Inter-rater reliability of occupational exposure assessment in a case-control study of female breast cancer.

The objective of this paper was to estimate the inter-rater reliability of expert assessments of occupational exposures. An inter-rater reliability sub-study was conducted within a population-based case-control study of postmenopausal breast cancer. Detailed information on lifetime occupational histories was obtained from participants and two industrial hygienists assigned exposures to 185 jobs using a checklist of 293 agents. Experts rated exposure for each job-agent combination according to exposure status (unexposed/exposed), confidence that the exposure occurred (possible/probable/definite), intensity (low/medium/high), and frequency (% time per week). The statistical unit of observation was each job-agent assessment (185 jobs × 293 agents = 54,205 assessments per expert). Crude agreement, Gwet AC1/2 statistics, and Cohen's Kappa were used to estimate inter-rater agreement for confidence and intensity; for frequency, the intra-class correlation coefficient (ICC) was used. The majority of job-agent combinations were evaluated by the two experts to be not exposed (crude agreement >98% of decisions). The degree of agreement between the experts for the confidence of exposure status was Gwet AC1/2 = 0.99 (95% CI: 0.99-0.99), and for intensity, a Gwet AC2 = 0.99 (95% CI: 0.99-0.99). For frequency, an ICC of 0.31 (95% CI: 0.26-0.35) was found. A sub-analysis restricted to job-agent combinations for which the two experts agreed on exposure status revealed a moderate agreement for confidence of exposure (Gwet AC2 = 0.66) and high agreement for intensity (Gwet AC2 = 0.96). For frequency, the ICC was 0.52 (95% CI: 0.47-0.57). A high level of inter-rater agreement was found for identifying exposures and for coding intensity, but agreement was lower for the coding of frequency of exposure.

Workplace measurements by the U.S. Occupational Safety and Health Administration since 1979: Descriptive analysis and potential uses for exposure assessment.

This letter summarizes modifications to the results presented in Lavoué et al. (2012): Lavoué, J., Burstyn, I.,Friesen, M. (2012) Workplace Measurements by the US Occupational Safety and Health Administration since1979: Descriptive Analysis and Potential Uses for Exposure Assessment. Annals of occupational hygiene57(1):77­97. Although several results were altered, the conclusions were not affected by the changes.

Workplace measurements by the US Occupational Safety and Health Administration since 1979: descriptive analysis and potential uses for exposure assessment.

BACKGROUND: Inspectors from the US Occupational Safety and Health Administration (OSHA) have been collecting industrial hygiene samples since 1972 to verify compliance with Permissible Exposure Limits. Starting in 1979, these measurements were computerized into the Integrated Management Information System (IMIS). In 2010, a dataset of over 1 million personal sample results analysed at OSHA's central laboratory in Salt Lake City [Chemical Exposure Health Data (CEHD)], only partially overlapping the IMIS database, was placed into public domain via the internet. We undertook this study to inform potential users about the relationship between this newly available OSHA data and IMIS and to offer insight about the opportunities and challenges associated with the use of OSHA measurement data for occupational exposure assessment. METHODS: We conducted a literature review of previous uses of IMIS in occupational health research and performed a descriptive analysis of the data recently made available and compared them to the IMIS database for lead, the most frequently sampled agent. RESULTS: The literature review yielded 29 studies reporting use of IMIS data, but none using the CEHD data. Most studies focused on a single contaminant, with silica and lead being most frequently analysed. Sixteen studies addressed potential bias in IMIS, mostly by examining the association between exposure levels and ancillary information. Although no biases of appreciable magnitude were consistently reported across studies and agents, these assessments may have been obscured by selective under-reporting of non-detectable measurements. The CEHD data comprised 1 450 836 records from 1984 to 2009, not counting analytical blanks and erroneous records. Seventy eight agents with >1000 personal samples yielded 1 037 367 records. Unlike IMIS, which contain administrative information (company size, job description), ancillary information in the CEHD data is mostly analytical. When the IMIS and CEHD measurements of lead were merged, 23 033 (39.2%) records were in common to both IMIS and CEHD datasets, 10 681 (18.2%) records were only in IMIS, and 25 012 (42.6%) records were only in the CEHD database. While IMIS-only records represent data analysed in other laboratories, CEHD-only records suggest partial reporting of sampling results by OSHA inspectors into IMIS. For lead, the percentage of non-detects in the CEHD-only data was 71% compared to 42% and 46% in the both-IMIS-CEHD and IMIS-only datasets, respectively, suggesting differential under-reporting of non-detects in IMIS. CONCLUSIONS: IMIS and the CEHD datasets represent the biggest source of multi-industry exposure data in the USA and should be considered as a valuable source of information for occupational exposure assessment. The lack of empirical data on biases, adequate interpretation of non-detects in OSHA data, complicated by suspected differential under-reporting, remain the principal challenges to the valid estimation of average exposure conditions. We advocate additional comparisons between IMIS and CEHD data and discuss analytical strategies that may play a key role in meeting these challenges.

Comparison of formaldehyde exposure levels in two multi-industry occupational exposure databanks using multimodel inference.

The recent increase in international multicenter epidemiologic studies of occupational hazards and the current development of risk analysis tools based on multinational exposure data provide an incentive to evaluate the extrapolation of exposure data across countries. We compared formaldehyde measurements in the French (COLCHIC) and U.S. (IMIS) occupational exposure databases for 1986-2001 using multimodel inference. In this approach, conclusions are based on a set of plausible models rather than on a single model. Modeled variables included data source, industry, year, and sample type (short-term or long-term). The model set included 72 models corresponding to testing all variables and 5 interactions and estimation of 93 coefficients. As it is impossible to classify non-detects (ND) as short-term or long-term samples, they were removed from IMIS prior to analysis. Respectively, 3143 and 2646 IMIS and COLCHIC data, spread across 28 common industries, were analyzed. The full model explained 21% of the total variability. All fixed effects and the source-industry interaction were shown as strong predictors of exposure. The temporal trend (5% yearly decrease) and difference between short-term and long-term data (short-term greater than long-term by two-fold) were similar across the two databanks. Industry-specific differences between IMIS and COLCHIC were detected only for health-related activities for which COLCHIC levels were higher (~ four-fold). Sensitivity analyses were conducted to assess the impact of excluding NDs from IMIS. They showed initial predicted industry-specific geometric means for IMIS data potentially multiplied by factors from 0.42 to 0.98 (long-term data) and from 0.11 to 1.4 (short-term data) when NDs were included with various distributions across the short-term and long-term categories. The most realistic scenario yielded 0.82 for long-term (18% overestimation) data and 1.05 (5% underestimation) for short-term data. Although tempered by a probable non-detect bias, our analysis showed that both databanks provide a similar multi-industry portrait of formaldehyde exposure despite a potential for very different occupational settings. Our results offer encouraging insight about extrapolation of exposure data across countries.

Multimodel inference and multimodel averaging in empirical modeling of occupational exposure levels.

Empirical modeling of exposure levels has been popular for identifying exposure determinants in occupational hygiene. Traditional data-driven methods used to choose a model on which to base inferences have typically not accounted for the uncertainty linked to the process of selecting the final model. Several new approaches propose making statistical inferences from a set of plausible models rather than from a single model regarded as 'best'. This paper introduces the multimodel averaging approach described in the monograph by Burnham and Anderson. In their approach, a set of plausible models are defined a priori by taking into account the sample size and previous knowledge of variables influent on exposure levels. The Akaike information criterion is then calculated to evaluate the relative support of the data for each model, expressed as Akaike weight, to be interpreted as the probability of the model being the best approximating model given the model set. The model weights can then be used to rank models, quantify the evidence favoring one over another, perform multimodel prediction, estimate the relative influence of the potential predictors and estimate multimodel-averaged effects of determinants. The whole approach is illustrated with the analysis of a data set of 1500 volatile organic compound exposure levels collected by the Institute for work and health (Lausanne, Switzerland) over 20 years, each concentration having been divided by the relevant Swiss occupational exposure limit and log-transformed before analysis. Multimodel inference represents a promising procedure for modeling exposure levels that incorporates the notion that several models can be supported by the data and permits to evaluate to a certain extent model selection uncertainty, which is seldom mentioned in current practice.

Comparison of indices proposed as criteria for assigning skin notation.

OBJECTIVES: Skin notations are used as a hazard identification tool to flag chemicals associated with a potential risk related to transdermal penetration. The transparency and rigorousness of the skin notation assignment process have recently been questioned. We compared different approaches proposed as criteria for these notations as a starting point for improving and systematizing current practice. METHODS: In this study, skin notations, dermal acute lethal dose 50 in mammals (LD(50)s) and two dermal risk indices derived from previously published work were compared using the lists of Swiss maximum allowable concentrations (MACs) and threshold limit values (TLVs) from the American Conference of Governmental Industrial Hygienists (ACGIH). The indices were both based on quantitative structure-activity relationship (QSAR) estimation of transdermal fluxes. One index compared the cumulative dose received through skin given specific exposure surface and duration to that received through lungs following inhalation 8 h at the MAC or TLV. The other index estimated the blood level increase caused by adding skin exposure to the inhalation route at kinetic steady state. Dermal-to-other route ratios of LD(50) were calculated as secondary indices of dermal penetrability. RESULTS: The working data set included 364 substances. Depending on the subdataset, agreement between the Swiss and ACGIH skin notations varied between 82 and 87%. Chemicals with a skin notation were more likely to have higher dermal risk indices and lower dermal LD(50) than chemicals without a notation (probabilities between 60 and 70%). The risk indices, based on cumulative dose and kinetic steady state, respectively, appeared proportional up to a constant independent of chemical-specific properties. They agreed well with dermal LD(50)s (Spearman correlation coefficients -0.42 to -0.43). Dermal-to-other routes LD(50) ratios were moderately associated with QSAR-based transdermal fluxes (Spearman correlation coefficients -0.2 to -0.3). CONCLUSIONS: The plausible but variable relationship between current skin notations and the different approaches tested confirm the need to improve current skin notations. QSAR-based risk indices and dermal toxicity data might be successfully integrated in a systematic alternative to current skin notations for detecting chemicals associated with potential dermal risk in the workplace.

Monte Carlo simulation to reconstruct formaldehyde exposure levels from summary parameters reported in the literature.

OBJECTIVES: This study presents a procedure allowing the numerical synthesis of exposure data reported in different ways in the literature, including summary parameters and single measurements. The procedure was applied to literature regarding formaldehyde exposure in the reconstituted wood panels industry, including oriented-strand board (OSB), medium density fibre board (MDF) and particle board (PB). METHODS: For each publication providing summary parameters we estimated geometric means (GM) and geometric standard deviations (GSD) by assuming lognormality of exposure levels. Monte Carlo simulation was performed to re-create datasets from the sample sizes and estimated GMs and GSDs, allowing their subsequent formatting together with the single measurements. The precision and bias of the methods used to estimate GMs and GSDs were evaluated. RESULTS: Altogether, the 13 articles included in our study yielded a final database of 874 data, of which 732 were simulated. For both area and personal data, exposures corresponding to MDF and PB were similar while OSB levels were lower. The most recent available personal levels (1985-1994) were highest in PB for jobs performed in the vicinity of the press (GM=0.63 mg m-3). Corresponding area levels were highest for PB in the main production zone (GM=0.43 mg m-3). Mixed-effects models fitted to area PB data explained 38% of the total variability. A 6-fold decrease in exposures from 1965 to 1995 was estimated. Replication of the simulation process yielded relative standard deviations of the calculated GMs and GSDs between 10 and 20%. The relative biases of the methods used to estimate GMs and GSDs varied across methods and decreased with higher sample sizes (from approximately 15% for n=5 to less than 5% for n=30, in absolute value). The precision also varied across methods and improved with higher sample sizes (from approximately 30% for n=5 to approximately 10% for n=30). DISCUSSION: This methodology constitutes a new meta-analysis tool that should improve the interpretation of industrial hygiene literature data, but needs to be further validated.