Article Text
Abstract
Objectives To design and construct a standardised tool to provide exposure information associated with commonly used asbestos products and their related tasks in New South Wales (NSW), Australia.
Methods Asbestos dust exposure measurements taken during workplace inspections in the 1970s and 1980s were collected and stored in an exposure database. Measurements were assigned to specific asbestos product and task groups and divided into two sampling periods 1970–1979 and 1980–1989.
Results A total of 1578 asbestos air measurements collected from WorkCover and Dust Diseases Board company records were entered into a custom built exposure database. An asbestos-specific exposure matrix (ASTEM) was constructed in Microsoft Excel 2000, consisting of 3 axes incorporating 12 tasks, 8 asbestos products and the 2 time periods based on 872 individual measurements extracted from the exposure database. Each matrix cell contains the mean asbestos exposure levels measured in fibres/ml, 5th and 95th percentiles and number of data points in the set.
Conclusion An ASTEM has been developed which provides exposure levels for different task/product combinations. When used in conjunction with a detailed occupational history, it will improve exposure estimates of a worker's cumulative asbestos exposure.
- Asbestos
- matrix
- exposure
Statistics from Altmetric.com
For certain asbestos-related diseases the likelihood that dust exposure has made a significant contribution to disease development increases with exposure.1 Determining a worker's cumulative asbestos exposure is therefore critical in the assessment of compensation claims for these diseases. In order to quantitatively assess a worker's cumulative exposure, detailed information regarding the type and frequency of tasks performed and the types of asbestos product handled is required. Historical task exposure levels can then be used to calculate past exposure of individual workers. Industrial hygienists with expert knowledge in asbestos exposure are often relied upon to provide exposure assessments. The latter formulated using a limited pool of published and unpublished reference exposure levels. As a result hygienist estimates can vary dramatically depending on the reference sources selected for their calculations. Other shortcomings associated with using hygienists for exposure assessments include: shortage of experienced expert hygienists, cost and time constraints.
The aim of this study was to develop an asbestos-specific task exposure matrix (ASTEM) based on quantitative measurements from historical workplace reports. Task exposure matrices (TEMs) are an extension of the original job exposure matrix (JEM) methodology developed in the early 1980s2 JEMs use job titles to analyse possible exposures and have been designed for general population and industry specific assessments.3 4 The advantage of JEMs is that they allow a rapid low cost method of exposure assessment when only limited occupational information is available.5 Their weakness however, is their inability to account for variability between workers within an occupation as well as day-to-day variability an individual might experience in the same job.6 7 8 TEMs overcome this limitation by taking into consideration the variability of exposure within a job title and therefore reduce the potential for non-differential misclassification inherent in the JEM method.7
To date, no study has attempted to review Australian asbestos exposure measurements for the purpose of providing reference exposure levels that may assist with determining a worker's retrospective asbestos exposure. Our newly developed ASTEM is a summary of average asbestos air measurements associated with particular task and product combinations. Exposure is calculated by using the proportion of time spent on each task multiplied by the ASTEM value for each task and product combination. All resulting values are added to provide a cumulative asbestos exposure for a worker in fibre/ml×year (f/ml.year) units. Thus, a detailed occupational history is essential to the process. The purpose of this paper is to describe the process behind the construction of the ASTEM.
Methods
The New South Wales (NSW) Dust Diseases Board (DDB) is a NSW Government statutory authority, which provides compensation to workers with dust diseases attributable to dust exposure while employed in NSW. Compensation is awarded after the diagnosis of an occupational lung disease has been made and it has been established that sufficient dust exposure has occurred to contribute to disease development. WorkCover NSW is also a statutory authority, which administers and enforces compliance with occupational health and safety,9 injury management, return to work and workers compensation legislation.
Existing files on NSW companies held at the DDB and WorkCover were searched for workplace asbestos fibre measurements. These files were originally kept on companies required to pay a ‘dust levy’ to the Board. From time to time companies were inspected by the former Department of Labour and Industry (now WorkCover) representatives and air monitoring was performed to ensure compliance with workplace exposure standards. As such, asbestos fibre measurements found in these files are known to have been taken by a small group of trained industrial hygienists, using comparable sampling methods and equipment.
A dust exposure measurement database was created in Microsoft Access from the task descriptions and measurements extracted from the files. Details entered into the database included descriptions of the activity being monitored, the date samples were taken, the sample type (ie, personal or atmospheric), time of sample (ie, short term, long term or time-weighted average (TWA)), and the flow rate with which the sample was collected. Complete details were not available for all measurements. Sample time and flow rate in particular were often not recorded, however in the case of sample time this was often evident from the description of how the measurements were performed and the type of activity being monitored. Company information such as industry, location and type of product manufactured was also entered into the database together with any description of processes related to the handling of asbestos.
All air samples were collected and analysed in accordance with the appropriate National Occupational Health & Safety Commission (NOHSC) standards of the time.10 In brief, a sample was collected by drawing a measured quantity of air through a membrane filter by means of a sampling pump. The filter is later transformed from an opaque membrane into a transparent, optically homogeneous specimen. The fibres are then sized and counted, using a phase contrast microscope and eyepiece graticule. The result is expressed as fibres per ml (f/ml) of air, calculated from the number of fibres on the filter and the measured volume of air sampled.
Short-term sampling, mostly 10 min each filter, but sometimes 30 min, was the main method used until about 1975. The samplers were manually held by the hygienist close to the nose of the worker and reported as breathing zone samples. After about 1975, personal sampling was introduced and the sampler was attached on the worker with the sampling filter in the breathing zone. The result was reported as a personal, breathing zone sample. In very dusty jobs the filters would be changed and consecutive sampling was used so that the filters could be counted. If there was too much dust on a filter then the sample would be rejected. (E Francis, personal communication, 2006. Industrial hygienist and former NSW WorkCover Inspector.)
Based on sampling times, measurements were grouped into two categories as either (1) task-based measurements (long term or short term) or (2) full shift TWAs. For the purpose of the matrix only the task-based measurements were analysed. The main reason for this was that they formed the majority of our measurements while only 17% of measurements were full shift TWAs. A TWA can be calculated from a task-based measurement if the duration with which the task is performed is known. Our intention was to combine the matrix measurements with information obtained from detailed work histories, which specify task duration. Atmospheric or static samples were also excluded as there was generally no information given as to the distance from the source or the emission pathway. Thus all measurements in the database represent personal exposures in f/ml.
The need an asbestos-specific matrix was reinforced by the lack of task associated fibre ml measurements in the published literature. PubMed searches were performed for Australian and International data. Published task exposure measurements were entered into a separate part of the database and reference details were recorded. Our only criteria were that results had to be published post 1970 and recorded in fibres per ml. Where published data was presented as a range, we converted this to a geometric mean (GM) using the following equation GM=exp(ln(a)+ln(b)/2) (where a and b are the range's minimum and maximum data points).11Sampling methods and dates were also recorded where available. In addition to PubMed searches, case reports prepared by NSW industrial hygienists for the DDB were reviewed for asbestos fibre measurements. In the past these reports were requested by the DDB to assist with the processing of compensation claims. Regular crosschecks were performed to avoid duplication from the different sources.
The asbestos task exposure matrix was created in a Microsoft Excel 2000 spreadsheet. Measurements were assigned to product or task groups according to task description and industry and then to one of two time periods: 1970–1979 or 1980–1989. These time frames were selected to incorporate the majority of our measurements in equal periods and to reflect the effect of the change in NSW exposure standards from 4 f/ml to 2 f/ml, which occurred in 1978.12 Earlier data were excluded from the matrix as prior to 1970 measurements were recorded in millions of particles per cubic foot using a very different sampling technique and cannot be accurately converted to fibres/ml.13 14
Results
A total of 1578 asbestos air measurements collected from WorkCover and DDB company records were entered into the exposure database. After excluding measurements that did not meet our criteria, (ie, recorded in f/ml, personal sampling, task based and collected between 1970 and 1989) we were left with 1010 samples. For the purpose of the matrix we only wanted to select tasks that would be common to many different occupations and asbestos products that we knew were frequently used and for which we had the most data. Thus the number of samples falling into our 19 task–product combinations chosen for the matrix was 872. Of these 76% fell into the first time period 1970–1979 (table 1).
A search of the published literature and DDB held industrial hygienist reports identified a total 263 measurements (35% from hygienist reports) matching our matrix's 19 task–product combinations. Dates, sampling times and sampling methods and number of samples were not available for the majority of measurements. Approximately 30% of published data had to be converted from range to GM as discussed in the methods section.
The complete matrix is presented as table 2 and has three components. The columns list 8 asbestos products, the rows list 12 asbestos-related tasks. For DDB data the rows are further divided into two time periods (ie, 1970–1979 or 1980–1989). Each matrix cell contains the arithmetic mean exposure levels for each task in f/ml and their 5th and 95th percentiles. The number of samples from which these values were derived appears in the right hand side of each cell.
Tasks that one would expect a high value, such as spraying asbestos insulation and installing asbestos lagging, did show large values (99.7 f/ml and 9.9 f/ml, respectively). Similarly, tasks that were not expected to be of a dusty nature were low, such as handling brakes and clutches and cement products (each 0.8 fibres/ml in 1980–1989). We observed that the calculated exposure values decreased between the two time periods for all tasks. This was expected as the exposure standard decreased across this period leading to changes to workplace practices in order to achieve compliance. It is also possible that asbestos content had decreased in certain products. Due to limited detail in publications our literature data could not be separated into the two distinct time periods, however in the majority of cases higher results were calculated from literature data than from DDB data.
An example of the ASTEM in use is given in table 3. We have used the point estimate of exposure for each task in this calculation. In estimating exposure for compensation purposes it is probably more appropriate to use upper 95% CI for each task to prevent underestimation of exposure given the high variability in exposure measures and inability to account for control measures.
Discussion
An ASTEM has been developed which provides exposure levels for different task and product combinations. The advantage of this matrix is that it includes a large number of asbestos fibre measurements that have come from NSW worksites and were collected by a small group of trained industrial hygienists using comparable sampling methods and equipment. When used in conjunction with a detailed occupational history collected at interview, improved estimates of a worker's lifetime asbestos exposure may be possible.
Simple reliance upon published exposure data for quantitative exposure estimates in health studies has been found to be limited, due to the inconsistency of authors in reporting historical exposures.15 Studies relying on historical data faced similar problems of poor quality and missing data as well as incomplete reporting. There have been a number of reports describing the design and usage of exposure databases for the purpose of evaluating worker's health risks.16 17 18 These studies reported as we did, that use of a database with automated features allowed flexible data entry, faster collation of results and a reduction of opportunities for human error such as with exposure calculations and assignment of products and tasks.
In recent years a number of studies have employed TEMs to assess disease risk related to occupational exposures. A study investigating an association between respiratory function and bauxite exposure in miners employed the use of a bauxite specific TEM.19 Aluminium smelter workers with respiratory symptoms were reviewed using the TEM approach in relation to occupational exposure to several different occupational substances.20 A TEM was also used to estimate exposure in former workers from a beryllium processing facility using routine monitoring data collected at the beryllium processing plant.21 A nickel specific matrix was designed to assess exposure in a cohort of Norwegian nickel refinery workers in light of the known association with respiratory cancer in a variation on the conventional TEM.22
Most studies faced similar problems related to missing data, particularly for earlier years when samples were either not collected or of questionable quality. Various approaches in overcoming this included: calculating a mean-based measurement from the earliest years available, using a multiplication factor in a retrograde calculation or simply retaining estimates from the most recent period. We are fortunate in having data available over a relatively long period of time. We did not have adequate data available for exposure prior to 1970; the TEM will hence be less suitable for those with significant periods of exposure prior to 1970.
The ASTEM produced in this study is unique in that it is the first consolidated summary of a significant number of asbestos dust measurements taken from NSW workplaces. One major advantage is that samples were collected in a uniform manner by a small number of staff from the same organisation. Due to relatively limited information we were unable to restrict our measurements to those collected in routine monitoring as others have done.17 Generally, the worksites from which measurements were collected did not perform this kind of workplace surveillance. This may mean that our measurements are non-representative or rather more likely to be ‘worst case scenario’ as Department of Labour and Industry inspections were sometimes conducted in response to employee complaints about working conditions. The majority of readings have come from the manufacturing industry, which may introduce problems when applying them to other industries as conditions may have differed. Control measures in place were not recorded in most instances and may partly account for the variability in measurements taken for similar tasks. Most samples were taken in factory environments and therefore may have limited applicability to other situations.
Our review of dust measurements in the literature highlighted difficulties associated with using this information. Only 55% of total DDB dust measurements fell (N=872) into our selected matrix cells however this was still over three times as much data as could be found in published sources (N=263). In the majority of task product combinations, the calculated average exposure was a lot higher in literature measurements than our NSW data. This may be due to the fact that the measurements from the literature were taken in earlier years or in different countries where dust controls and exposure standards differed. It may also be a consequence of different sampling techniques being employed. It would be impossible to draw any conclusions from such a poorly described data set except to suggest that our measurements appear by comparison to not be gross overestimates - due, for example to sampling bias of workplaces with the highest exposure.
We chose to base our matrix on task-based rather than full shift TWA measurements primarily due to a lack of TWA data. In other studies where TEMs have been constructed using TWAs, tasks monitored have mostly been performed for the duration of a full shift. Unfortunately, the descriptions of TWA measurements in the company reports from which our data were collected rarely included information on task duration within the monitored period. Thus our concern was that tasks of short-term duration with high exposures could potentially be misrepresented by these TWAs. Nieuwenhuijsen et al who looked at dust exposure in flour mills shared this concern and argued one of the benefits of using short-term task exposures was that they avoided problems of dilution of exposure intensity which can occur when full shift average exposure is calculated.23 A limitation of using task-based matrices is that in instances where there are a limited number of exposure measurements there is a risk that due to intraday and interday variability in work practices, samples may not adequately reflect the range of worker exposures for particular tasks.24 Another problem associated with task approach is categorising multiple tasks measurements. This has been partly addressed by grouping tasks within the matrix based on the more common multitask samples. Use of the TEM depends on good data on tasks and duration of tasks performed. This will usually come from interviews with workers. The quality of data recalled by the worker will be an important factor to consider in the confidence to be placed in the resulting cumulative exposure estimates.
In the next stage of this study the ASTEM will be used to estimate workers exposure to asbestos using information collected from face-to-face occupational histories. Job specific questionnaires have been designed for over 20 different job titles with a specific focus on the duration and frequency of tasks involving contact with asbestos as well as start year and end year of each asbestos-related task. Questions about the use of personal protective equipment, work place ventilation and bystander exposure are also included. Lung tissue from these same workers will be analysed for asbestos fibres to allow for comparison against the ASTEM estimation.
What this study adds
There are limited data available of the exposure workers experienced when working with asbestos.
Accurate exposure assessment is important for diagnosis, epidemiological investigations and for medicolegal purposes.
An asbestos-specific task exposure matrix (ASTEM) has been designed using historical workplace dust monitoring data.
The ASTEM is a summary of exposure estimates associated with common asbestos products and their related tasks, this can be used with information taken from workers about their occupational history to calculate a lifetime cumulative asbestos exposure.
Using a standard set of exposure estimates will allow consistency in the calculation of workers' exposure to asbestos.
Acknowledgments
The authors would like to thank WorkCover NSW for providing access to archived company dust reports. The authors also wish to thank Sandra Ware and Marian Barker for editorial assistance and data entry work.
References
Footnotes
policy-implications Information taken from workers about their occupational history can be used together with the asbestos-specific exposure matrix (ASTEM) to calculate a lifetime cumulative asbestos exposure. Using a standard set of exposure estimates will allow consistency in the calculation of workers' exposure to asbestos.
Funding This study was funded by the Workers' Compensation (Dust Diseases) Board of New South Wales. This funding body did not have involvement in any area of the study's design, implementation, analysis or write up.
Competing interests None.
Provenance and peer review Not commissioned; externally peer reviewed.