Most of the studies of work-related fatal injury are essentially cohort studies of the population of interest, with the population being all working persons or subgroups of the population identified by characteristics such as occupation, industry, age, or sex. However, other approaches have been used (e.g., case-control studies, capture-recapture studies, and proportional mortality studies) or could be used (e.g., case-crossover design). Sometimes the analysis focuses on frequencies and percentages, but usually rates are presented, at least at an overall level. Comparisons are made on the basis of frequency, percent, and/or rate, with the effect of multiple exposures of interest controlled by stratification or exclusion. Other approaches have included Poisson regression or other regression methods, the use of standardized or proportional mortality ratios, and the calculation of lifetime risk or years of life lost.
Disease studies commonly focus on individual worker groups. Cohort approaches are often used, but case-control designs, and to a lesser extent cross-sectional designs, are more commonly used than they are for injury studies. The main outcomes of interest are relative risks (or odds ratios), standardized mortality ratios, proportional mortality ratios, and DALYs, which have become a commonly used measure in some settings. As mentioned earlier, studies assessing burden commonly make use of the population attributable risk approach. More detail on these approaches is available in many good epidemiology references (Rothman and Greenland, 1998; Checkoway et al., 2003).
Data Sources For Identifying Work-Related Injuries And Disease
In most developed countries and some developing countries, routine data systems probably provide reasonable coverage of severe and fatal injuries for the employees and industry sectors that they do cover (although some underenumeration can be expected, even for deaths). Disease deaths and serious disease cases may also be recorded, although the routine systems rarely identify the work-relatedness of an individual case. Cases recorded by available data systems and reported to the WHO or the International Labor Organization have been used as the basis for estimating the global number of work-related injury (and disease) cases. The main examples of useful data sources are summarized here.
Death Certificates
Since a death certificate should be completed for all deaths, death certificates form a potential source of information on work-related injury and disease deaths if information on the certificate is able to be used to identify such deaths. This information is usually only of direct use for injuries, unless the disease is of a sort that is explicitly connected to work (such as coal workers’ pneumoconiosis or silicosis). For injury, if work-relatedness is explicitly coded, it is usually identified by a box that can be checked if the death is related to work or occurs at work. The U.S. National Traumatic Occupational Fatalities surveillance system, which is run by the National Institute for Occupational Safety and Health, is the best example of the use of death certificate information for surveillance purposes. Some information is coded, but most of the available information is in narrative form. Although this has disadvantages, narrative information is very useful because it allows searches to be conducted on a wide variety of particular topics.
Death certificates provide a relatively low-resource approach to work-related fatal injury surveillance. Coverage of work-related deaths should not be influenced by employment arrangement, job type, or marital status, factors that by design or practice can exclude deaths from being included in compensation systems. However, there are limitations to the usefulness of death certificates for injury. These include incomplete coverage, inconsistent information on certificates because of the many different people involved in completing them, lack of uniform coding of information provided on the certificate, lack of relevant information on occupation and industry of the person, and lack of detailed information.
Workers’ Compensation Information
Information systems based on compensation claims usually have the advantage of reasonably up-to-date and independently verified information, but the disadvantage of deliberate limitations in the coverage of some workers and the practical exclusion of some others. Examples include the Berufsgenossenschaften in Germany, provincial-based systems in Canada, and the National Data-Set for Compensation-Based Statistics in Australia. All these systems provide information on both injury and disease, although injuries are generally better covered. For each of these systems, and for other compensation systems, there are concerns that the coverage is incomplete and that the extent of this incompleteness is uncertain, although it is of the order of 40% in Australia for work-related fatal injury and probably is similar elsewhere (Driscoll et al., 2003).
OHS Authority Reports
Most countries have one or more state or federal government occupational health and safety authorities that investigate some work-related incidents and keep statistics regarding them. These are a potentially very useful source of information on work-related injury, since they often contain detailed information. However, the information provided by these agencies can have many shortcomings, being influenced by changes in administrative policy, varying approaches to coding, and variable coverage.
A good example of work-related fatality surveillance using information from official statistics held by OHS authorities is the Reporting of Injuries, Disease and Dangerous Occurrences Regulations (RIDDOR) system. The RIDDOR data are used as the basis of information published by the Health and Safety Executive (HSE) in the United Kingdom. RIDDOR is based on notifications to the HSE required by statute. Like most systems that rely on notifications to OHS authorities, or investigation by the OHS authorities, work-related deaths that are covered by other agencies for the purposes of reporting or investigation are not included within RIDDOR statistics. These include deaths related to the supply or use of flammable gas, most deaths related to marine fishing and merchant shipping, fatal injury associated with civil aircraft flight or with premises covered by the relevant explosives act, and deaths arising from motor vehicle crashes on public roads. Therefore, again like most systems based on OHS notifications and inspections, the RIDDOR system covers the majority of work-related injury deaths of workers, while probably missing a substantial minority of those deaths. Some information on bystander deaths is included, but coverage is probably poor.
Coroners’ Data
The advantages and disadvantages of using coroners’ data as a source of information on work-related injury deaths have already been mentioned. In brief, all injury deaths, including those that are related to work, are due to some form of external cause, and so can be expected to be reported to the coroner (as required by law in most developed countries). All deaths reported to the coroner are the subject of some form of investigation, and the information collected during a coroner’s investigation is often very detailed and can contain useful information for descriptive and preventive purposes. The size and completeness of the data set allows patterns of circumstances to be seen that would not be noticed when a more limited group of incidents was examined; and the information contained in coroners’ files is theoretically available to persons with bona fide reasons for accessing it. Problems with using the coroners’ system (or medical examiner system in some places in the United States and Canada) as a source of information include a lack of standardized classification and investigation approaches, lack of mutually exclusive classification categories, incomplete coverage, lack of relevant information in the coronial files, lack of a centralized file storage system, and reliance on paper records. Some countries are moving to overcome these difficulties by developing a computerized database of all deaths reported to the coroner, although Australia is probably the only country where such a system is fully functional (Monash University National Centre for
Coronial Information (MUNCCI), 2001). The coroners’ system is not a good source of information on work-related disease deaths because, with some minor exceptions, these do not fall under the area of interest of the coroners’ system. Coroners’ records have been most comprehensively used as a source of information on work-related fatal injuries in Australia and New Zealand.
Hospital Admissions
Many countries have begun collecting information on the type and cause of hospital admissions. Some of these systems use codes that identify a connection to work. This is primarily of use for (nonfatal) work-related injury cases, because the connection between work and disease is often not recognized.
Emergency Department Presentations
Emergency department data systems are also increasingly being used as sources of information on various types of injury, including work-related injury, in persons that present to emergency departments. Like the hospital data, the emergency department systems appear to be of little use as a source of information on work-related disease.
Multiple Sources
Some fatal injury data systems attempt to use information from multiple sources. Probably the best of these is the Census of Fatal Occupational Injuries (CFOI), which is run by the U.S. Bureau of Labor and Industry and was introduced in 1992. This system uses death certificate information, but complements it with information from workers’ compensation reports, coroners, the Occupational Safety and Health Administration, police, and the media. A unique aspect of the CFOI system is that some follow-up of surviving persons is undertaken. There is no equivalent system collecting information on work-related disease.
Company Data
Data maintained by companies in personnel and job assignment records are often used in studies to assess the relationship of exposure to a particular disease outcome. Exposure measures can be based on the duration of exposure, the level of exposure, or some combination of the two. Large-size cohort studies can be based on collection, and a common analysis, of data from several companies making the same products. The International Agency for Research on Cancer has broadened these analyses to a global level for a number of work-related (and other) exposures, such as sulfuric acid mist, wood dust, manmade mineral fibers, and formaldehyde.
Surveys
Many countries conduct national or regional surveys that include questions on occupational injuries and illnesses. For example, the U.S. Bureau of Labor Statistics conducts annual surveys of occupational injuries and illnesses, and the Australian Bureau of Statistics conducts national health surveys approximately every five years. Such surveys are usually based on self-report.
The Press
Several studies have used press clippings and radio reports as a source of information on work-related fatal injury, but usually as an adjunct to other information sources rather than as the primary data source. Some information on work-related disease is available from the press, but this is ad hoc and not amenable to use for surveillance or other purposes.
Conclusion
Work-related injury and disease is a significant cause of injury and disease at national and global levels. It is important to obtain reliable information on the burden arising from work, and the exposures leading to this burden, in order to prioritize and plan appropriate interventions. Such interventions are considered in various general texts (e.g., see Agius and Seaton, 2005).
Bibliography:
- Agius R and Seaton A (2005) Practical Occupational Medicine. London: Hodder Arnold.
- Checkoway H, Pearce N, and Kriebel D (2003) Research Methods in Occupational Epidemiology. New York: Oxford University Press