized vehicle, work indoors, work outdoors, other indoors, and other outdoors. Season classifications were determined using the average temperature over the specific sampling period. If the average temperature was 418 °C, the season was classified as “warm”; otherwise, the season was “cool”.3 For comparison with the MAQ, we classified “warm” seasons as summer and “cool” seasons as winter. In the event that we were not able to capture two distinct seasons for an individual participant, results were averaged to obtain a single set of results for each participant for a season. Methods for addressing missing TLD data are presented in the Supplementary Materials.
Personal History and Physical Activity Questionnaires The Personal History Questionnaire asked all participants about their current employment status in ten job categories: (1) homemaker, (2) employed full-time, (3) employed part-time, (4) employed — on leave for health reasons, (5) employed — temporarily away from job, (6) unemployed o6 months, (7) unemployed 46 months, (8) retired — not working, (9) retired — working, and (10) retired — volunteering. For comparison with the MAQ, categories 2, 3 and 9 were assigned as working outside of the home, whereas the remaining seven categories were classified as not working outside of the home. In addition, as part of the Physical Activity Questionnaire, participants provided information on time spent in 10 activities: (1) working at their job, (2) conducting volunteer work, (3) walking, (4) doing yard work, (5) traveling via car, subway, or bus, (6) performing household chores, (7) taking care of others, (8) dancing/practicing a sport, (9) doing conditioning activities (e.g., aerobics), and (10) spending time in leisure activities (e.g., reading).
Data Analysis In order to address our primary aim of presenting the time–location data from the MESA Air cohort, we calculated summary statistics for the amount of time spent in each microenvironment based on the MAQ for the cohort as a whole and by demographic group (i.e., age, race/ethnicity, gender, study site, employment status, education, and income, as reported on the Personal History Questionnaire). This cross-sectional analysis included baseline MAQs only. A second aim of this analysis was to compare the time–location patterns obtained from the MAQ to behaviors reported by the same participants on repeated administration of the MAQ and additional survey instruments: the TLD, the Personal History Questionnaire, and the Physical Activity Questionnaire. To accomplish this aim, the time–location data for each participant who completed both a TLD and a MAQ in the same season were compared using Spearman correlation coefficients and Wilcoxon rank-sum tests. For participants who filled out multiple MAQs, we selected © 2015 Nature America, Inc.
Time–location patterns: MESA Air Spalt et al
3 the MAQ with the closest date to the TLD. We also investigated the impact of gender, age, race/ethnicity, and employment status on observed differences between the different data collection methods. Consistency across MAQs, for those without lifestyle changes, was assessed with Spearman correlation coefficients. Agreement between the MAQ and the Personal History and Physical Activity Questionnaires was assessed for several measures using Wilcoxon Rank Sum tests. Direct comparisons between the MAQ and the Physical Activity Questionnaire were not made because participants reported time in locations on the MAQ and time in activities on this second survey. Time spent at work was compared among working and nonworking persons as classified by the Personal History Questionnaire and among persons reporting time spent at work versus none on the Physical Activity Questionnaire. Time spent outdoors at home was then contrasted between participants reporting some versus no time spent working in the yard on the Physical Activity Questionnaire. Finally, a similar contrast was made for time in transit based on transit and walking behaviors reported in the Physical Activity Questionnaire. All MAQs were matched to Personal History and Physical Activity Questionnaires administered at the same clinical exam or the exam prior for instances when the questionnaire was not administered. Finally, we compared average values of MESA Air time–location data to those reported by NHAPS and CHAD. All statistical analyses were conducted using SAS Software Version 9.3 of the SAS System for Windows (Copyright 2010 SAS Institute, Cary, NC, USA).
RESULTS Participant Characteristics A total of 6209 participants completed the time–location section of at least one MAQ with 5996 participants doing so at baseline. Participants completed the MAQ a mean of 1.9 times and a maximum of six times for a total of 9704 MAQs. Of the 9,704 MAQs administered, 7831 (81%) took place in person during MESA clinic exams, and 1873 (19%) occurred during follow-up calls. A total of 5618 participants at baseline answered the question regarding seasonal differences in time–location patterns, and of these, 45% reported the same patterns in winter and summer. Personal History and Physical Activity Questionnaires were only administered at in-clinic exams. For the 7835 MAQs administered at exams, 7819 (99.8%) of the Personal History Questionnaires collected at the same exam included responses to the question on job status. For the Physical Activity Questionnaires administered during the same exam visit as the MAQ, 7680 (98%) of participants provided information on time engaged in various physical activities. Eighty-nine individuals participated in the 2-week personal monitoring sampling campaigns and provided at least one TLD. Eighty of these participants participated in a second sampling campaign and completed two TLDs, for a total of 169 TLDs. Table 1 presents the demographic characteristics of the MESA Air participants who completed baseline MAQs. The average age of participants that completed the baseline MAQ was 65 ± 10 years, and 53% of the population was female. Fifty percent of the cohort worked full or part-time, and another 7% volunteered (data not shown). Demographic characteristics for those that completed the MAQ are very similar to MESA overall,27 and demographic characteristics for participants who completed a baseline MAQ are very similar to those that completed MAQs overall (data not shown). Time–Activity Patterns in MESA Air On average, participants reported spending the majority (≥70%) of their time indoors at home (Table 2). After time indoors at home, the most time was spent at work/volunteer/school locations indoors; however, 450% of the participants reported that they spent no time indoors at work/volunteer/school. More than 50% of participants also reported spending no time in the three outdoor locations. Participants reported spending more time in indoor locations and less time in outdoor locations in winter than © 2015 Nature America, Inc.
Table 1. Characteristics of MESA Air participants who completed MESA Air questionnaires and time–location diaries. Baseline MAQs N = 5,996
TLDs N = 89
Number
%
Number
%
Gender Female Male
3194 2802
53 47
46 43
52 48
Race/ethnicity White Chinese Black, African-American Hispanic
2258 650 1667 1421
38 11 28 24
54 2 23 10
61 2 26 11
Age category 39–44a 45–54 55–64 65–74 75–84 85+
18 976 1967 1815 1079 141
0 16 33 30 18 2
0 11 37 30 9 2
0 12 42 34 10 2
Site Forsyth County New York Cityb Baltimore St. Paul Chicago Los Angelesc
888 1195 721 872 1136 1184
15 20 12 15 19 20
13 21 14 13 13 15
15 24 16 15 15 17
Socioeconomic status High school education or more Family income ≥ $30,000 year − 1 Single family home
4999d 3671e 3266f
84 65 54
84 68 57
94 82 64
Abbreviations: MAQ, MESA Air Questionnaire; Mesa Air, Multi-Ethnic Study of Atherosclerosis and Air Pollution; TLD, time–location diaries. aMESA family included a small number of participants under the age of 45, and 18 completed MAQs. bIncludes Rockland County participants. cIncludes Riverside participants. dValues missing for 13 participants. eValues missing for 340 participants. fValues missing for three participants.
in summer (Table 2). Figure 1 and Supplementary Table 2a and g present time–location patterns by age, gender, race/ethnicity, site, employment status, education, and income. Consistency of MAQ and TLD Reponses There were a smaller proportion of Chinese and Hispanics participants in the personal monitoring campaigns, so there are fewer TLDs captured in these groups (Table 1). TLD participants also tended to have higher incomes and were more likely to have at least a high school education. The age distributions of the TLD participants and the cohort as a whole were roughly similar except that TLD participants were more likely to be between the ages of 55 and 64 years and less likely to be 75–84 years. The average time between administration of the MAQ and the start of the TLD was 398 days (range: 13–966 days). The highest correlations between the TLD and MAQ responses for each individual were for time spent indoors at work and at home (r = 0.61–0.80), and the weakest correlations were for time spent in other locations (r = − 0.04 to 0.28; Table 3). Agreement between the MAQ and TLD for time spent outdoors was higher in the summer than in the winter. Participants tended to report more time indoors on average (based on their MAQ) than was observed during the 2-week personal monitoring period based on their TLDs (data not shown). Percent difference values between the Journal of Exposure Science and Environmental Epidemiology (2015), 1 – 7
Time–location patterns: MESA Air Spalt et al
4 Table 2.
Summary of reported time (h/week, % of total time) by location reported by participants administered the baseline MAQ (n = 5,996).
Location
Season
Mean
Indoor locations Home Home Work/volunteer/school Work/volunteer/school Other Other
Summer Winter Summer Winter Summer Winter
118 125 15 16 13 14
Outdoor locations Home Home Work/volunteer/school Work/volunteer/school Other Other In transit In transit
Summer Winter Summer Winter Summer Winter Summer Winter
6 2 2 1 6 2 8 7
SD
5th percentile
(70%) (74%) (9%) (10%) (8%) (8%)
27 26 20 21 14 15
(16%) (16%) (12%) (12%) (8%) (9%)
75 84 0 0 0 0
(45%) (50%) (0%) (0%) (0%) (0%)
(4%) (1%) (1%) (1%) (4%) (1%) (5%) (4%)
11 6 8 7 11 6 6 6
(6%) (3%) (5%) (4%) (7%) (4%) (4%) (4%)
0 0 0 0 0 0 0 0
(0%) (0%) (0%) (0%) (0%) (0%) (0%) (0%)
50th percentile 119 126 0 0 10 11 0 0 0 0 1 0 7 7
95th percentile
(71%) (75%) (0%) (0%) (6%) (6%)
158 162 50 51 38 39
(94%) (96%) (30%) (30%) (23%) (23%)
(0%) (0%) (0%) (0%) (1%) (0%) (4%) (4%)
29 12 10 6 28 13 18 17
(17%) (7%) (6%) (4%) (17%) (8%) (11%) (10%)
Abbreviation: MAQ, MESA Air Questionnaire.
40% 35% 30% 25% 20% 15% 10% 5% 0%
In Transit
Other Outdoors
Other Indoors
Work Outdoors
Work Indoors
Home Outdoors
Figure 1. Mean percent time spent in six microenvironments by demographic, socioeconomic group, and site. The remaining time (up to 100%) represents the time spent indoors at home. For example, female participants reported spending 74% of their time indoors at home.
Table 3.
Spearman correlation coefficients (with 95% confidence intervals) between MAQ and TLD responses.
Location
Summer (n = 65)a
Winter (n = 78)a
Both (n = 143)a
Indoor locations Home Work/volunteer/school Other All
0.61 0.80 0.25 0.24
(0.42, 0.74) (0.68, 0.87) (0.00, 0.46) (-0.01, 0.45)
0.65 0.68 0.15 0.30
(0.49, 0.76) (0.54, 0.78) (-0.07, 0.36) (0.08, 0.48)
0.63 0.73 0.20 0.35
(0.51, (0.65, (0.04, (0.20,
Outdoor locations Home Work/volunteer/school Other All In transit
0.19 0.28 0.15 0.08 0.31
(−0.06, 0.41) (0.03, 0.49) (−0.10, 0.38) (−0.17, 0.32) (0.06, 0.51)
0.04 0.13 −0.04 0.01 0.49
(−0.18, 0.26) (−0.10, 0.34) (−0.26, 0.18) (−0.21, 0.24) (0.30, 0.64)
0.14 0.20 0.10 0.10 0.39
(−0.02, 0.30) (0.03, 0.35) (−0.06, 0.26) (−0.07, 0.26) (0.24, 0.52)
0.72) 0.80) 0.35) 0.49)
Abbreviations: MAQ, MESA Air Questionnaire; TLD, time–location diaries. aA total of 89 participants completed TLDs and MAQs. Of these, 65 were in the summer and 78 were in the winter.
Journal of Exposure Science and Environmental Epidemiology (2015), 1 – 7
© 2015 Nature America, Inc.
Time–location patterns: MESA Air Spalt et al
5 Table 4.
Average time spent at home for NHAPS and MESA Air participants by season, gender, and race (h/day). All
Time spent indoors at residence NHAPS MESA Air Time spent outdoors at residencea NHAPS MESA Air
65+ Winter Summer Male Female White Black Hispanic Chinese
16.7 19.6 17.3 18.8
2.3 1.7
2.4 1.8
Employed full-time
Employed part-time
Not employed
17.2 17.8
16.3 16.9
15.8 16.7
17.5 17.9
16.7 17.0
16.9 17.0
16.8 17.6
NR 18.6
14.7 14.7
16.4 17.1
19.3 19.3
1.9 1.2
2.4 2.0
2.6 1.8
1.9 1.6
2.3 1.9
2.1 1.5
2.1 1.8
NR 1.1
2.2 1.4
2.1 1.7
2.4 1.9
Abbreviations: MESA Air, Multi-Ethnic Study of Atherosclerosis and Air Pollution; NHAPS, National Human Activity Pattern Survey; NR, not reported. aIncludes only those participants that report some time outdoors at home for both cohorts.
MAQ and TLD for participants who reported on the MAQ having the same time–location patterns in winter and summer were not different from those who reported different seasonal patterns (Wilxon; P40.05). We also examined whether demographic factors affected the strength of the correlations between responses on the MAQ and the TLD. Supplementary Figure 1 shows the Spearman correlation coefficients between the amount of time that each individual reported in each microenvironment on their MAQ and their TLD, by gender, age, race/ethnicity, and employment status. For most locations, men and women provided similarly consistent data on their MAQs and TLDs (Supplementary Figure 1a). However, the responses for the outdoor work location were less consistent for women than men. Although none of the women in the personal monitoring campaign reported any time in this location on the MAQ, some women did report spending time at work outdoors on their TLD. For individuals o 65 years, the 2-week periods captured by the TLDs were more highly correlated with the typical patterns of behavior reported in the MAQs than for older individuals (Supplementary Figure 1b). On average, individuals ≥ 65 years reported more time indoors and less time outdoors on their MAQ compared to their TLD (data not shown). The correlation of 0 for work outdoors was due to personal monitoring participants ≥ 65 years old reporting no time on the MAQ and some time on the TLDs. All race/ethnic groups were relatively consistent in their responses regarding time spent in transit. There were differences in the consistency between the MAQ and TLD for the rest of the locations, but these differences were not significant (Supplementary Figure 1c). Working individuals were more consistent in their responses related to work locations but not other locations (Supplementary Figure 1d). Consistency of Repeated MAQ Administration A total of 394 participants with no reported lifestyle changes filled out the time–location portion of the MAQ to assess changes in responses over time. The average time between administration of the two MAQs was 1621 days (range: 22–2393 days). Correlation between responses for the time–location patterns was high (0.78) with higher correlations for indoors (0.89) than transit (0.40) or outdoors (0.35). Consistency of MAQ Results with other Survey Instruments available in MESA Comparisons were made between the MAQs and other MESA questionnaires using information on employment status and the time spent at work, working in the yard, in transit, and walking. Not surprisingly, working participants reported spending significantly more time at work compared with individuals in the © 2015 Nature America, Inc.
nonworking categories based on the Personal History Questionnaire (means of 30 and 3 h/week, respectively; P o0.0001) and on the Physical Activity Questionnaire (29 and 4 h/week; P o 0.0001). The amount of time spent outdoors at home from the MAQ was also higher among participants who reported time working in the yard on the Physical Activity Questionnaire than those that did not (6 vs 2 h/week; P o 0.0001). Participants who reported some transit time on the Physical Activity Questionnaire reported an average of 8 h per week of transportation time on the MAQ, whereas those who reported no time in transit on the Physical Activity form reported 5 h per week on the MAQ (P o 0.0001). Participants reporting walking in the Physical Activity Questionnaire also had a higher weekly average amount of time in transit (8 h) compared with those who did not report walking (6 h, P o 0.0001). Comparison with Other Studies of Time–Location Patterns Time–location patterns in the MESA Air cohort were compared with other national cohorts. Comparisons of demographic characteristics for MESA Air, NHAPS, and CHAD are provided in Supplementary Table 3. Although the overall time spent at home indoors was greater for MESA Air than NHAPS participants, similar patterns across seasons and among genders and employment status were observed (Table 4). Overall, CHAD participants reported spending less time indoors than MESA Air participants. CHAD participants over the age of 65 years spent a similar amount of time indoors and outdoors compared with MESA Air participants over 65 (Table 5). For both CHAD and MESA Air, males reported spending less time indoors and more time outdoors than females, and white participants reported spending the least time indoors and the most outdoors (Table 5). DISCUSSION In this study, we successfully deployed and evaluated a time– activity questionnaire in a multiethnic cohort for incorporation into epidemiological analyses. The MAQ-collected information on time spent in microenvironments from MESA Air participants, and agreement between the MAQ and other data collection tools provides evidence of the reliability of this survey. The evaluation of repeated administration of the MAQ and the comparison of the MAQ with other data collection tools demonstrated consistency across data collection methods, and increases our confidence that the MAQ captured reliable information on where MESA Air participants spend their time. Although variable for location, MAQ and TLD responses were positively correlated. Comparisons between the MAQ and other questionnaires administered in MESA provided additional evidence that the MAQ succeeded in capturing time–location patterns. Journal of Exposure Science and Environmental Epidemiology (2015), 1 – 7
Time–location patterns: MESA Air Spalt et al
6 Table 5.
Comparison of average total time spent indoors and outdoors between CHAD and MESA Air (h/day). All
Male
Female
65+
Male: 65+
Female: 65+
White
Black
Hispanic
Chinese
Total time indoors CHAD 21.2a MESA Air 21.5
20.7 21.0
21.5 22.0
21.8 21.8
NR 21.3
NR 22.2
21.1 21.2
21.5 21.5
21.4 21.6
NR 22.3
Total time outdoorsb CHAD 2.3a MESA Air 2.4
2.7 2.8
1.9 2.0
2.0 2.3
2.7 2.7
1.5 1.9
2.4 2.7
2.3 2.5
2.3 2.3
NR 1.4
Abbreviations: CHAD, Consolidated Human Activity Database; Mesa Air, Multi-Ethnic Study of Atherosclerosis and Air Pollution; NR, not reported. aWeighted average of male and female participants. bIncludes only those participants that report some time outdoors for both cohorts.
Although TLD and MAQ responses were correlated, there were many differences between the MAQ and the TLD. The TLDs represent a 2-week snapshot in time, whereas the MAQs are intended to capture typical activities over a season. Requesting information on the “typical” amount of time spent in locations is likely to result in the mode for each participant rather than the mean.35,36 For participants with less routine schedules, disparity between the MAQ and TLD is more likely. In addition, because the MESA Air population is older with a large proportion of retired individuals, the time–location pattern for this population is likely less routine than for a working population. If true, we would not necessarily expect any 2-week period to be reflective of what is typical but can expect to find similarities in aggregate responses. Overall, we found greater consistency for locations where participants spend more time than for those locations where less time is spent. Previous researchers found age- and gender-specific differences for the reporting of time spent working.36,37 Similarly, we found some reporting inconsistencies by gender, age, race/ethnicity, and employment status. Results were similar by gender for all categories except for work outdoors. When we investigated responses by age, we found that an agreement between responses to the MAQ regarding typical patterns of behavior and the specific patterns during the 2-week period captured by the TLD was lower for most locations for individuals aged ≥ 65 years. Accuracy of recalled information is likely to decrease as people age, but differences may also be because individuals over the age of 65 are less likely to be working (Wilcoxon, P o0.0001) and therefore may have a less routine schedule. Differences between MESA Air and NHAPS may be due to differences in demographic composition, collection methodology, and/or dates of collection. In MESA Air, white participants spent the most time outdoors at their residence, and Chinese participants spent the least. Because NHAPS had a large percentage of Whites and a very small percentage of Chinese participants (assumed to be some fraction of the 1.7% Asian) and a smaller percentage of black and Hispanic participants, we would expect that the MESA Air population would have a lower amount of time outdoors at their residence. NHAPS asked participants to record every minute for a 24-h period, whereas the MAQ asked participants to record the average time in each location for each season rounded to the nearest hour. Therefore, some differences between NHAPS and MESA Air time–location patterns could be due to this discrepancy in reporting. Because baseline MESA Air data collection occurred 11–15 years after NHAPS, some differences may be due to behavioral changes in the general population over time as the result of changes in technology, weather, socioconomics, and other factors. Participants in MESA Air reported spending more time away from home than those in NHAPS, and this change may be due to changes in lifestyle over this timespan. MESA Air participants spent more time indoors compared with CHAD participants, which is likely due to differences in the Journal of Exposure Science and Environmental Epidemiology (2015), 1 – 7
demographics of these two cohorts. For participants at least 65 years of age, the time–location patterns were more similar, indicating that the differences between MESA Air and CHAD participants are likely because MESA Air participants are older. Although this study provides reliable estimates of the time– location patterns for a diverse, older population, there are some limitations to this analysis. As the population in MESA is an older population, we cannot use these data to understand time– location in younger groups. The MAQ asked participants to round their time to the nearest hour, so short-term events, even if typical, may not be captured by the MAQ. This approach may result in an underestimation of time spent in certain locations (e.g., home outdoors) and an overestimation of time spent in others (e.g., home indoors). This analysis was limited to a single administration of the MAQ and did not take into account any changes over time. In addition, too few Chinese participants completed TLDs, so we were not able to assess the MAQ in comparison with the second instrument for this population. Overall, the MAQ has provided useful and reliable estimates of the time spent in multiple microenvironments for the MESA Air cohort, and having information on the amount of time participants spend inside and outside at different locations allows for more precision in air pollution concentration estimates. Future work will assess differences in the time–location patterns among demographic groups and investigate how various individuallevel characteristics impact total time spent in specific microenvironments. CONFLICT OF INTEREST The authors declare no conflict of interest.
ACKNOWLEDGEMENTS This publication was developed under a STAR research assistance agreement, No. RD831697 (MESA Air) and grant No. RD-83479601-0 (CCAR) by the U.S. Environmental Protection Agency. It has not been formally reviewed by the EPA. The views expressed in this document are solely those of the authors and the EPA does not endorse any products or commercial services mentioned in this publication. Support for MESA is provided by contracts N01-HC-95159, N01-HC-95160, N01-HC-95161, N01-HC-95162, N01-HC-95163, N01-HC-95164, N01-HC-95165, N01-HC-95166, N01HC-95167, N01-HC-95168, and N01-HC-95169 and CTSA UL1-RR-024156. Funding for the MESA Family study is provided by grants R01-HL-071051, R01-HL-071205, R01HL-071250, R01-HL-071251, R01-HL-071252, R01-HL-071258, R01-HL-071259, and UL1-RR-025005 by the National Center for Research Resources, Grant UL1RR033176, and is now at the National Center for Advanced Translational Sciences, Grant UL1TR000124. Additional support was provided from the National Institute of Environmental Health Sciences through grants K24ES013195, P50ES015915, and P30ES07033.
REFERENCES 1 NRC. Human Exposure Assessment for Airborne Pollutants: Advances and Opportunities. The National Academies Press: Washington, DC: National Research Council, 1991.
© 2015 Nature America, Inc.
Time–location patterns: MESA Air Spalt et al
7 2 Sexton K, Ryan PB. Assessment of human exposure to air pollution: methods, measurements, and models. In: Watson AY, Bates RR, Kennedy D (eds). Air Pollution, the Automobile, and Public Health. National Academy Press: New York, 1988, 207–238. 3 Allen RW, Adar SD, Avol E, Cohen M, Curl CL, Larson T et al. Modeling the residential infiltration of outdoor PM(2.5) in the Multi-Ethnic Study of Atherosclerosis and Air Pollution (MESA Air). Environ Health Perspect 2012; 120: 824–830. 4 Chen C, Zhao B. Review of relationship between indoor and outdoor particles: I/O ratio, infiltration factor and penetration factor. Atmos Environ 2011; 45: 275–288. 5 Allen R, Larson T, Sheppard L, Wallace L, Liu L-JS. Use of real-time light scattering data to estimate the contribution of infiltrated and indoor-generated particles to indoor air. Environ Sci Technol 2003; 37: 3484–3492. 6 Jo W-K, Park K-H. Concentrations of volatile organic compounds in the passenger side and the back seat of automobiles. J Expo Anal Environ Epidemiol 1999; 9: 217. 7 Leung P-L, Harrison RM. Roadside and in-vehicle concentrations of monoaromatic hydrocarbons. Atmos Environ 1999; 33: 191–204. 8 Westerdahl D, Fruin S, Sax T, Fine PM, Sioutas C. Mobile platform measurements of ultrafine particles and associated pollutant concentrations on freeways and residential streets in Los Angeles. Atmos Environ 2005; 39: 3597–3610. 9 Chan CC, Ozkaynak H, Spengler JD, Sheldon L. Driver exposure to volatile organic compounds, carbon monoxide, ozone and nitrogen dioxide under different driving conditions. Environ Sci Technol 1991; 25: 964–972. 10 Adar SD, Davey M, Sullivan JR, Compher M, Szpiro A, Liu LJ. Predicting Airborne Particle Levels Aboard Washington State School Buses. Atmos Environ 2008; 42: 7590–7599. 11 Adar SD, Gold DR, Coull BA, Schwartz J, Stone PH, Suh H. Focused exposures to airborne traffic particles and heart rate variability in the elderly. Epidemiology 2007; 18: 95–103. 12 Behrentz E, Sabin LD, Winer AM, Fitz DR, Pankratz DV, Colome SD et al. Relative importance of school bus-related microenvironments to children’s pollutant exposure. J Air Waste Manag Assoc 2005; 55: 1418–1430. 13 Zhang KM, Wexler AS, Zhu YF, Hinds WC, Sioutas C. Evolution of particle number distribution near roadways. Part II: the ‘Road-to-Ambient’process. Atmos Environ 2004; 38: 6655–6665. 14 Dockery DW, Pope CA 3rd, Xu X, Spengler JD, Ware JH, Fay ME et al. An association between air pollution and mortality in six U.S. cities. N Engl J Med 1993; 329: 1753–1759. 15 Miller KA, Siscovick DS, Sheppard L, Shepherd K, Sullivan JH, Anderson GL et al. Long-term exposure to air pollution and incidence of cardiovascular events in women. N Engl J Med 2007; 356: 447–458. 16 Pope CA 3rd, Burnett RT, Thun MJ, Calle EE, Krewski D, Ito K et al. Lung cancer, cardiopulmonary mortality, and long-term exposure to fine particulate air pollution. JAMA 2002; 287: 1132–1141. 17 Hoek G, Brunekreef B, Goldbohm S, Fischer P, van den Brandt PA. Association between mortality and indicators of traffic-related air pollution in the Netherlands: a cohort study. Lancet 2002; 360: 1203–1209. 18 Ostro B, Lipsett M, Reynolds P, Goldberg D, Hertz A, Garcia C et al. Long-term exposure to constituents of fine particulate air pollution and mortality: results from the California Teachers Study. Environ Health Perspect 2010; 118: 363–369.
19 Gehring U, Heinrich J, Krämer U, Grote V, Hochadel M, Sugiri D et al. Long-term exposure to ambient air pollution and cardiopulmonary mortality in women. Epidemiology 2006; 17: 545–551. 20 Rosenlund M, Berglind N, Pershagen G, Hallqvist J, Jonson T, Bellander T. Longterm exposure to urban air pollution and myocardial infarction. Epidemiology 2006; 17: 383–390. 21 Puett RC, Hart JE, Yanosky JD, Paciorek C, Schwartz J, Suh H et al. Chronic fine and coarse particulate exposure, mortality, and coronary heart disease in the Nurses' Health Study. Environ Health Perspect 2009; 117: 1697–1701. 22 Graham SE, McCurdy T. Developing meaningful cohorts for human exposure models. J Expo Sci Environ Epidemiol 2004; 14: 23–43. 23 US EPA Exposure factors handbook 2011 edition. National Center for Environmental Assessment: Washington, DC, USA, 2011. Report No. EPA/600/R-09/052 F. 24 EPA US Descriptive Statistics Tables from a Detailed Analysis of the National Human Activity Pattern Survey (NHAPS) Data. U.S. Environmental Protection Agency: Las Vegas, NV, 1996. 25 Leech JA, Nelson WC, Burnett RT, Aaron S, Raizenne ME. It's about time: a comparison of Canadian and American time-activity patterns. J Expo Anal Environ Epidemiol 2002; 12: 427–432. 26 Setton E, Marshall JD, Brauer M, Lundquist KR, Hystad P, Keller P et al. The impact of daily mobility on exposure to traffic-related air pollution and health effect estimates. J Expo Sci Environ Epidemiol 2011; 21: 42–48. 27 Kaufman JD, Adar SD, Allen RW, Barr RG, Budoff MJ, Burke GL et al. Prospective study of particulate air pollution exposures, subclinical atherosclerosis, and clinical cardiovascular disease: The Multi-Ethnic Study of Atherosclerosis and Air Pollution (MESA Air). Am J Epidemiol 2012; 176: 825–837. 28 US EPA Total Risk Integrated Methodology (TRIM) Air Pollutants Exposure Model Documentation (TRIM.Expo/APEX, version 4.5), Volume II: Technical Support Document. EPA: Research Triangle Park, NC, USA, 2012. Report No. EPA-452/B-12-001b. 29 US EPA SHEDS-Multimedia Model Version 3 Technical Manual. U.S. Environmental Protection Agency: Washington, DC: 2008. 30 Nethery E, Leckie SE, Teschke K, Brauer M. From measures to models: an evaluation of air pollution exposure assessment for epidemiological studies of pregnant women. J Occup Environ Med 2008; 65: 579–586. 31 Wu XM, Fan ZT, Ohman-Strickland P. Time-location patterns of a population living in an air pollution hotspot. J Environ Public Health 2010; 2010: 625461. 32 US EPA Data Sources Available for Modeling Environmental Exposures in Older Adults. EPA: Washington, DC, USA, 2011. Report No. EPA/600/R-12/013. 33 Bild DE, Bluemke DA, Burke GL, Detrano R, Diez Roux AV, Folsom AR et al. Multiethnic study of atherosclerosis: objectives and design. Am J Epidemiol 2002; 156: 871–881. 34 Cohen MA, Adar SD, Allen RW, Avol E, Curl CL, Gould T et al. Approach to estimating participant pollutant exposures in the Multi-Ethnic Study of Atherosclerosis and Air Pollution (MESA Air). Environ Sci Technol 2009; 43: 4687–4693. 35 Lin K-H. Revisiting the gap between stylized and diary estimates of market work time. Soc Sci Res 2012; 41: 380–391. 36 Juster FT, Ono H, Stafford FP. An assessment of alternative measures of time use. Sociol Methodol 2003; 33: 19–54. 37 Bonke J. Paid work and unpaid work: diary information versus questionnaire information. Soc Indic Res 2005; 70: 349–368.
Supplementary Information accompanies the paper on the Journal of Exposure Science and Environmental Epidemiology website (http:// www.nature.com/jes)
© 2015 Nature America, Inc.
Journal of Exposure Science and Environmental Epidemiology (2015), 1 – 7
www.nature.com/jes
ORIGINAL ARTICLE
Time–location patterns of a diverse population of older adults: the Multi-Ethnic Study of Atherosclerosis and Air Pollution (MESA Air) Elizabeth W. Spalt1, Cynthia L. Curl1, Ryan W. Allen2, Martin Cohen1, Sara D. Adar3, Karen H. Stukovsky4, Ed Avol5, Cecilia Castro-Diehl6, Cathy Nunn7, Karen Mancera-Cuevas8 and Joel D. Kaufman1,9,10 The primary aim of this analysis was to present and describe questionnaire data characterizing time–location patterns of an older, multiethnic population from six American cities. We evaluated the consistency of results from repeated administration of this questionnaire and between this questionnaire and other questionnaires collected from participants of the Multi-Ethnic Study of Atherosclerosis and Air Pollution (MESA Air). Participants reported spending most of their time inside their homes (average: 121 h/ week or 72%). More than 50% of the participants reported spending no time in several of the location options, including at home outdoors, at work/volunteer/school locations indoors or outdoors, or in “other” locations outdoors. We observed consistency between self-reported time–location patterns from repeated administration of the time–location questionnaire and compared with other survey instruments. Comparisons with national cohorts demonstrated the differences in time–location patterns in the MESA Air cohort due to differences in demographics, but the data showed similar trends in patterns by age, gender, season, and employment status. This study was the first to explicitly examine the time–location patterns in an older, multiethnic population and the first to add data on Chinese participants. These data can be used to inform future epidemiological research of MESA Air and other studies that include diverse populations. Journal of Exposure Science and Environmental Epidemiology advance online publication, 29 April 2015; doi:10.1038/jes.2015.29 Keywords: epidemiology; personal exposure; population-based studies
INTRODUCTION A standard method for assessing the individual-level exposure to environmental pollutants is to calculate the “time-weighted” average of microenvironmental concentrations. This approach can estimate exposure to many different types of pollutants, including air pollutants.1,2 The more variable the pollutant concentrations are across microenvironments, the greater the need to accurately characterize both the pollutant concentrations and the time spent in each location. Air pollutant concentrations can be quite variable across proximal microenvironments. Ambient-generated particulate matter concentrations can decrease 50% or more going from outside to inside a home.3–5 Concentrations of particles, volatile organic compounds, nitric oxide, and nitrogen dioxide inside vehicles traveling on roadways have also been found to be significantly higher than ambient levels.6–13 Despite this knowledge, most air pollution epidemiology relies on the assumption that residential outdoor concentrations of air pollutants are good proxies for individual exposure.14–21 This assumption is based on data from studies suggesting that
individuals spend the majority of time at home.22–25 However, failing to characterize time spent in other microenvironments, especially the indoor environment at home, will likely introduce measurement error. Setton et al.26 found that residence-only estimates of air pollution exposure have considerable bias compared with estimates that take into account the time spent working or traveling. State-of-the-art air pollution epidemiology is now moving from simply using outdoor concentration data to incorporating information on person-specific characteristics to more accurately estimate individual exposures.27 However, this approach requires more accurate microenvironment time–location data. Several large-scale studies of time–location patterns provide time–location data for use in exposure assessments. The National Human Activity Pattern Survey (NHAPS) includes 24-h diary data in 82 different locations for 9,386 participants in 48 states.24 The Consolidated Human Activity Database (CHAD)22 includes NHAPS, along with data from other national and regional studies. Data from these types of large-scale studies or databases can be a useful alternative for collecting study-specific data. Moreover, these data are used in national exposure models including USEPA air pollution exposure models: Air Pollution Exposure Model28
1 Department of Environmental and Occupational Health Sciences, University of Washington, Seattle, Washington, USA; 2Faculty of Health Sciences, Simon Fraser University, Burnaby, British Columbia, Canada; 3Department of Epidemiology, University of Michigan, Ann Arbor, Michigan, USA; 4Department of Biostatistics, University of Washington, Seattle, Washington, USA; 5Department of Preventive Medicine, University of Southern California, Los Angeles, California, USA; 6Department of Medicine, Columbia University, New York, New York, USA; 7Department of Public Health Sciences, Wake Forest University School of Medicine, Winston-Salem, North Carolina, USA; 8Division of Rheumatology, Feinberg School of Medicine, Northwestern University, Chicago, Illinois, USA; 9Department of Epidemiology, University of Washington, Seattle, Washington, USA and 10 Department of Medicine, University of Washington, Seattle, Washington, USA. Correspondence: Elizabeth W. Spalt, Department of Environmental and Occupational Health Sciences, University of Washington, Campus Box 354695, 4225 Roosevelt Way NE, Suite 302, Seattle WA 98105, USA. Tel.: 206-897-1436. Fax: 206-897-1991. E-mail: [email protected] Received 6 May 2014; revised 25 February 2015; accepted 26 February 2015
Time–location patterns: MESA Air Spalt et al
2 and Stochastic Human Exposure and Dose Simulation model (SHEDS).29 However, existing information may not be representative of unique populations. Studies of very specific cohorts (e.g., pregnant women;30 minorities, and populations living below the poverty line31) have found it necessary to collect data on patterns of time–location specific to their populations. These data not only are valuable for the study in which they were collected, but also broaden the population for which the time–location patterns are understood. USEPA has recently made specific recommendations for the collection of additional time–location data for older populations.32 The primary aim of this analysis is to present the time–location questionnaire data of an older, multiethnic cohort of American adults in the Multi-Ethnic Study of Atherosclerosis and Air Pollution (MESA Air).27 We also evaluate the reliability of time– location data by comparison of repeated administration of the questionnaire and across several survey instruments. Finally, we compare the time–location patterns of MESA Air participants to those in national cohorts. This work adds to the literature as the MESA Air cohort is older and more racially/ethnically diverse than other cohorts in which the time–location patterns have previously been described.
was not administered at the same time as the MAQ, results from the previous clinic exam (230-1,122 days prior to the MAQ) were used. Each of these questionnaires is described in more detail below.
MESA Air Questionnaire The MAQ was the primary data collection tool for gathering information on home characteristics relevant to pollutant infiltration efficiencies and about behaviors related to individual exposures. Participants were asked specific questions about their typical time–location patterns in winter and summer and could designate if their patterns were the same in both seasons. For each day of the week, the MAQ included questions documenting time spent (rounded to the nearest hour) in each of the seven locations — home indoors, home outdoors, work/volunteer/school indoors, work/ volunteer/school outdoors, in transit (e.g., car, bike), other indoors, and other outdoors. Participants also designated which days of the week they considered weekends and weekdays. The amount of time by transit mode (e.g., walking/biking, car/taxi, bus, train/subway), road types traveled (e.g., freeways, residential streets), and traffic conditions experienced (e.g., light traffic/moving at the speed limit) were also documented. Additional questions asked about home characteristics related to building type, building age, the presence of an attached garage, and other factors relevant to infiltration. Methods for addressing missing MAQ data are presented in the Supplementary Materials.
Time–Location Diary METHODS Study Population MESA Air is an ancillary study to the Multi-Ethnic Study of Atherosclerosis (MESA33), a long-term study of the progression of cardiovascular disease (CVD) in adults. MESA included 6814 participants from six US communities: Baltimore, MD; Chicago, IL; Forsyth County, NC; Los Angeles, CA; New York, NY; and St. Paul, MN. MESA participants were aged 45–84 years at enrollment between 2000 and 2002, with an approximately equal gender ratio, and were free of recognized CVD at baseline. Four ethnic/racial groups were targeted for inclusion: white, black, Hispanic, and Chinese, but recruitment of racial/ethnic groups varied by study site. All MESA Air participants provided informed consent before participation. The primary aim of MESA Air is to understand the relationship between the individual-level ambient-derived air pollution exposure and the progression of subclinical CVD and the incidence of clinical disease events.27 MESA Air includes 6424 participants, primarily recruited between 2005 and 2007 from the MESA cohort. Additional participants were recruited from a second ancillary study to MESA, the MESA Family Study (n = 490), and directly for MESA Air (n = 257) in three additional areas near existing MESA communities, two in the Los Angeles basin and one near New York City.27 An additional 1127 MESA participants were included in MESA Air analyses but did not complete all aspects of the MESA Air study and were not included in this analysis.
Overview of Participant Questionnaires Every MESA Air participant was asked to complete a comprehensive MESA Air Questionnaire (MAQ, see Supplementary Materials) at recruitment (hereafter referred to as the baseline MAQ) to describe typical activity patterns in the winter and summer. These questions were asked during a face-to-face interview during a MESA clinic exam between 2005 and 2007. The MAQ was repeated up to six times during follow-up phone calls and a later clinic exam; repeated administration was triggered for participants who indicated a major change in lifestyle (change in residence, work or school status, caregiver status, or in household members). At a subsequent exam, the MAQ was also re-administered to a small subset of participants (n = 10%) who did not report a major lifestyle change to assess changes in responses over time. For participants with multiple MAQs, the time between administration ranged from o3 months to 6.7 years. Detailed daily time–location diaries (TLDs, see Supplementary Materials) were also completed by a small subset of participants (n = 89) who took part in a personal air pollution exposure study conducted between October 2006 and July 2008.34 In addition, MESA gathered information on various participant activities including employment and physical activity from all participants via technician-administered Personal History and Physical Activity Questionnaires. Personal History Questionnaires were administered at all clinic exams, whereas Physical Activity Questionnaires were administered at all but one exam. When the Physical Activity Questionnaire Journal of Exposure Science and Environmental Epidemiology (2015), 1 – 7
In the TLDs, the subset of participants undergoing personal sampling recorded hourly time spent in specified locations for each day during two 2-week sampling periods using the following classifications — home indoors, home outdoors, motorized vehicle, work indoors, work outdoors, other indoors, and other outdoors. Season classifications were determined using the average temperature over the specific sampling period. If the average temperature was 418 °C, the season was classified as “warm”; otherwise, the season was “cool”.3 For comparison with the MAQ, we classified “warm” seasons as summer and “cool” seasons as winter. In the event that we were not able to capture two distinct seasons for an individual participant, results were averaged to obtain a single set of results for each participant for a season. Methods for addressing missing TLD data are presented in the Supplementary Materials.
Personal History and Physical Activity Questionnaires The Personal History Questionnaire asked all participants about their current employment status in ten job categories: (1) homemaker, (2) employed full-time, (3) employed part-time, (4) employed — on leave for health reasons, (5) employed — temporarily away from job, (6) unemployed o6 months, (7) unemployed 46 months, (8) retired — not working, (9) retired — working, and (10) retired — volunteering. For comparison with the MAQ, categories 2, 3 and 9 were assigned as working outside of the home, whereas the remaining seven categories were classified as not working outside of the home. In addition, as part of the Physical Activity Questionnaire, participants provided information on time spent in 10 activities: (1) working at their job, (2) conducting volunteer work, (3) walking, (4) doing yard work, (5) traveling via car, subway, or bus, (6) performing household chores, (7) taking care of others, (8) dancing/practicing a sport, (9) doing conditioning activities (e.g., aerobics), and (10) spending time in leisure activities (e.g., reading).
Data Analysis In order to address our primary aim of presenting the time–location data from the MESA Air cohort, we calculated summary statistics for the amount of time spent in each microenvironment based on the MAQ for the cohort as a whole and by demographic group (i.e., age, race/ethnicity, gender, study site, employment status, education, and income, as reported on the Personal History Questionnaire). This cross-sectional analysis included baseline MAQs only. A second aim of this analysis was to compare the time–location patterns obtained from the MAQ to behaviors reported by the same participants on repeated administration of the MAQ and additional survey instruments: the TLD, the Personal History Questionnaire, and the Physical Activity Questionnaire. To accomplish this aim, the time–location data for each participant who completed both a TLD and a MAQ in the same season were compared using Spearman correlation coefficients and Wilcoxon rank-sum tests. For participants who filled out multiple MAQs, we selected © 2015 Nature America, Inc.
Time–location patterns: MESA Air Spalt et al
3 the MAQ with the closest date to the TLD. We also investigated the impact of gender, age, race/ethnicity, and employment status on observed differences between the different data collection methods. Consistency across MAQs, for those without lifestyle changes, was assessed with Spearman correlation coefficients. Agreement between the MAQ and the Personal History and Physical Activity Questionnaires was assessed for several measures using Wilcoxon Rank Sum tests. Direct comparisons between the MAQ and the Physical Activity Questionnaire were not made because participants reported time in locations on the MAQ and time in activities on this second survey. Time spent at work was compared among working and nonworking persons as classified by the Personal History Questionnaire and among persons reporting time spent at work versus none on the Physical Activity Questionnaire. Time spent outdoors at home was then contrasted between participants reporting some versus no time spent working in the yard on the Physical Activity Questionnaire. Finally, a similar contrast was made for time in transit based on transit and walking behaviors reported in the Physical Activity Questionnaire. All MAQs were matched to Personal History and Physical Activity Questionnaires administered at the same clinical exam or the exam prior for instances when the questionnaire was not administered. Finally, we compared average values of MESA Air time–location data to those reported by NHAPS and CHAD. All statistical analyses were conducted using SAS Software Version 9.3 of the SAS System for Windows (Copyright 2010 SAS Institute, Cary, NC, USA).
RESULTS Participant Characteristics A total of 6209 participants completed the time–location section of at least one MAQ with 5996 participants doing so at baseline. Participants completed the MAQ a mean of 1.9 times and a maximum of six times for a total of 9704 MAQs. Of the 9,704 MAQs administered, 7831 (81%) took place in person during MESA clinic exams, and 1873 (19%) occurred during follow-up calls. A total of 5618 participants at baseline answered the question regarding seasonal differences in time–location patterns, and of these, 45% reported the same patterns in winter and summer. Personal History and Physical Activity Questionnaires were only administered at in-clinic exams. For the 7835 MAQs administered at exams, 7819 (99.8%) of the Personal History Questionnaires collected at the same exam included responses to the question on job status. For the Physical Activity Questionnaires administered during the same exam visit as the MAQ, 7680 (98%) of participants provided information on time engaged in various physical activities. Eighty-nine individuals participated in the 2-week personal monitoring sampling campaigns and provided at least one TLD. Eighty of these participants participated in a second sampling campaign and completed two TLDs, for a total of 169 TLDs. Table 1 presents the demographic characteristics of the MESA Air participants who completed baseline MAQs. The average age of participants that completed the baseline MAQ was 65 ± 10 years, and 53% of the population was female. Fifty percent of the cohort worked full or part-time, and another 7% volunteered (data not shown). Demographic characteristics for those that completed the MAQ are very similar to MESA overall,27 and demographic characteristics for participants who completed a baseline MAQ are very similar to those that completed MAQs overall (data not shown). Time–Activity Patterns in MESA Air On average, participants reported spending the majority (≥70%) of their time indoors at home (Table 2). After time indoors at home, the most time was spent at work/volunteer/school locations indoors; however, 450% of the participants reported that they spent no time indoors at work/volunteer/school. More than 50% of participants also reported spending no time in the three outdoor locations. Participants reported spending more time in indoor locations and less time in outdoor locations in winter than © 2015 Nature America, Inc.
Table 1. Characteristics of MESA Air participants who completed MESA Air questionnaires and time–location diaries. Baseline MAQs N = 5,996
TLDs N = 89
Number
%
Number
%
Gender Female Male
3194 2802
53 47
46 43
52 48
Race/ethnicity White Chinese Black, African-American Hispanic
2258 650 1667 1421
38 11 28 24
54 2 23 10
61 2 26 11
Age category 39–44a 45–54 55–64 65–74 75–84 85+
18 976 1967 1815 1079 141
0 16 33 30 18 2
0 11 37 30 9 2
0 12 42 34 10 2
Site Forsyth County New York Cityb Baltimore St. Paul Chicago Los Angelesc
888 1195 721 872 1136 1184
15 20 12 15 19 20
13 21 14 13 13 15
15 24 16 15 15 17
Socioeconomic status High school education or more Family income ≥ $30,000 year − 1 Single family home
4999d 3671e 3266f
84 65 54
84 68 57
94 82 64
Abbreviations: MAQ, MESA Air Questionnaire; Mesa Air, Multi-Ethnic Study of Atherosclerosis and Air Pollution; TLD, time–location diaries. aMESA family included a small number of participants under the age of 45, and 18 completed MAQs. bIncludes Rockland County participants. cIncludes Riverside participants. dValues missing for 13 participants. eValues missing for 340 participants. fValues missing for three participants.
in summer (Table 2). Figure 1 and Supplementary Table 2a and g present time–location patterns by age, gender, race/ethnicity, site, employment status, education, and income. Consistency of MAQ and TLD Reponses There were a smaller proportion of Chinese and Hispanics participants in the personal monitoring campaigns, so there are fewer TLDs captured in these groups (Table 1). TLD participants also tended to have higher incomes and were more likely to have at least a high school education. The age distributions of the TLD participants and the cohort as a whole were roughly similar except that TLD participants were more likely to be between the ages of 55 and 64 years and less likely to be 75–84 years. The average time between administration of the MAQ and the start of the TLD was 398 days (range: 13–966 days). The highest correlations between the TLD and MAQ responses for each individual were for time spent indoors at work and at home (r = 0.61–0.80), and the weakest correlations were for time spent in other locations (r = − 0.04 to 0.28; Table 3). Agreement between the MAQ and TLD for time spent outdoors was higher in the summer than in the winter. Participants tended to report more time indoors on average (based on their MAQ) than was observed during the 2-week personal monitoring period based on their TLDs (data not shown). Percent difference values between the Journal of Exposure Science and Environmental Epidemiology (2015), 1 – 7
Time–location patterns: MESA Air Spalt et al
4 Table 2.
Summary of reported time (h/week, % of total time) by location reported by participants administered the baseline MAQ (n = 5,996).
Location
Season
Mean
Indoor locations Home Home Work/volunteer/school Work/volunteer/school Other Other
Summer Winter Summer Winter Summer Winter
118 125 15 16 13 14
Outdoor locations Home Home Work/volunteer/school Work/volunteer/school Other Other In transit In transit
Summer Winter Summer Winter Summer Winter Summer Winter
6 2 2 1 6 2 8 7
SD
5th percentile
(70%) (74%) (9%) (10%) (8%) (8%)
27 26 20 21 14 15
(16%) (16%) (12%) (12%) (8%) (9%)
75 84 0 0 0 0
(45%) (50%) (0%) (0%) (0%) (0%)
(4%) (1%) (1%) (1%) (4%) (1%) (5%) (4%)
11 6 8 7 11 6 6 6
(6%) (3%) (5%) (4%) (7%) (4%) (4%) (4%)
0 0 0 0 0 0 0 0
(0%) (0%) (0%) (0%) (0%) (0%) (0%) (0%)
50th percentile 119 126 0 0 10 11 0 0 0 0 1 0 7 7
95th percentile
(71%) (75%) (0%) (0%) (6%) (6%)
158 162 50 51 38 39
(94%) (96%) (30%) (30%) (23%) (23%)
(0%) (0%) (0%) (0%) (1%) (0%) (4%) (4%)
29 12 10 6 28 13 18 17
(17%) (7%) (6%) (4%) (17%) (8%) (11%) (10%)
Abbreviation: MAQ, MESA Air Questionnaire.
40% 35% 30% 25% 20% 15% 10% 5% 0%
In Transit
Other Outdoors
Other Indoors
Work Outdoors
Work Indoors
Home Outdoors
Figure 1. Mean percent time spent in six microenvironments by demographic, socioeconomic group, and site. The remaining time (up to 100%) represents the time spent indoors at home. For example, female participants reported spending 74% of their time indoors at home.
Table 3.
Spearman correlation coefficients (with 95% confidence intervals) between MAQ and TLD responses.
Location
Summer (n = 65)a
Winter (n = 78)a
Both (n = 143)a
Indoor locations Home Work/volunteer/school Other All
0.61 0.80 0.25 0.24
(0.42, 0.74) (0.68, 0.87) (0.00, 0.46) (-0.01, 0.45)
0.65 0.68 0.15 0.30
(0.49, 0.76) (0.54, 0.78) (-0.07, 0.36) (0.08, 0.48)
0.63 0.73 0.20 0.35
(0.51, (0.65, (0.04, (0.20,
Outdoor locations Home Work/volunteer/school Other All In transit
0.19 0.28 0.15 0.08 0.31
(−0.06, 0.41) (0.03, 0.49) (−0.10, 0.38) (−0.17, 0.32) (0.06, 0.51)
0.04 0.13 −0.04 0.01 0.49
(−0.18, 0.26) (−0.10, 0.34) (−0.26, 0.18) (−0.21, 0.24) (0.30, 0.64)
0.14 0.20 0.10 0.10 0.39
(−0.02, 0.30) (0.03, 0.35) (−0.06, 0.26) (−0.07, 0.26) (0.24, 0.52)
0.72) 0.80) 0.35) 0.49)
Abbreviations: MAQ, MESA Air Questionnaire; TLD, time–location diaries. aA total of 89 participants completed TLDs and MAQs. Of these, 65 were in the summer and 78 were in the winter.
Journal of Exposure Science and Environmental Epidemiology (2015), 1 – 7
© 2015 Nature America, Inc.
Time–location patterns: MESA Air Spalt et al
5 Table 4.
Average time spent at home for NHAPS and MESA Air participants by season, gender, and race (h/day). All
Time spent indoors at residence NHAPS MESA Air Time spent outdoors at residencea NHAPS MESA Air
65+ Winter Summer Male Female White Black Hispanic Chinese
16.7 19.6 17.3 18.8
2.3 1.7
2.4 1.8
Employed full-time
Employed part-time
Not employed
17.2 17.8
16.3 16.9
15.8 16.7
17.5 17.9
16.7 17.0
16.9 17.0
16.8 17.6
NR 18.6
14.7 14.7
16.4 17.1
19.3 19.3
1.9 1.2
2.4 2.0
2.6 1.8
1.9 1.6
2.3 1.9
2.1 1.5
2.1 1.8
NR 1.1
2.2 1.4
2.1 1.7
2.4 1.9
Abbreviations: MESA Air, Multi-Ethnic Study of Atherosclerosis and Air Pollution; NHAPS, National Human Activity Pattern Survey; NR, not reported. aIncludes only those participants that report some time outdoors at home for both cohorts.
MAQ and TLD for participants who reported on the MAQ having the same time–location patterns in winter and summer were not different from those who reported different seasonal patterns (Wilxon; P40.05). We also examined whether demographic factors affected the strength of the correlations between responses on the MAQ and the TLD. Supplementary Figure 1 shows the Spearman correlation coefficients between the amount of time that each individual reported in each microenvironment on their MAQ and their TLD, by gender, age, race/ethnicity, and employment status. For most locations, men and women provided similarly consistent data on their MAQs and TLDs (Supplementary Figure 1a). However, the responses for the outdoor work location were less consistent for women than men. Although none of the women in the personal monitoring campaign reported any time in this location on the MAQ, some women did report spending time at work outdoors on their TLD. For individuals o 65 years, the 2-week periods captured by the TLDs were more highly correlated with the typical patterns of behavior reported in the MAQs than for older individuals (Supplementary Figure 1b). On average, individuals ≥ 65 years reported more time indoors and less time outdoors on their MAQ compared to their TLD (data not shown). The correlation of 0 for work outdoors was due to personal monitoring participants ≥ 65 years old reporting no time on the MAQ and some time on the TLDs. All race/ethnic groups were relatively consistent in their responses regarding time spent in transit. There were differences in the consistency between the MAQ and TLD for the rest of the locations, but these differences were not significant (Supplementary Figure 1c). Working individuals were more consistent in their responses related to work locations but not other locations (Supplementary Figure 1d). Consistency of Repeated MAQ Administration A total of 394 participants with no reported lifestyle changes filled out the time–location portion of the MAQ to assess changes in responses over time. The average time between administration of the two MAQs was 1621 days (range: 22–2393 days). Correlation between responses for the time–location patterns was high (0.78) with higher correlations for indoors (0.89) than transit (0.40) or outdoors (0.35). Consistency of MAQ Results with other Survey Instruments available in MESA Comparisons were made between the MAQs and other MESA questionnaires using information on employment status and the time spent at work, working in the yard, in transit, and walking. Not surprisingly, working participants reported spending significantly more time at work compared with individuals in the © 2015 Nature America, Inc.
nonworking categories based on the Personal History Questionnaire (means of 30 and 3 h/week, respectively; P o0.0001) and on the Physical Activity Questionnaire (29 and 4 h/week; P o 0.0001). The amount of time spent outdoors at home from the MAQ was also higher among participants who reported time working in the yard on the Physical Activity Questionnaire than those that did not (6 vs 2 h/week; P o 0.0001). Participants who reported some transit time on the Physical Activity Questionnaire reported an average of 8 h per week of transportation time on the MAQ, whereas those who reported no time in transit on the Physical Activity form reported 5 h per week on the MAQ (P o 0.0001). Participants reporting walking in the Physical Activity Questionnaire also had a higher weekly average amount of time in transit (8 h) compared with those who did not report walking (6 h, P o 0.0001). Comparison with Other Studies of Time–Location Patterns Time–location patterns in the MESA Air cohort were compared with other national cohorts. Comparisons of demographic characteristics for MESA Air, NHAPS, and CHAD are provided in Supplementary Table 3. Although the overall time spent at home indoors was greater for MESA Air than NHAPS participants, similar patterns across seasons and among genders and employment status were observed (Table 4). Overall, CHAD participants reported spending less time indoors than MESA Air participants. CHAD participants over the age of 65 years spent a similar amount of time indoors and outdoors compared with MESA Air participants over 65 (Table 5). For both CHAD and MESA Air, males reported spending less time indoors and more time outdoors than females, and white participants reported spending the least time indoors and the most outdoors (Table 5). DISCUSSION In this study, we successfully deployed and evaluated a time– activity questionnaire in a multiethnic cohort for incorporation into epidemiological analyses. The MAQ-collected information on time spent in microenvironments from MESA Air participants, and agreement between the MAQ and other data collection tools provides evidence of the reliability of this survey. The evaluation of repeated administration of the MAQ and the comparison of the MAQ with other data collection tools demonstrated consistency across data collection methods, and increases our confidence that the MAQ captured reliable information on where MESA Air participants spend their time. Although variable for location, MAQ and TLD responses were positively correlated. Comparisons between the MAQ and other questionnaires administered in MESA provided additional evidence that the MAQ succeeded in capturing time–location patterns. Journal of Exposure Science and Environmental Epidemiology (2015), 1 – 7
Time–location patterns: MESA Air Spalt et al
6 Table 5.
Comparison of average total time spent indoors and outdoors between CHAD and MESA Air (h/day). All
Male
Female
65+
Male: 65+
Female: 65+
White
Black
Hispanic
Chinese
Total time indoors CHAD 21.2a MESA Air 21.5
20.7 21.0
21.5 22.0
21.8 21.8
NR 21.3
NR 22.2
21.1 21.2
21.5 21.5
21.4 21.6
NR 22.3
Total time outdoorsb CHAD 2.3a MESA Air 2.4
2.7 2.8
1.9 2.0
2.0 2.3
2.7 2.7
1.5 1.9
2.4 2.7
2.3 2.5
2.3 2.3
NR 1.4
Abbreviations: CHAD, Consolidated Human Activity Database; Mesa Air, Multi-Ethnic Study of Atherosclerosis and Air Pollution; NR, not reported. aWeighted average of male and female participants. bIncludes only those participants that report some time outdoors for both cohorts.
Although TLD and MAQ responses were correlated, there were many differences between the MAQ and the TLD. The TLDs represent a 2-week snapshot in time, whereas the MAQs are intended to capture typical activities over a season. Requesting information on the “typical” amount of time spent in locations is likely to result in the mode for each participant rather than the mean.35,36 For participants with less routine schedules, disparity between the MAQ and TLD is more likely. In addition, because the MESA Air population is older with a large proportion of retired individuals, the time–location pattern for this population is likely less routine than for a working population. If true, we would not necessarily expect any 2-week period to be reflective of what is typical but can expect to find similarities in aggregate responses. Overall, we found greater consistency for locations where participants spend more time than for those locations where less time is spent. Previous researchers found age- and gender-specific differences for the reporting of time spent working.36,37 Similarly, we found some reporting inconsistencies by gender, age, race/ethnicity, and employment status. Results were similar by gender for all categories except for work outdoors. When we investigated responses by age, we found that an agreement between responses to the MAQ regarding typical patterns of behavior and the specific patterns during the 2-week period captured by the TLD was lower for most locations for individuals aged ≥ 65 years. Accuracy of recalled information is likely to decrease as people age, but differences may also be because individuals over the age of 65 are less likely to be working (Wilcoxon, P o0.0001) and therefore may have a less routine schedule. Differences between MESA Air and NHAPS may be due to differences in demographic composition, collection methodology, and/or dates of collection. In MESA Air, white participants spent the most time outdoors at their residence, and Chinese participants spent the least. Because NHAPS had a large percentage of Whites and a very small percentage of Chinese participants (assumed to be some fraction of the 1.7% Asian) and a smaller percentage of black and Hispanic participants, we would expect that the MESA Air population would have a lower amount of time outdoors at their residence. NHAPS asked participants to record every minute for a 24-h period, whereas the MAQ asked participants to record the average time in each location for each season rounded to the nearest hour. Therefore, some differences between NHAPS and MESA Air time–location patterns could be due to this discrepancy in reporting. Because baseline MESA Air data collection occurred 11–15 years after NHAPS, some differences may be due to behavioral changes in the general population over time as the result of changes in technology, weather, socioconomics, and other factors. Participants in MESA Air reported spending more time away from home than those in NHAPS, and this change may be due to changes in lifestyle over this timespan. MESA Air participants spent more time indoors compared with CHAD participants, which is likely due to differences in the Journal of Exposure Science and Environmental Epidemiology (2015), 1 – 7
demographics of these two cohorts. For participants at least 65 years of age, the time–location patterns were more similar, indicating that the differences between MESA Air and CHAD participants are likely because MESA Air participants are older. Although this study provides reliable estimates of the time– location patterns for a diverse, older population, there are some limitations to this analysis. As the population in MESA is an older population, we cannot use these data to understand time– location in younger groups. The MAQ asked participants to round their time to the nearest hour, so short-term events, even if typical, may not be captured by the MAQ. This approach may result in an underestimation of time spent in certain locations (e.g., home outdoors) and an overestimation of time spent in others (e.g., home indoors). This analysis was limited to a single administration of the MAQ and did not take into account any changes over time. In addition, too few Chinese participants completed TLDs, so we were not able to assess the MAQ in comparison with the second instrument for this population. Overall, the MAQ has provided useful and reliable estimates of the time spent in multiple microenvironments for the MESA Air cohort, and having information on the amount of time participants spend inside and outside at different locations allows for more precision in air pollution concentration estimates. Future work will assess differences in the time–location patterns among demographic groups and investigate how various individuallevel characteristics impact total time spent in specific microenvironments. CONFLICT OF INTEREST The authors declare no conflict of interest.
ACKNOWLEDGEMENTS This publication was developed under a STAR research assistance agreement, No. RD831697 (MESA Air) and grant No. RD-83479601-0 (CCAR) by the U.S. Environmental Protection Agency. It has not been formally reviewed by the EPA. The views expressed in this document are solely those of the authors and the EPA does not endorse any products or commercial services mentioned in this publication. Support for MESA is provided by contracts N01-HC-95159, N01-HC-95160, N01-HC-95161, N01-HC-95162, N01-HC-95163, N01-HC-95164, N01-HC-95165, N01-HC-95166, N01HC-95167, N01-HC-95168, and N01-HC-95169 and CTSA UL1-RR-024156. Funding for the MESA Family study is provided by grants R01-HL-071051, R01-HL-071205, R01HL-071250, R01-HL-071251, R01-HL-071252, R01-HL-071258, R01-HL-071259, and UL1-RR-025005 by the National Center for Research Resources, Grant UL1RR033176, and is now at the National Center for Advanced Translational Sciences, Grant UL1TR000124. Additional support was provided from the National Institute of Environmental Health Sciences through grants K24ES013195, P50ES015915, and P30ES07033.
REFERENCES 1 NRC. Human Exposure Assessment for Airborne Pollutants: Advances and Opportunities. The National Academies Press: Washington, DC: National Research Council, 1991.
© 2015 Nature America, Inc.
Time–location patterns: MESA Air Spalt et al
7 2 Sexton K, Ryan PB. Assessment of human exposure to air pollution: methods, measurements, and models. In: Watson AY, Bates RR, Kennedy D (eds). Air Pollution, the Automobile, and Public Health. National Academy Press: New York, 1988, 207–238. 3 Allen RW, Adar SD, Avol E, Cohen M, Curl CL, Larson T et al. Modeling the residential infiltration of outdoor PM(2.5) in the Multi-Ethnic Study of Atherosclerosis and Air Pollution (MESA Air). Environ Health Perspect 2012; 120: 824–830. 4 Chen C, Zhao B. Review of relationship between indoor and outdoor particles: I/O ratio, infiltration factor and penetration factor. Atmos Environ 2011; 45: 275–288. 5 Allen R, Larson T, Sheppard L, Wallace L, Liu L-JS. Use of real-time light scattering data to estimate the contribution of infiltrated and indoor-generated particles to indoor air. Environ Sci Technol 2003; 37: 3484–3492. 6 Jo W-K, Park K-H. Concentrations of volatile organic compounds in the passenger side and the back seat of automobiles. J Expo Anal Environ Epidemiol 1999; 9: 217. 7 Leung P-L, Harrison RM. Roadside and in-vehicle concentrations of monoaromatic hydrocarbons. Atmos Environ 1999; 33: 191–204. 8 Westerdahl D, Fruin S, Sax T, Fine PM, Sioutas C. Mobile platform measurements of ultrafine particles and associated pollutant concentrations on freeways and residential streets in Los Angeles. Atmos Environ 2005; 39: 3597–3610. 9 Chan CC, Ozkaynak H, Spengler JD, Sheldon L. Driver exposure to volatile organic compounds, carbon monoxide, ozone and nitrogen dioxide under different driving conditions. Environ Sci Technol 1991; 25: 964–972. 10 Adar SD, Davey M, Sullivan JR, Compher M, Szpiro A, Liu LJ. Predicting Airborne Particle Levels Aboard Washington State School Buses. Atmos Environ 2008; 42: 7590–7599. 11 Adar SD, Gold DR, Coull BA, Schwartz J, Stone PH, Suh H. Focused exposures to airborne traffic particles and heart rate variability in the elderly. Epidemiology 2007; 18: 95–103. 12 Behrentz E, Sabin LD, Winer AM, Fitz DR, Pankratz DV, Colome SD et al. Relative importance of school bus-related microenvironments to children’s pollutant exposure. J Air Waste Manag Assoc 2005; 55: 1418–1430. 13 Zhang KM, Wexler AS, Zhu YF, Hinds WC, Sioutas C. Evolution of particle number distribution near roadways. Part II: the ‘Road-to-Ambient’process. Atmos Environ 2004; 38: 6655–6665. 14 Dockery DW, Pope CA 3rd, Xu X, Spengler JD, Ware JH, Fay ME et al. An association between air pollution and mortality in six U.S. cities. N Engl J Med 1993; 329: 1753–1759. 15 Miller KA, Siscovick DS, Sheppard L, Shepherd K, Sullivan JH, Anderson GL et al. Long-term exposure to air pollution and incidence of cardiovascular events in women. N Engl J Med 2007; 356: 447–458. 16 Pope CA 3rd, Burnett RT, Thun MJ, Calle EE, Krewski D, Ito K et al. Lung cancer, cardiopulmonary mortality, and long-term exposure to fine particulate air pollution. JAMA 2002; 287: 1132–1141. 17 Hoek G, Brunekreef B, Goldbohm S, Fischer P, van den Brandt PA. Association between mortality and indicators of traffic-related air pollution in the Netherlands: a cohort study. Lancet 2002; 360: 1203–1209. 18 Ostro B, Lipsett M, Reynolds P, Goldberg D, Hertz A, Garcia C et al. Long-term exposure to constituents of fine particulate air pollution and mortality: results from the California Teachers Study. Environ Health Perspect 2010; 118: 363–369.
19 Gehring U, Heinrich J, Krämer U, Grote V, Hochadel M, Sugiri D et al. Long-term exposure to ambient air pollution and cardiopulmonary mortality in women. Epidemiology 2006; 17: 545–551. 20 Rosenlund M, Berglind N, Pershagen G, Hallqvist J, Jonson T, Bellander T. Longterm exposure to urban air pollution and myocardial infarction. Epidemiology 2006; 17: 383–390. 21 Puett RC, Hart JE, Yanosky JD, Paciorek C, Schwartz J, Suh H et al. Chronic fine and coarse particulate exposure, mortality, and coronary heart disease in the Nurses' Health Study. Environ Health Perspect 2009; 117: 1697–1701. 22 Graham SE, McCurdy T. Developing meaningful cohorts for human exposure models. J Expo Sci Environ Epidemiol 2004; 14: 23–43. 23 US EPA Exposure factors handbook 2011 edition. National Center for Environmental Assessment: Washington, DC, USA, 2011. Report No. EPA/600/R-09/052 F. 24 EPA US Descriptive Statistics Tables from a Detailed Analysis of the National Human Activity Pattern Survey (NHAPS) Data. U.S. Environmental Protection Agency: Las Vegas, NV, 1996. 25 Leech JA, Nelson WC, Burnett RT, Aaron S, Raizenne ME. It's about time: a comparison of Canadian and American time-activity patterns. J Expo Anal Environ Epidemiol 2002; 12: 427–432. 26 Setton E, Marshall JD, Brauer M, Lundquist KR, Hystad P, Keller P et al. The impact of daily mobility on exposure to traffic-related air pollution and health effect estimates. J Expo Sci Environ Epidemiol 2011; 21: 42–48. 27 Kaufman JD, Adar SD, Allen RW, Barr RG, Budoff MJ, Burke GL et al. Prospective study of particulate air pollution exposures, subclinical atherosclerosis, and clinical cardiovascular disease: The Multi-Ethnic Study of Atherosclerosis and Air Pollution (MESA Air). Am J Epidemiol 2012; 176: 825–837. 28 US EPA Total Risk Integrated Methodology (TRIM) Air Pollutants Exposure Model Documentation (TRIM.Expo/APEX, version 4.5), Volume II: Technical Support Document. EPA: Research Triangle Park, NC, USA, 2012. Report No. EPA-452/B-12-001b. 29 US EPA SHEDS-Multimedia Model Version 3 Technical Manual. U.S. Environmental Protection Agency: Washington, DC: 2008. 30 Nethery E, Leckie SE, Teschke K, Brauer M. From measures to models: an evaluation of air pollution exposure assessment for epidemiological studies of pregnant women. J Occup Environ Med 2008; 65: 579–586. 31 Wu XM, Fan ZT, Ohman-Strickland P. Time-location patterns of a population living in an air pollution hotspot. J Environ Public Health 2010; 2010: 625461. 32 US EPA Data Sources Available for Modeling Environmental Exposures in Older Adults. EPA: Washington, DC, USA, 2011. Report No. EPA/600/R-12/013. 33 Bild DE, Bluemke DA, Burke GL, Detrano R, Diez Roux AV, Folsom AR et al. Multiethnic study of atherosclerosis: objectives and design. Am J Epidemiol 2002; 156: 871–881. 34 Cohen MA, Adar SD, Allen RW, Avol E, Curl CL, Gould T et al. Approach to estimating participant pollutant exposures in the Multi-Ethnic Study of Atherosclerosis and Air Pollution (MESA Air). Environ Sci Technol 2009; 43: 4687–4693. 35 Lin K-H. Revisiting the gap between stylized and diary estimates of market work time. Soc Sci Res 2012; 41: 380–391. 36 Juster FT, Ono H, Stafford FP. An assessment of alternative measures of time use. Sociol Methodol 2003; 33: 19–54. 37 Bonke J. Paid work and unpaid work: diary information versus questionnaire information. Soc Indic Res 2005; 70: 349–368.
Supplementary Information accompanies the paper on the Journal of Exposure Science and Environmental Epidemiology website (http:// www.nature.com/jes)
© 2015 Nature America, Inc.
Journal of Exposure Science and Environmental Epidemiology (2015), 1 – 7
