Several studies have examined methods of accounting for the effects of error associated with the use of ambient particulate matter (PM) concentrations as proxies for personal PM exposure.1–3 Investigators from the National Morbidity and Mortality and Air Pollution Study,4 for example, compared regression calibration and multistage Poisson regression. Although such strategies are potentially useful, the comparison relied on estimates of the cross-sectional association between personal and ambient PM10 concentrations in a convenience sample of 5 panel studies representing 292 participants from 4 geographic locations. A 5-study convenience sample is potentially problematic, as nonrandom study selection may provide biased inferences.5 Also, cross-sectional PM correlations may be weaker than longitudinal, within-person PM correlations, due to interindividual variation in behaviors influencing exposure.6–8
There has not been a systematic and quantitative review of studies of the ambient-personal PM2.5 correlation and, perhaps more importantly, the modifying effects of participant, study and environment characteristics. Thus, there are no summary estimates of the correlation that could be used to adjust for the error inherent in using ambient PM2.5 concentration as proxies for personal PM exposure. To address this gap, we systematically and quantitatively reviewed the literature estimating within-participant, ambient-personal PM2.5 correlations and determined the extent and sources of measurement error inherent in using ambient PM2.5 as a surrogate for personal exposure. These results will facilitate quantification of bias resulting from the use of ambient PM2.5 as a proxy for personal exposure in a Women's Health Initiative (WHI) ancillary study, the Environmental Epidemiology of Arrhythmogenesis in the WHI.
Systematic Review Strategy
We searched for studies of the within-participant, ambient-personal or outdoor-personal PM2.5 correlation considering all document types, languages, and publication dates. On 12 November 2007 we searched PubMed (1950 to date), ISI Web of Science (1955 to date), ISI BIOSIS Previews (1969 to date), CSA Environmental Sciences and Pollution Management (1967 to date), Toxline (1965 to date), and Proquest Dissertations & Theses (1861 to date). STN EMBASE (1974 to date) was searched on 14 December 2007.
We used the following strategy to search PubMed: (PM 2.5 OR PM2.5 OR PM25 OR PM 25 OR fine particle*) AND (ambient OR outdoor OR outdoors OR outside OR exterior OR external OR background OR fixed site*) AND (individual OR personal) AND (correlat* OR associat* OR relat* OR compar* OR pearson OR spearman). The same 4 sets of keywords were adapted for input into Web of Science, BIOSIS, Environmental Sciences, Toxline, and EMBASE. The Dissertations & Theses search required only the first 3 sets of keywords to create a result set small enough for review.
Citations were downloaded to an electronic reference manager (EndNote X1, Thomson Reuters), duplicates were removed, and secondary references were added. The citations were independently reviewed with respect to 3 inclusion criteria: measurement of ambient PM2.5, measurement of personal PM2.5, and estimation of the within-participant, ambient-personal PM2.5 correlation. We extracted study, participant and environment characteristics from all articles meeting inclusion criteria. Study characteristics included journal of publication, publication date, setting, study dates, sample size, duration, timing (consecutive, nonconsecutive), lower limit of PM2.5 detection, number (minimum, mean) of paired PM2.5 measures, and correlation metric (Pearson, Spearman). Participant characteristics included age (mean, minimum, maximum), percent female, and the presence of comorbidities (pulmonary, cardiovascular, multiple, none). Environmental characteristics included the mean, median and standard deviation of PM2.5 concentrations (ambient, personal); the within-participant, ambient-personal PM2.5 correlation coefficients and corresponding number of paired measurements; season; average distance to ambient monitor; monitor type; air exchange rate; percent of time using air conditioning; and percent of time with windows open. Discrepant exclusions and extractions were adjudicated by consensus. We requested supplemental data from authors by electronic mail as needed. City-specific longitudes and latitudes were obtained from the GEOnet Names Server (http://earth-info.nga.mil/gns/html/whatsnew.htm#C3). Meteorologic data were obtained from the National Climatic Data Center (http://www.ncdc.noaa.gov/oa/climateresearch.html).
Uniform measures of association for the jth study were estimated from the personal-ambient PM2.5 correlations measured within each of the ith participants. We converted each within-participant correlation coefficient (ri) to its variance-stabilizing, Fisher's z-transform
Estimates of the within-participant variance
and between-participant variance
for the jth study were estimated from the number of paired personal-ambient PM2.5 measurements for each participant (ni), the number of participants per study (kj), the weighted sum of squared errors (Qj) = ∑i=1k(ni − 3)(Zri − Zri)2, and a constant
. The transformed effect size for the jth study is given by
with weights (wi)
, standard errors (Sj)
and study-specific weights
.10 Negative τ2 estimates were set to 0. Fixed-effects summary estimates were approximated using the median correlation coefficient and the average number of paired measurements for 2 studies11,12 that did not provide participant-specific correlation coefficients.
Funnel plot asymmetry was assessed using a plot of Wj versus j, the adjusted rank correlation and regression asymmetry tests,13,14 and a nonparametric “trim and fill” method that imputes hypothetically missing results due to publication bias.15 In the absence of publication bias, plots of Wj versus j usually resemble a symmetrical funnel with the more precise estimates forming the spout and the less precise estimates forming the cone, while low P values associated with the former tests (PBegg; PEgger) give evidence of asymmetry.
Interstudy heterogeneity was evaluated using a plot16 of
and Cochran's Q test.17 The plot and test are related, in that the position of the jth study along the vertical axis illustrates its contribution to the Q test statistic. In the absence of heterogeneity, all studies fall within the 95% confidence limits and for the Cochran Q test PCochran > 0.1.
Variation in the strength and precision of j across levels of the study, environment, and participant characteristics was first assessed by estimating a summary random-effects estimate of within each study, environment and participant category.18 We also constructed a series of univariable random-effects meta-regression models to relate each study, environment, and participant characteristic to differences in . Lastly, a multivariable random-effects meta-regression model and a backwards-elimination strategy were used to evaluate 10 characteristics of study, participant, and environment routinely available in epidemiologic studies of PM2.5 health effects: latitude, longitude, presence of comorbidities, mean age, percent female, mean ambient PM2.5, relative humidity, sea level pressure, and mean temperature. Interval-scale characteristics were analyzed before and after dichotomization at their medians unless noted otherwise. All analyses were performed using STATA (College Station, TX). To facilitate interpretation, estimates of were back-transformed to their original metric after data analysis.
Our systematic review identified 567 candidate studies for screening. Of these studies, 18 (3%) met criteria for critical appraisal and were abstracted. Abstracted studies were published between 1999 and 2008 (Table 1). The studies they described were set in 17 North American and European cities, 10 states or provinces and 4 countries, with 68% performed in the U.S. (eFigure, http://links.lww.com/EDE/A363). The studies were conducted between 1995 and 2002. The mean study duration was 2.0 months (range 0.9 to 18.2), a period in which 79% of the studies collected PM2.5 data over consecutive days. During data collection, the studies recorded an average of 8 (range 5 to 20) ambient and personal PM2.5 concentration pairs per participant on which their Pearson (37%) and Spearman (63%) correlation coefficients were based (Table 1).
The studies represented 619 nonsmoking participants aged 6–93 (median = 70) years, 60% of whom were female and 41% of whom did not report chronic pulmonary or cardiovascular disease (Table 2). Ambient PM2.5 concentrations (range 8.3 to 25.2 μg/m3) were lower than personal PM2.5 concentrations (range 9.3 to 28.6 μg/m3) overall, with a median personal-ambient PM2.5 difference of 0 (range −9.0 to 16.3) (Table 3). The estimated (median 0.54; range 0.09 to 0.83) and its standard deviation (median 0.12; range 0.04 to 0.31) varied widely (Table 3, Fig. 1), the latter reflecting variability in sample weights (median 82.7; range 10.3 to 552.0). Estimates of ri were similarly variable among studies (median interquartile range (IQR) 0.38; range 0.22 to 1.04), as were temperature (range −6.0 to 24.6°C) and relative humidity (range 44% to 87%), especially when comparing medians from single-season studies (44%).
Figure 2 provides a funnel plot of j, which suggests little evidence of asymmetry. This result is consistent with PBegg = 0.9 and PEgger = 0.2, but the “trim and fill” method imputed 4 hypothetically missing studies with r <0.15. Figure 3, a Galbraith plot in which 12 correlation coefficients (44%) fell outside the 95% confidence bounds, provided strong evidence of heterogeneity. This evidence was consistent with PCochran < 0.001.
The interquartile range (IQR) of ri (range dichotomized as ≥0.41 vs. < 0.41) was the characteristic associated with the greatest difference in : −0.37 (95% CI = −0.53 to −0.20) (Fig. 4). Other factors associated with increased were studies conducted in Europe, studies with eastern longitudes in North America, higher mean ambient PM2.5 concentrations, and higher relative humidity, although imprecision was noted. After restricting to North American studies, given the considerable heterogeneity by study locale and small number of European studies (n = 2), higher mean ambient PM2.5 concentrations and higher relative humidity were the only characteristics predictive of in multivariable meta-regression models (P < 0.05).
Surveys of human activity patterns suggest that people spend more than 85% of their time indoors,19 where they are exposed to numerous sources of PM2.5, with physical and chemical properties and toxicities that often differ from those of ambient PM2.5.20,21 Thus, estimates of personal PM2.5 exposure based on ambient concentrations are associated with some degree of uncertainty. This has led to the suggestion that epidemiologic studies should use ambient PM as a surrogate only for outdoor PM exposure, not total exposure.22–26 Nonetheless, certain studies are often cited to justify using ambient PM2.5 concentrations as proxies for total personal PM2.5 exposures.8,27–29 These studies report strong within-participant ambient-personal PM2.5 correlations. Other studies, which report ambient-personal PM2.5 correlations as low as 0.1030 are rarely cited.
Motivated by this apparent pattern of citations, we reviewed studies of the within-participant correlation between ambient and personal PM2.5, examining them for publication bias and heterogeneity. We found low potential for publication bias, although the trim and fill analysis imputed 4 hypothetically missing studies. These hypothetically missing studies most likely represent unpublished findings because they differed considerably from the majority of the published literature, they could not be used to justify reliance on ambient PM2.5 as a proxy for personal PM2.5 exposure, or they were considered implausible. An alternative explanation is that they represent studies this meta-analysis did not identify, although this is less likely since we reviewed 7 electronic reference databases, evaluated secondary sources, and did not apply any document type, language, or publication starting-date limitations. Indeed, this systematic review had all the features of a meta-analysis deemed necessary to ensure its sensitivity.31–33
Although there was little evidence for pronounced publication bias, we found strong evidence for heterogeneity in —evidence that contraindicated the estimation of a single summary measure to represent the entire literature. The direct associations between European locales and eastern longitudes in North America with may reflect regional factors including higher urban PM2.5 concentrations34 or closer proximities to regulatory monitors. Furthermore, the direct associations of ambient PM2.5 concentrations and relative humidity with suggest an increased contribution of ambient PM2.5 to personal exposures through activity patterns or increased air exchange. Regional differences in geographic, household, and personal factors may explain the indirect association between variation in ri and , but further investigation was limited because these factors were uncommon, uncollected, or inconsistently reported. Similarly, we were unable to determine whether small ranges in personal or ambient PM2.5 concentrations were associated with , as few studies reported participant-specific concentrations.
We did not find a strong association between temperature and , but the investigation included several multiseason studies. On the other hand, the scope of our investigation was limited by exclusion of 12 studies of the cross-sectional ambient-personal PM2.5 correlation. Cross-sectional correlations are thought to be weaker than longitudinal, within-person correlations due to interindividual variation in activities affecting exposure (eg, spending time near smokers, cooking, or cleaning).6–8 A series of studies conducted in the Netherlands also found that ambient-personal PM correlations were stronger when analyses were conducted longitudinally.35 Because studies of within- versus between-participant correlations address systematically different questions—the recognition of which precludes simultaneous evaluation36—the a priori exclusion of cross-sectional correlations was appropriate.
We were unable to determine whether associations based on summary data were good proxies for associations estimated using individual participant data.37 One method to assess the validity of our conclusions and eliminate the potential for ecologic bias38 would have been to evaluate individual participant data. A meta-analysis based on individual participant data also would allow for increased flexibility in analyses of heterogeneity and greater consistency of reporting.39 Although such data were unavailable, the findings reported here were based on a large number of studies and have been interpreted cautiously.
The present meta-analysis focused on PM2.5 although several European countries also regulate PM10. It remains unclear whether findings for PM2.5 extend to PM10. The current meta-analysis also did not evaluate the association between ambient and personal concentrations of sulfate or elemental carbon, although these combustion products may better represent the influence of outdoor particles because their indoor sources are uncommon.40–42 Nonetheless, the results presented here have potentially important implications for studies examining the health effects of PM2.5 because methods for modifying regression equations to account for normally distributed measurement error are well established. Although the uniformity with which these results can be applied across study designs deserves additional consideration,3,6–8,43 Crooks et al44 recently described a Bayesian method for incorporating log-normal measurement error in a cross-sectional study of PM health effects. Log-normal distributions are believed more appropriate for PM, but the application of these methods requires knowledge of the conditional distribution of personal exposure given ambient exposure (specifically, the mean and standard deviation of the personal exposure distribution) as well as the ambient-personal PM2.5 correlations described here.
Limitations notwithstanding, the present report reinforces the view that characteristics of participants, studies and the environments in which the studies are conducted affect the accuracy of ambient PM2.5 as a proxy for personal exposure. The wide range in estimated correlations between personal and ambient PM2.5, as well as the associations with participant, study, and environment characteristics, suggest that the potential for exposure misclassification can be substantial. Thus, these factors warrant greater scrutiny in studies utilizing ambient PM2.5 as a proxy for personal exposure.
We thank Carry Croghan (US EPA) for assisting the authors in providing additional data analyses used in this report. Although this work has been reviewed by EPA and approved for publication, it may not necessarily reflect official Agency policy.
1. Brauer M, Brumm J, Vedal S, Petkau AJ. Exposure misclassification and threshold concentrations in time series analyses of air pollution health effects. Risk Anal
2. Yeh S, Small MJ. Incorporating exposure models in probabilistic assessment of the risks of premature mortality from particulate matter. J Expo Anal Environ Epidemiol
3. Zeger SL, Thomas D, Dominici F, Samet JM, Schwartz J, Dockery D, Cohen A. Exposure measurement error in time-series studies of air pollution: concepts and consequences. Environ Health Perspect
4. Dominici F, Zeger SL, Samet JM. A measurement error model for time-series studies of air pollution and mortality. Biostatistics
5. Hunter JE, Schmidt FL. Methods of Meta-Analysis. Correcting Error and Bias in Research Findings
. 2nd ed. Thousand Oaks, CA: Sage Publications Inc; 2004.
6. Janssen NA, Hoek G, Brunekreef B, Harssema H, Mensink I, Zuidhof A. Personal sampling of particles in adults: relation among personal, indoor, and outdoor air concentrations. Am J Epidemiol
7. Janssen NA, Hoek G, Harssema H, Brunekreef B. Childhood exposure to PM10: relation between personal, classroom, and outdoor concentrations. Occup Environ Med
8. Janssen NA, Hoek G, Harssema H, Brunekreef B. Personal exposure to fine particles in children correlates closely with ambient fine particles. Arch Environ Health
9. Fisher RA. Statistical Methods for Research Workers
. Edinburgh: Oliver and Boyd; 1925.
10. Field AP. Meta-analysis of correlation coefficients: a Monte Carlo comparison of fixed- and random-effects methods. Psychol Methods
11. Kim D, Sass-Kortsak A, Purdham JT, Dales RE, Brook JR. Associations between personal exposures and fixed-site ambient measurements of fine particulate matter, nitrogen dioxide, and carbon monoxide in Toronto, Canada. J Expo Sci Environ Epidemiol
12. Reid CM. Assessment of Exposure to Selected Criteria Pollutants for Two Sensitive Population Cohorts in Atlanta, Georgia. Emory University. Atlanta, Georgia. 2003.
13. Begg CB, Mazumdar M. Operating characteristics of a rank correlation test for publication bias. Biometrics
14. Egger M, Davey Smith G, Schneider M, Minder C. Bias in meta-analysis detected by a simple, graphical test. BMJ
15. Duval S, Tweedie R. Trim and fill: A simple funnel-plot-based method of testing and adjusting for publication bias in meta-analysis. Biometrics
16. Galbraith RF. A note on graphical presentation of estimated odds ratios from several clinical trials. Stat Med
17. Cochran WG. The combination of estimates from different experiments. Biometrics
18. Berkey CS, Hoaglin DC, Mosteller F, Colditz GA. A random-effects regression model for meta-analysis. Stat Med
19. Robinson J, Nelson WC. National human activity pattern survey data base. Research Triangle Park, NC: USEPA; 1995.
20. Monn C, Becker S. Cytotoxicity and induction of proinflammatory cytokines from human monocytes exposed to fine (PM2.5) and coarse particles (PM10–2.5) in outdoor and indoor air. Toxicol Appl Pharmacol
21. Wainman T, Zhang J, Weschler CJ, Lioy PJ. Ozone and limonene in indoor air: a source of submicron particle exposure. Environ Health Perspect
22. Mage D, Wilson W, Hasselblad V, Grant L. Assessment of human exposure to ambient particulate matter. J Air Waste Manag Assoc
23. McBride S, Williams R, Creason J. Bayesian hierarchical modeling of personal exposure to particulate matter. Atmospheric Environment
24. Wallace L, Williams R, Rea A, Croghan C. Continuous week long measurements of personal exposures and indoor concentrations of fine particles for 37 health-impaired North Carolina residents for up to four seasons. Atmospheric Environment
25. Wallace L, Williams R, Suggs JC, Jones P. Estimating contributions of outdoor fine particles to indoor concentrations and personal exposures: effects of household characteristics and personal activities. Washington, DC: US EPA Environmental Protection Agency; 2006.
26. Wilson WE, Suh HH. Fine particles and coarse particles: concentration relationships relevant to epidemiologic studies. J Air Waste Manag Assoc
27. Janssen NA, de Hartog JJ, Hoek G, Brunekreef B, Lanki T, Timonen KL, Pekkanen J. Personal exposure to fine particulate matter in elderly subjects: relation between personal, indoor, and outdoor concentrations. J Air Waste Manag Assoc
28. Sarnat JA, Koutrakis P, Suh HH. Assessing the relationship between personal particulate and gaseous exposures of senior citizens living in Baltimore, MD. J Air Waste Manag Assoc
29. Sarnat JA, Schwartz J, Catalano PJ, Suh HH. Gaseous pollutants in particulate matter epidemiology: confounders or surrogates? Environ Health Perspect
30. Adgate JL, Ramachandran G, Pratt GC, Waller LA, Sexton K. Longitudinal variability in outdoor, indoor, and personal PM2.5 exposure in healthy non-smoking adults. Atmospheric Environment
31. Moher D, Cook DJ, Eastwood S, Olkin I, Rennie D, Stroup DF. Improving the quality of reports of meta-analyses of randomised controlled trials: the QUOROM statement. Quality of Reporting of Meta-analyses. Lancet
32. Oxman AD, Guyatt GH, Singer J, Goldsmith CH, Hutchison BG, Milner RA, Streiner DL. Agreement among reviewers of review articles. J Clin Epidemiol
33. Shea BJ, Grimshaw JM, Wells GA, et al. Development of AMSTAR: a measurement tool to assess the methodological quality of systematic reviews. BMC Med Res Methodol
34. Rom WN, Markowitz S. Environmental and Occupational Medicine. 4th ed. Philadelphia: Lippincott Williams & Wilkins; 2006.
35. Janssen NA, Lanki T, Hoek G, et al. Associations between ambient, personal, and indoor exposure to fine particulate matter constituents in Dutch and Finnish panels of cardiovascular patients. Occup Environ Med
36. Bland JM, Altman DG. Calculating correlation coefficients with repeated observations: Part 2–Correlation between subjects. BMJ
37. Lau J, Ioannidis JP, Schmid CH. Summing up evidence: one answer is not always enough. Lancet
38. Morgenstern H. Uses of ecologic analysis in epidemiologic research. Am J Public Health
39. Stewart LA, Tierney JF. To IPD or not to IPD? Advantages and disadvantages of systematic reviews using individual patient data. Eval Health Prof
40. Brown KW, Sarnat JA, Suh H, Coull BA, Spengler JD, Koutrakis P. Ambient site, home outdoor and home indoor particulate concentrations as proxies of personal exposure. J Environ Monitor
41. Sarnat JA, Brown KW, Schwartz J, Coull BA, Koutrakis P. Ambient gas concentrations and personal particulate matter exposures: implications for studying the health effects of particles. Epidemiology
42. Sarnat SE, Coull BA, Schwartz J, Gold DR, Suh HH. Factors affecting the association between ambient concentrations and personal exposures to particles and gases. Environ Health Perspect
43. Schwartz J, Dockery DW, Neas LM. Is daily mortality associated specifically with fine particles? J Air Waste Manag Assoc
44. Crooks JL, Whitsel EA, Quibrera PM, Catellier DJ, Laio D, Smith RL. The effect of ignoring interpolation error on the inferred relationship between ambient particulate matter exposure and median RR interval in post-menopausal women. Epidemiology
45. Ebelt ST, Petkau AJ, Vedal S, Fisher TV, Brauer M. Exposure of chronic obstructive pulmonary disease patients to particulate matter: relationships between personal and ambient air concentrations. J Air Waste Manag Assoc
46. Williams R, Suggs J, Creason J, et al. The 1998 Baltimore Particulate Matter Epidemiology-Exposure Study: part 2. Personal exposure assessment associated with an elderly study population. J Expo Anal Environ Epidemiol
. 2000;10(6 pt 1):533–543.
47. Evans GF, Highsmith RV, Sheldon LS, et al. The 1999 Fresno particulate matter exposure studies: comparison of community, outdoor, and residential PM mass measurements. J Air Waste Manag Assoc
48. Rodes CE, Lawless PA, Evans GF, et al. The relationships between personal PM exposures for elderly populations and indoor and outdoor concentrations for three retirement center scenarios. J Expo Anal Environ Epidemiol
49. Suh H, Koutrakis P, Chang L. Characterization of the composition of personal, indoor, and outdoor particulate exposures. Report to the California Air Resources Board. Harvard School of Public Health, Environmental Science and Engineering Program, Boston, MA. 2003. Contract No. 98–330.
50. Liu LJ, Box M, Kalman D, et al. Exposure assessment of particulate matter for susceptible populations in Seattle. Environ Health Perspect
51. Williams R, Suggs J, Rea A, et al. The Research Triangle Park particulate matter panel study: PM mass concentration relationships. Atmos Environ
52. Noullett M, Jackson PL, Brauer M. Winter measurements of children's personal exposure and ambient fine particle mass, sulphate and light absorbing components in a northern community. Atmos Environ
53. Wu CF, Jimenez J, Claiborn C, Gould T, Simpson CD, Larson T, Liu LJS. Agricultural burning smoke in Eastern Washington: Part II. Exposure assessment. Atmos Environ
Supplemental Digital Content
© 2010 Lippincott Williams & Wilkins, Inc.