Estimating Population Level Exposure

In this issue of the Annals, Driscoll and colleagues present their methods and results for estimating the prevalence and intensity of occupational exposure to lead, formaldehyde, and polycyclic aromatic hydrocarbons (PAHs) in Australia (Driscoll et al., 2015a,b,c). Over the years, various methods have been used to estimate population-level exposure to a range of workplace hazards, particularly carcinogens. Over 30 years ago NIOSH conducted the National Occupational Exposure Survey to estimate exposure in the United States (Sundin and Fraser, 1989). Walkthrough surveys of 4490 randomly chosen workplaces over 10 employees were conducted to estimate population prevalence to a very large number of chemicals. One challenge of this approach was its reliance on material safety data sheets to identify toxic substances, which resulted in by-products of production, such as PAHs from combustion and raw materials often not being captured. Internationally, the best known population-level exposure surveillance effort has been CAREX (acronym for CARcinogen EXposure), which was developed by the Finnish Institute for Occupational Health (FIOH), in collaboration with the International Agency for Research on Cancer (IARC), as part of an effort to estimate the burden of occupational cancer in the European Union (Kauppinen et al., 2000). CAREX estimated the prevalence of exposure to 85 agents (counting PAHs as a single agent) in 55 industry sectors in each EU member state using assessments from Europe’s leading workplace exposure experts. Since that time, the CAREX model has been applied to other countries, including Costa Rica, which enhanced its assessment of pesticides (Partanen et al., 2003). FINJEM was developed by the FIOH as a general exposure information system for hazard control, risk quantification, and hazard surveillance using that agencies expertise and exposure data (Kauppinen et al., 1998). Unlike previous models, FINJEM was designed to estimate both prevalence and levels of exposure. More recently, the CAREX Canada project was modeled after the original CAREX project, but attempted to adopt some aspects of FINJEM including assigning levels of exposure as well as prevalence (Peters et al., 2015). The methods used by Driscoll and colleagues, which were largely developed for population-based case-control studies, differ in many respects from previous approaches (Driscoll et al., 2015a). The assessment is based on interviews with 4993 people who participated in the Australian Workplace Exposure Study (Carey et al., 2014). The randomly selected working participants received a computer-assisted interview to assess exposure to carcinogens in their current job. Potentially exposed people were assigned to job-specific modules that are part of OccIDEAS, a tool developed for retrospective exposure assessment in epidemiologic studies (Carey et al., 2014). The job-specific modules collected information on the general work environment, specific tasks, and control measures. The OccIDEAS system was then used to apply decision rules that linked response patterns to expert-based estimates of their probability and qualitative level of exposure to 38 known and suspected carcinogens. Previous projects have relied on occupation and industry titles alone to develop exposure estimates. The approach used by Driscoll and colleagues offers the opportunity to identify previously unrecognized exposure circumstances through identifying tasks and using the questions in the relevant job-specific modules to assess exposure. A challenge is conducting a large enough survey with sufficient job-specific modules to identify rarer exposure circumstances. Projects such as these and others that have focused on a single carcinogen (such as WoodEx; Kauppinen et al., 2006) have contributed to prevention by raising awareness of the number of workers potentially impacted by workplace carcinogens. The numbers generated are often quoted in government reports and IARC monographs to provide a gauge of the potential impact of regulations or evaluations, respectively. CAREX and these similar projects are based on the concept of hazard. Thus, prevalence estimates include all workers potentially exposed above background (ambient environmental) levels. However, with the addition of either qualitative or quantitative level of exposure intensity, the data generated by these projects is now more useful for disease surveillance, epidemiology, and risk assessment (Kauppinen et al., 2014). For example, both FINJEM and CAREX data have been used to create job exposure matrixes for epidemiologic purposes (for example, Guo et al., 2004; Veglia et al., 2007; Offermans et al., 2014). There are now efforts underway to compare these different systems with those designed for use with epidemiologic studies (Lavoué et al., 2012). In recent years, a number of projects have used population-level assessments of exposure to estimate the burden of occupational cancer, where the estimates are used to model prevalence, and sometimes intensity, of exposure in the past in order to estimate the proportion of cancers due to these exposures. In particular, CAREX has been used to model exposure incorporating expert assessment of time trends in both the Global Burden of Disease project, and country-specific efforts, such as in the UK (Driscoll et al., 2005; Van Tongeren et al., 2012). Efforts to estimate population-level exposure to occupational hazards, especially carcinogens, have played an important role in raising awareness of toxic substances in the workplace. From a prevention perspective, perhaps the most important use of population-level estimates is their use in the latest generation of burden of cancer projects (Rushton et al., 2008), but assessments from these projects are now being used in a number of applications. Developing new and innovative methods to improve our assessment of population-level exposure, such as those described here, is essential for these efforts to promote prevention, as well as contributing to a wide range of epidemiologic uses.

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.