Article Text


Minisymposium 5

Statistics from

Community based job exposure matrices


G. Benke1.1Department of Epidemiology and Preventive Medicine, Monash University, Melbourne, Australia

A job exposure matrix (JEM) can be defined as a cross classification between a list of job titles and a list of agents (which may be chemicals, physical or biological agents, or psychosocial or ergonomic factors) to which persons carrying out the jobs may be exposed. A time axis is often included to account for the temporal changes in exposures in jobs. JEMs can be either industry or community based, with the latter primarily used in community based case–control studies. Community based or general population JEMs have been used widely in occupational epidemiology over the past 25 years. The first reported JEM was published by Reed and Harcourt in 1941, but it was not until the early 1980s that systematic JEMs were first employed. A brief literature review indicates that there are at least 20 community based JEMs currently available for researchers. These JEMs use a range of standard or national job/industry classification schemes, with the ISCO/ISIC and SOC/SIC combinations being the most common. Key issues with the application of these JEMs are the quality of the job/industry information for coding, quality of the coders, and non-differential misclassification resulting from the use of the JEM. Studies that compare JEMs are invaluable in understanding the limitations and possible non-differential misclassification of particular JEMs. An example of a successful community based JEM that has been employed in Europe and Australia is FINJEM. Although FINJEM has been mainly used in cancer studies, recent use in respiratory studies has indicated the versatility of this JEM, a feature of many community based JEMs. With the increased use and understanding of JEMs now available, refinement and reduction in misclassification should be achieved in the next decade.


H. Kromhout1, R. Vermeulen2, J. P. Zock3.1Institute for Risk Assessment Sciences, Utrecht University, Utrecht, The Netherlands; 2Occupational and Environmental Epidemiology Branch, DCEG, NCI, DHHS, Rockville, MD, USA; 3Respiratory and Environmental Health Research Unit, IMIM, Barcelona, Spain

A revised community based job exposure matrix (ALOHA JEM) that yields estimates for exposure to mineral and biological dusts and to gases and fumes was elaborated for the follow up of the European Respiratory Health Survey (ECRHS) among young adults. The original JEM had a different job axis, and because the follow up study used ISCO-88 codes for job histories, a revised JEM had to be constructed. Two of us (HK, RV) used the same original basic principles, resulting in a more specific than sensitive assignment of exposures. Double coding of Spanish job histories made comparisons between the original ad hoc JEM and the ALOHA JEM possible. Overall, both JEMs generated similar exposure prevalences. However, for the ALOHA JEM a higher estimate was found more often than a lower estimate, implying a somewhat higher sensitivity than the original ad hoc JEM. Differences are related to different structure of the job coding systems, and some were due to changes in insights of occupational exposures in certain sectors. Consequently, the ALOHA JEM was used to create exposure profiles for of the ECRHS II population. Surprisingly, slight positive trends in frequency of exposure were seen over time. Increase in exposure over time (from no to low or high exposure or from low to high exposure) was consequently modelled with logistic regression analyses. For exposure to biological dust, increase in exposure was less likely for men and individuals from s higher socioeconomic class, such as managers and foremen. Older individuals were less likely to increase exposure to biological dust. For exposure to mineral dust, a somewhat different picture emerged, with men and current smokers more likely to increase their exposure during follow up. For health status at the beginning of follow up, only individuals with asthma diagnosed by a physician at the start of follow up were less likely to increase their exposure to biological and mineral dust.


A. ’t Mannetje, D. McLean, N. Pearce.Centre for Public Health Research, Massey University, Wellington, New Zealand

Introduction: In New Zealand, a demand exists from epidemiologists as well as policy makers for a comprehensive and accessible database with national occupational exposure information. A multipurpose information system, such as the Finnish job exposure matrix (FINJEM), would meet this demand. However, few New Zealand specific exposure measurement data exist upon which to build a general population job-exposure matrix (GPJEM). In New Zealand, we therefore rely on existing GPJEM from other countries in combination with national expertise. Here we present the methods for developing a GPJEM for New Zealand (NZJEM).

Methods: Existing GPJEMs from other countries were used as a starting point for the NZJEM. Occupational classifications used in these GPJEM were recoded to the New Zealand Standard Classification of Occupations. The GPJEMs were linked by occupation and assembled in one database. This database provided a simultaneous view of all JEM evaluations of a certain agent for each job, to consider by the experts. The experts then provided their own estimate of the situation in New Zealand. A nationwide survey is currently being conducted in New Zealand, asking a random sample of the working population about exposures they experience at their job. This will be used to validate and further develop the NZJEM.

Results and conclusions: We plan to complete the first version of NZJEM by the end of 2004. We also plan to include the following additional features in NZJEM: (a) demographic data: number of men and women employed in each occupation in New Zealand; (b) direct links with the other GPJEM; (c) inclusion of textual definitions of exposures, occupations, and industries; and (d) easy data finding and report generating possibilities.


K. Karipidis1,2, G. Benke1, M. Sim1, L. Fritschi3, M. Yost4, A. Hughes5, A. Kricker5, B. Armstrong5.1Department of Epidemiology and Preventive Medicine, Monash University, Melbourne, Victoria, Australia; 2Australian Radiation Protection and Nuclear Safety Agency, Yallambie, Victoria, Australia; 3School of Population Health, University of Western Australia, Crawley, WA, Australia; 4Department of Environmental and Occupational Health Sciences, University of Washington, Seattle, USA; 5School of Public Health, University of Sydney, Sydney, NSW, Australia

Introduction: The possibility that exposure to extremely low frequency electric and magnetic fields increases the risk of non-Hodgkin’s lymphoma (NHL) continues to be widely debated. We used data from a population based case–control study to examine the association between NHL and occupational exposure to 50 Hz magnetic fields.

Methods: The study population consisted of 704 incident cases of NHL and 694 controls matched by age, sex, and region of residence. A detailed occupational history was taken from each subject. Magnetic field exposure was assessed using two separate job exposure matrices (JEMs). JEM 1 was constructed from workplace magnetic field measurements in the USA. JEM 2 was constructed in Sweden. The cumulative exposure distribution was divided into quartiles and the odds ratios (OR) were calculated using the lowest quartile as the referent group.

Results: In JEM 1, subjects in the second, third, and upper quartiles of cumulative exposure compared with those in the lowest quartile had odds ratios of 0.84 (95% confidence interval 0.62 to 1.13), 1.16 (0.86 to 1.56) and 1.26 (0.94 to 1.7) respectively. In JEM 2 subjects in the second, third, and upper quartile compared with those in the lowest quartile had odds ratios of 1.01 (0.75 to 1.36), 1.34 (0.99 to 1.8), and 1.14 (0.84 to 1.53) respectively.

Conclusions: These preliminary findings provide weak support for the possibility that occupational exposure to 50 Hz magnetic fields increases the risk of NHL.

View Abstract

Request permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.