Ann. occup. Hyg., Vol. 46, No. 8, pp. 649-652, 2002
© 2002 British Occupational Hygiene Society
Published by Oxford University Press
Editorial |
The Babel of Multicenter Exposure Assessment
Institute for Risk Assessment Sciences (IRAS), Utrecht University, Yalelaan 2, De Uithof, 3508TD Utrecht, The Netherlands
Received 14 May 2002; in final form 20 August 2002
INTRODUCTION
Now the whole earth had one language and the same words. And as they migrated from the east, they came upon a plain in the land of Shinar and settled there. And they said to one another, ‘Come, let us make bricks, and burn them thoroughly’. And they had brick for stone, and bitumen for mortar. Then they said, ‘Come, let us build ourselves a city, and a tower with its top in the heavens, and let us make a name for ourselves; otherwise we shall be scattered abroad upon the face of the whole earth’.
The LORD came down to see the city and the tower, which mortals had built. And the LORD said, ‘Look, they are one people, and they have all one language; and this is only the beginning of what they will do; nothing that they propose to do will now be impossible for them. Come, let us go down, and confuse their language there, so that they will not understand one another’s speech’. So the LORD scattered them abroad from there over the face of all the earth, and they left off building the city. Therefore it was called Babel, because there the LORD confused the language of all the earth; and from there the LORD scattered them abroad over the face of all the earth.
(Genesis ch. 11 v. 1–9, New Revised Standard Version of the Bible, © 1989 Division of Christian Education of the National Council of the Churches of Christ in the USA. Used by permission. All rights reserved.)
An international multicenter study is similar in its ambition to the construction of the Tower of Babel, in that skills and resources of many are combined to yield an effort impossible for any one individual research group to undertake alone. Such a combined effort is the ideal scenario when individual centers have the expertise to carry out an investigation, but lack populations large enough to give sufficient power for the study of small risks. Small-scale studies are often praised for the quality of the information they collect, but they tend to be restricted to populations that are too homogeneous and/or too limited in size to make reliable inferences. Pooling data gives sufficient power to observe clearer associations between specific exposures and diseases. However, as the builders of the Tower of Babel found out, large cooperative efforts can prove perilous. Any extra gain in statistical power due to the increase in the studied population can be lost due to substantial misclassification of both exposure and health outcomes. Therefore, the central question is to consider whether a large multicenter venture can achieve a quality of data comparable to factory-based studies. How does one prevent a multicenter international study in occupational epidemiology from disintegrating into a cacophony?
ASSESSMENT METHODOLOGY
Occupational exposure assessment in particular remains the least standardized part of multicenter studies, and the main challenge (Alexon, 1985). It has only begun to receive rigorous mathematical treatment in the last 20 yr, probably as a consequence of both the accumulation of exposure measurement data and the increasing demand for higher certainty in exposure estimates from epidemiologists, risk assessors and other stakeholders such as industry and labor representatives. An outcome is methods that make better use of the resources available to exposure assessors (Seixas and Checkoway, 1995; Cherrie et al., 1996; IARC, 1996; Stewart et al., 1996). International databases (Kauppinen et al., 2000) are likely to further facilitate coherent exposure assessment in multicenter studies, but much work remains to be done in this arena.
Exposure assessment can be seen as consisting of the following tasks: documenting the occupational history of subjects, estimating exposure intensity for the circumstances in the occupational histories and combining the previous two sources of information to obtain an estimate of exposure. In recreating the subject’s occupational history, one must trace how determinants of exposure (job performed, production volume, machinery, personal protective equipment, work schedules, etc.) have changed with the passage of time in the workplaces under investigation. In some epidemiological study designs (e.g. case–control), workers themselves or their next of kin can be interviewed. However, this is not the case in cohort studies, in which case knowledgeable individuals are not always available to answer these questions and their opinions are difficult to validate. In multicenter studies, such data gathering can be standardized through creating classification schemes for determinants of exposure [e.g. the job class system developed for the study of European asphalt workers by Burstyn et al. (2002)] and through designing standard questionnaire forms for conducting the interviews. Back-translations of any classification schemes and questionnaires are essential to ensure that they are uniformly interpreted. Differences between countries in the stage of evolution attained by the profession of occupational hygiene add to the challenge of finding ‘experts’ of comparable qualifications. Therefore, additional training of persons conducting primary data collection seems prudent, although we know little about whether it improves the accuracy of the collected information. Finally, it is essential to investigate both reproducibility and validity of occupational histories collected in multicenter studies, as such data, although crucial to understanding uncertainties in exposure assessment, are generally lacking.
Exposure intensity is best estimated from the available exposure measurements. There has been considerable effort in recent years to standardize sampling and analysis methods (see for example Kenny et al., 1997), but because of past and persisting differences, data that is recovered as part of retrospective exposure assessment in international multicenter studies is often difficult to compare across countries (Burstyn et al., 2000). This is a problem in the construction of measurement databases, but these are still the best available sources of information about past exposure intensities.
HELPING THE EXPERTS
In an attempt to overcome difficulties with sparse and uncertain exposure information, researchers sometimes rely on the opinions of trusted and/or specially trained experts to establish, on the basis of interviews, the exposure status of an individual or a group with respect to the agents of interest. This was the approach developed for hypothesis-generating case–control studies in the general population (Gerin et al., 1985; Siemiatycki et al., 1997). One workplace example was a study of cancer risk among European welders, in which ‘quantitative estimates were derived from consultation of literature sources and of some company data’, relying primarily on the judgement of experts to reconstruct exposures (Gerin et al., 1993). In the example of retrospective exposure assessment for an international cohort exposed to phenoxy herbicides, chlorophenols and dioxins, the investigators relied on a combination of expert judgement, literature review and application of deterministic source–receptor models to reconstruct exposures (Kauppinen et al., 1994). There is a clear need to standardize the approaches of experts used by the different centres; expert assessors must be calibrated. Computers are being applied to this ('tMannetje et al., 2001a) and appear to provide some improvement in agreement in assessments performed by independently working, but centrally trained, assessors ('tMannetje et al., 2001b). However, Kauppinen et al. (2002) found that agreement between assessors was ‘difficult to reach’ during a computerized approach to exposure reconstruction in a multicenter cohort study of workers in the pulp, paper and paper product industries.
Experts must form their opinions based on some knowledge, presumably obtained through familiarity with the studied exposure circumstance, and ultimately, measurements of exposure. For the claim of an expert to superior knowledge is undermined if merely derived from subjective opinion. Thus, exposure measurement data forms the best basis for expert evaluation. An approach was recently proposed to formally combine information from ‘experts’ and exposure measurements by application of a Bayesian framework (Ramachandran and Vincent, 1999; Ramachandran, 2001; Wild et al., 2002). Although promising, it requires further theoretical development before it can be applied in epidemiological studies (Burstyn and Kromhout, 2002, 2003). Earlier, approaches that combined statistical exposure models with expert input were applied in studies of the European man-made mineral fiber (Cherrie and Dodgson, 1986; Dodgson et al., 1987; Cherrie et al., 1996) and asphalt (Burstyn et al., 2002) industries, although the latter placed more emphasis on the exposure measurements than the former.
HOPE FOR THE FUTURE
What is likely to be the future of exposure assessment in multicenter international studies in occupational epidemiology? Greater public concern about small risks and the need of regulators for quantitative risk assessment and identification of specific causative agents have both increased the importance of exposure assessment based on personal exposure measurement data. Three factors will increase the need for multicenter studies using the best possible exposure assessment. First, despite the success of occupational epidemiology in identifying workplace causes of cancer, some subgroups, such as small businesses and dispersed groups in large companies (e.g. cleaners, maintenance workers), have not been extensively studied (Blair et al., 1999). Second, it is also likely that outsourcing and automation will continually reduce the size of the workforce in firms that have been traditionally seen as ‘large’ enterprises. Third, ‘strong’ risk factors have already been identified, leaving occupational epidemiologists with the challenge of discovering and characterizing ‘weak’ associations (Doll, 1996).
International multicenter studies are therefore likely to remain in the forefront of occupational epidemiology and the transfer of methodology and experience from small company-based studies to international ones will remain a major problem of exposure assessment. However, our experience with the study of European asphalt workers leads us to believe that international collaboration can lead to a better quality of exposure assessment in large international studies than in smaller company-based ones. The quality of exposure assessment, first and foremost, depends on the effective application of new and existing quantitative exposure assessment methods. As Stewart et al. (1996) indicated, ‘...historical exposure assessment requires an opportunistic approach, taking advantage of what information is available and developing creative and innovative approaches to exploit that information’. We are convinced that multicenter studies provide wonderful opportunities for successfully implementing this philosophy.
FOOTNOTES
* Author to whom correspondence should be addressed. E-mail: i.burstyn@iras.uu.nl 
REFERENCES
Alexon O. (1985) Dealing with the exposure variable in occupational and environmental epidemiology. Scand J Soc Med; 13: 147–52.[Web of Science][Medline]
Blair A, Rothman N, Zahm SH. (1999) Occupational cancer epidemiology in the coming decades. Scand J Work Environ Health; 25: 419–97.
Burstyn I, Kromhout H. (2002) A critique of Bayesian methods for retrospective exposure assessment. (Letter to the editor and reply.) Ann Occup Hyg; 46: 429–32.
Burstyn I, Kromhout H. (2003) Who qualifies to be an expert? (Letter to the editor and reply.) Ann Occup Hyg; in press.
Burstyn I, Kromhout H, Kauppinen T, Heikkilä P, Boffetta P. (2000) Statistical modeling of determinants of historical exposure to bitumen and polycyclic aromatic hydrocarbons among paving workers. Ann Occup Hyg; 44: 43–56.
Burstyn I, Boffetta P, Kauppinen T et al. (2002) Estimating exposures in asphalt industry for an international epidemiological cohort study of cancer risk. Am J Ind Med; in press.
Cherrie JW, Dodgson J. (1986) Past exposures to airborne fibers and other potential risk factors in the European man-made mineral fiber production industry. Scand J Work Environ Health; 12 (suppl. 1): 26–33.
Cherrie JW, Schneider T, Spankie S, Quinn M. (1996) A new method for structured, subjective assessment of past concentrations. Occup Hyg; 3: 75–83.
Dodgson J, Cherrie JW, Groat S. (1987) Estimates of past exposure to respirable man-made mineral fibres in the European insulation wool industry. Ann Occup Hyg; 31: 567–82.
Doll RS. (1996) Weak associations in epidemiology: importance, detection, and interpretation. J Epidemiol; 6: S11–20.
Gerin M, Siemiatycki J, Kemper H, Begin D. (1985) Obtaining occupational exposure histories in epidemiologic case-control studies. J Occup Med; 27: 420–6.[Web of Science][Medline]
Gerin M, Fletcher AC, Gray C, Winkelmann R, Boffetta P, Simonato L. (1993) Development and use of a welding process exposure matrix in a historical prospective study of lung cancer risk in European welders. Int J Epidemiol; 22 (suppl. 2): S22–8.[Abstract]
IARC. (1996) Conference on retrospectove assessment of occupational exposures in epidemiology, 13–15 April 1994. Occup Hyg; 3(1–3): 1–208.
Kauppinen TP, Pannett B, Marlow DA, Kogevinas M. (1994) Retrospective assessment of exposure through modeling in a study on cancer risks among workers exposed to phenoxy herbicides, chlorophenols and dioxins. Scand J Work Environ Health; 20: 262–71.[Web of Science][Medline]
Kauppinen T, Toikkanen J, Pedersen D et al. (2000) Occupational exposures to carcinogens in the European Union. Occup Environ Med; 57: 10–18.
Kauppinen T, Teschke K, Astrakianakis G et al. (2002) Assessment of exposure in an international study on cancer risks among pulp, paper, and paper product workers. Am Ind Hyg Assoc J; 63: 254–61.
Kenny LC, Aitken R, Chalmers C et al. (1997) A collaborative European study of personal inhalable aerosol sampler performance. Ann Occup Hyg; 41: 135–53.
Ramachandran G. (2001) Retrospective exposure assessment using Bayesian methods. Ann Occup Hyg; 45: 651–67.
Ramachandran G, Vincent JH. (1999) A Bayesian approach to retrospective exposure assessment. Appl Occup Environ Hyg; 14: 547–57.[Medline]
Seixas NS, Checkoway H. (1995) Exposure assessment in industry specific retrospective occupational epidemiology studies. Occup Environ Med; 52: 625–33.
Siemiatycki J, Fritschi L, Nadon L, Gerin M. (1997) Reliability of an expert rating procedure for reptrospective assessment of occupational exposures in community-based case-control studies. Am J Ind Med; 31: 280–6.[Web of Science][Medline]
Stewart PA, Lees PSJ, Francis M. (1996) Quantification of historical exposures in occupational cohort studies. Scand J Work Environ Health; 22: 405–14.[Web of Science][Medline]
’tMannetje A, Fevotte J, Fletcher T, Brennan P. (2001a) A Micrisoft (r) application for expert assessment in multi-centre studies. In Marklund S, editor. X2001 – Exposure Assessment in Epidemiology and Practice. Stockholm: National Institute for Working Life. p.138.
’tMannetje A, Fevotte J, Fletcher T et al. (2001b) Expert assessment: inter-rater agreement in a multi-centre study. In Marklund S, editor. X2001 – Exposure Assessment in Epidemiology and Practice. Stockholm: National Institute for Working Life. p.11.
Wild P, Sauleau EA, Bourgkard E, Moulin J-J. (2002) Combining expert ratings and exposure measurements: a random effect paradigm. Ann Occup Hyg; 46: 479–87.
SELECTED FURTHER READING
Ashford JR. (1958) The design of a long-term sampling programme to measure the hazard associated with an industrial environment. J R Statist Soc Ser A; 3: 333–47.
Burstyn I, Kromhout H. (2000) Are all the members of a paving crew uniformly exposed to bitumen fume, organic vapour and benzo(a)pyrene? Risk Anal; 20: 653–64.[Web of Science][Medline]
Burstyn I, Teschke K. (1999) Studying the determinants of exposure: a review of methods. Am Ind Hyg Assoc J; 60: 57–72.[Web of Science][Medline]
Burstyn I, Kromhout H, Cruise PJ, Brennan P. (2000) Designing an international industrial hygiene database of exposures among workers in the asphalt industry. Ann Occup Hyg; 44: 57–66.
Burstyn I, Boffetta P, Burr GA et al. (2002) Validity of empirical models of exposure in asphalt paving. Occup Environ Med; 59: 620–4.
Hoar S. (1983) Job exposure matrix methodology. J Toxicol Clin Toxicol; 21: 9–26.[Web of Science][Medline]
Kauppinen T, Teschke K, Savela A, Kogevinas M, Boffetta P. (1997) International data base of exposure measurements in the pulp, paper and paper product industries. Int Arch Occup Environ Health; 70: 119–27.[Web of Science][Medline]
Kromhout H, Heederik D. (1995) Occupational epidemiology in the rubber industry: implications of exposure variability. Am J Ind Med; 27: 171–85.[Web of Science][Medline]
Kromhout H, Symanski E, Rappaport SM. (1993) A comprehensive evaluation of within- and between-worker components of occupational exposure to chemical agents. Ann Occup Hyg; 37: 253–70.
Loomis D, Salvan A, Kromhout H, Kriebel D. (1999) Selecting indices of occupational exposure for epidemiologic studies. Occup Hyg; 5: 73–91.
Oldham PD, Roach SA. (1952) A sampling procedure for measuring industrial dust exposure. Br J Ind Med; 9: 112–9.
Rappaport SM. (1991) Assessment of long-term exposures to toxic substances in air. Ann Occup Hyg; 35: 61–121.
Seixas NS, Sheppard L. (1996) Maximizing accuracy and precision using individual and grouped assessments. Scand J Work Environ Health; 22: 94–101.[Web of Science][Medline]
Seixas NS, Robins TG, Rice CH, Moulton LH. (1990) Assessment of potential biases in the application of MSHA respirable coal mine dust data to an epidemiologic study. Am Ind Hyg Assoc J; 51: 534–40.[Web of Science][Medline]
Teschke K, Ahrens W, Andersen A et al. (1999) Occupational exposure to chemical and biological agents in the nonproduction departments of pulp, paper, and paper product mills: an international study. Am Ind Hyg Assoc J; 60: 73–83.[Web of Science][Medline]
Tielemans E, Kupper LL, Kromhout H, Heederik D, Houba R. (1998) Individual-based and group-based occupational exposure assessment: some equations to evaluate different strategies. Ann Occup Hyg; 42: 115–9.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  I. BURSTYN, H-M. KIM, N. CHERRY, and Y. YASUI Metamodels of bias in Cox proportional-hazards and logistic regressions with heteroscedastic measurement error under group-level exposure assessment Ann. Hyg., April 1, 2006; 50(3): 271 - 279. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||