Validity of generic job–exposure matrices
| Authors, year | Study population | Generic JEM | Comparison measure | Method of assessing validity | Results |
|---|---|---|---|---|---|
| Hinds et al, 198529 | 261 male lung cancer case and 444 population-based controls in Hawaii | Hoar et al., 198020 | Occupational cancer literature | - Consideration of odds ratios and dose-response for known carcinogens | Coal tar/pitch OR = 1.9 |
| Petroleum pitch/tar OR = 2.0 | |||||
| Arsenic OR = 1.2 | |||||
| Chromium OR = 0.9 | |||||
| Asbestos OR = 12.6 | |||||
| Nickel OR = 1.6 | |||||
| Beryllium OR = 1.6 | |||||
| Linet et al, 198730 | 342 chronic lymphocytic leukemia cases and 342 hospital-based controls without cancer, in the US | Hoar et al., 198020; Sieber et al., 199125 | Self and surrogate reported exposures to butadiaene, asbestos, trichloroethylene, styrene, tetrachloroethylene, carbon tetrachloride, and benzene | - Sensitivity and specificity in comparison to self reports | - Sensitivities (benzene) from 0.10 to 0.24; specificities 0.85 to 0.91 |
| - Sensitivities (asbestos) from 0.23 to 0.47; specificities 0.87 to 0.91 | |||||
| - Kappa comparing the two JEMs | - Kappas from 0.01 to 0.60, median = 0.26 | ||||
| Cicioni et al, 199131 | 143 mesothelioma cases and 35753 other cancer controls, excluding lung cancer cases, in the US | Sieber et al., 199125 | Expert classification of asbestos exposure based on occupation and industry of subject | - Comparison of odds ratios for asbestos exposure | - JEM ORs = 2.0 and 2.4 for low and high exposure, respectively |
| - Expert ORs = 1.6 and 6.4 for low and high exposure, respectively | |||||
| Kauppinen et al, 199232 | 344 primary liver cancer and 861 controls with stomach cancer or myocardial infarction, in Finland | Pannett et al., 198521 | Expert (industrial hygienists) classification of exposure based on occupation and industry of subject | - Sensitivity and specificity in comparison to estimates of high exposure according to expert classification | - Sensitivities (for ∼30 agents) from 0.02 to 0.90, median = 0.41 |
| - Specificities 0.84 to 1.0, median = 1.0 | |||||
| Kromhout et al, 199233 | 878 males from Zutphen, The Netherlands, followed up for lung cancer | Hoar et al., 198020; Pannett et al., 198521 | Self reported exposures | - Sensitivity and specificity in comparison to self reports | - Sensitivities (12 agents) from 0 to 0.98, median ∼0.45 |
| - Specificities 0.17 to 1, median ∼0.97 | |||||
| - Kappa comparing the two JEMs | - Kappas from −0.07 to 0.87, median = 0.08 | ||||
| Ahrens et al, 199334; Orlowski et al, 199323 | 391 lung cancer cases and 391 population-based controls in Germany | Ferrario et al., 198822; Orlowski et al., 199323 | Self reported exposures to asbestos in 19 job specific questionnaires | - Kappa comparing the two JEMs, and the JEMs and self reports | - Kappas from 0.44 to 0.67 for inter-JEM comparison, from 0.15 to 0.44 for self-report-JEM comparison |
| Luce et al, 199335 | 616 subjects from a case-control study of sinonasal cancer in France | Ferrario et al., 198822 | Self reported exposure to formaldehyde and wood dust, with duration, intensity, and probability of exposure classified by expert (industrial hygienist) review | - Kappa comparing the JEM to the expert reviewed self reports | - Kappas (formaldehyde ) from 0.17 to 0.24 |
| - Kappas (wood dust) from 0.83 to 0.84 | |||||
| Roeleveld et al, 199336 | parents of 306 mentally retarded children and 322 children with other congenital handicaps (with known causes) from the Netherlands | Hoar et al., 198020; Pannett et al., 198521 | Self reported exposures to 42 agents | - Sensitivity and specificity in comparison to self reports | - Sensitivities from 0.18 to 0.32 |
| - Specificities 0.86 to 0.94 | |||||
| Stengel et al, 199337 | 765 bladder cancer cases and 765 hospital-based controls; 298 cases with glomeronephritis and 298 hospital-based controls in France | Ferrario et al., 198822 | Self reported exposure to organic solvents, with review by experts | - Sensitivity and specificity in comparison to expert reviewed self report | - Sensitivities from 0.23 to 0.63, median = 0.42 |
| - Specificities from 0.87 to 0.98, median = 0.94 | |||||
| - Kappa comparing the JEM to the expert reviewed self reports | - Kappas from 0.29 to 0.45, median = 0.36 | ||||
| Stucker et al, 199338 | 765 bladder cancer cases and 765 hospital-based controls, in France | Ferrario et al., 198822 | Expert classification of polycyclic aromatic hydrocarbon exposure based on occupation and industry of subject | - Sensitivity and specificity in comparison to expert classification | - Sensitivities from 0.13 to 0.96 and specificities from 0.57 to 0.99 depending on dichotomisation, specificity increased as sensitivity decreased |
| Le Moual et al, 199539 | 10046 adults living in one of 7 French cities in 1975 | Pannett et al., 198521; Ferrario et al., 198822 | - Comparison of the two JEMs to each other and to a third “French” JEM developed for this study, considering broad exposure to “dusts, gases, and chemical fumes”, and 28 more specific exposures | - Comparison of odds ratios and trend in exposure-lung function (FEV1) response of the three methods | - All three JEMs showed similar statistically significant decreasing trends in FEV1 with exposure to “dusts, gases, and chemical fumes” |
| - Results for the 28 specific hazards were much more variable | |||||
| McNamee, 199640 | 102 chronic pancreatitis cases and 204 population-based controls from the UK | Cherry et al., 199228 | - Self reported exposures to hydrocarbons using job specific questionnaires, with expert review by hygienists and occupational physicians | - Comparison of odds ratios between three methods | - Expert-reviewed self-reports ORs 1.9 and 3.7 for medium and high exposures respectively |
| - Internal JEM using mean exposure scores from above method for each job | - Internal JEM ORs 1.7 and 2.2 | ||||
| - Generic JEM ORs 2.2 and 1.1 | |||||
| Hawkes and Wilkins, 199741 | 214 agents common to both JEMs | Hoar et al., 198020; Sieber et al., 199125 | - Direct comparison of 2 JEMs, after conversion of all occupation codes to NIOSH-NOHS system, for 54 job groups in metal, paper and wood, and chemical industries | - Kappa comparing 2 JEMs | - Kappas from 0.02 to 0.27 in metal industry occupations |
| - Kappas from −0.07 to 0.24 in paper and wood industry | |||||
| - Kappas from −0.12 to 0.14 in chemical industry | |||||
| Rybicki et al, 199742 | 188 subjects in a US case-control study of neurologic disease, all with some occupational history in manufacturing | Sieber et al., 199125 | - Self reported exposures to copper, lead and iron, with expert review by an industrial hygienist | - Sensitivity and specificity in comparison to expert reviewed self report | - Mean sensitivities from 0 to 0.21 |
| - Mean specificities from 0.86 to 0.93 | |||||
| Tielemans et al, 199943 | subjects of 2 case-control studies of male infertility in the Netherlands | Hoar et al., 198020 | - Urine samples analysed for metabolites of toluene and xylene (n=267) and for chromium (n=156) | - Sensitivity and specificity in comparison to urine samples | - Sensitivities = 0.60 for toluene/xylene, 0.26 for chromium |
| - Specificities = 0.63 for toluene/xylene, 0.79 for chromium | |||||
| - Kappa in comparison to urine samples | - Kappas = 0.13 for toluene/xylene, 0.04 for chromium | ||||
| Louik et al, 200044 | 12188 mothers and 12017 fathers of children in a US case-control study of birth defects | Sieber et al., 199125 | - Two experts (occupational hygienist and physician) assessed exposures to dichlorodifluoromethane, propylene glycol, and amorphous fused silica to ∼200 industry/occupation combinations | - % agreement | - 20%, 3%, and 2% agreement between the JEM and at least one expert for dichlorodifluoromethane, propylene glycol, and amorphous fused silica, respectively |
| Benke et al, 200145 | 838 subjects of a case-control study of glioma in Australia | Kauppinen et al., 199826 | - Self reported exposures to 5 substances | - Kappa | - Kappas from 0 to 0.62 (median = 0.07) in comparison to self-reports |
| - Kappas from 0.07 to 0.46 (median = 0.28) in comparison to expert assessments |