years, epidemiologists will benefit from data about what the usual levels exposures in different jobs are now. Another important potential use of this data could be for setting priorities for standard-setting. OSHA has proven capable of promulgating an average of only one or two new standards a year. Under those circumstances, OSHA's impact on worker health is stronly influenced by whether it chooses to lower exposure limits for those hazards where the health effects will be greatest. That choice requires information about the dose-response curves for each hazard and the current distribution of exposures. While much scientific attention has focused on the first issue, relatively little has been given to the equally important second question. Of course, once OSHA makes a tentative decision to set a standard, an expensive regulatory analysis is conducted that does attempt to provide exposure information. But at the time when priorities are set, it is not available and the question of how many diseases would be prevented by a given reduction in exposure limits can not even be roughly answered. Better information has the potential to help OSHA increase its contribution to public health. A potential source for making rough estimates of the distribution of worker exposures can be found in the OSHA MIS. OSHA conducts roughly 10,000 health inspections a year and samples worker exposures in almost all of them. My studies for the Office of Technology Assessment showed that prior to 1979, however, only about 50% of the samples were recorded in the MIS. Since 1979, according to OSHA officials, the percentage has fallen to 15 or 20%, essentially because of administrative problems. Given the importance of the occupational health issue and the absence of alternative sources of data, this discarding of MIS samples is deeply troubling. Serious attention should be given to methods to guarantee full reporting. It may turn out, however, that additional data gathering efforts will be required to get an accurate picture of exposure levels and trends. Currently, the NIOSH National Occupational Hazard surveys have estimated how many workers were exposed to certain hazards, but not the crucial data about the levels of exposures. One option is to gather data on levels of exposures from a sample of the workplaces inspected. A general survey may not turn out to be the most efficient strategy if the real concern is with the 10 or 15 hazards which are the main candidates for regulatory attention over the coming years (to the extent that we are able to predict what they are--a more difficult task in health than in safety). The general conclusions are that we appear to have underinvested in data about health compared to data about safety hazards; and that we have failed to focus on providing the data that will be most useful for policymaking, now and in the future. U.S. DEPARTMENT OF LABOR SECRETARY OF LABOR WASHINGTON, D.C. Honorable Barney Frank Chairman Subcommittee on Manpower and Housing Dear Mr. Chairman: This is in response to your letter of July 10 concerning clarification of responsibility for the definition of occupational illnesses. The delegation of authority and assignment of responsibility for carrying out the provisions of the Occupational Safety and Health Act (the Act) was originally delegated in the Secretary of Labor's Order 12-71 of May 4, 1971, and most recently updated in the Secretary's Order 8-76 of April 30, 1976. Copies are enclosed for your reference. Section 4 (c) of Secretary's Order 8-76 delegates to the Commissioner of the Bureau of Labor Statistics (BLS) the responsibility for "...developing and maintaining an effective program of collection, compilation, analysis and publication of occupational safety and health statistics." This language clearly covers Occupational illnesses. The Bureau has been carrying out its responsibilities under the Secretary's Order since 1972. The instructions on the reverse side of the Form OSHA-200 (Log and Summary of Occupational Injuries and Illnesses) contain definitions for occupational injuries and illnesses. Occupational illness is defined, you will note, as any abnormal condition or disorder [of an A list is also given of examples of illnesses to be considered recordable illnesses. It is clear by inference that such illnesses can only be recorded by the employer when there is an evident connection between exposure and onset of symptoms. While this necessary condition does not rule out the recording of long-latent illnesses which may have been caused by "environmental factors associated with employment," as a practical matter, the recording of such illnesses is not likely to occur, given the mobility of the American workforce and the limitations of occupational epidemiology. The difficulties in assessing "the full level of occupational illness" are both policy-related, as BLS Commissioner Janet Norwood suggested at the hearing before your Subcommittee on June 20, and research-related, as Deputy Assistant Secretary Patrick Tyson noted on that occasion. Detection and attribution of occupational causes to long-latent diseases are complex problems. Among the complicating factors in determining whether a long-latent illness is in fact job-related are the following: the mobility of the American workforce already noted; the role of life-style (smoking, drinking, diet), and heredity; the difficulties scientists acknowledge in determining the dose-response relationship in man; the relative roles of primary carcinogens (promoters and initiators); and the possible synergistic effects of multiple exposures. Thus, even if exposure to a toxic substance is found to have occurred in an individual's workplace experience fifteen or twenty years earlier, imponderables may remain. For example, can that exposure be considered evidence that the illness the employee later contracted was due to that cause, or could the illness have been due, let us say, to other exposures in the individual's home environment? Unfortunately, even physicians are often unable to assess accurately the extent of an occupationally-related disease in an individual. For example, an article in the September 1983 American Journal of Public Health, titled "Sentinel Health Events...: A Basis for Physician Recognition and Public Health Surveillance," indicated that fewer than ten percent of board-certified pathologists were able to diagnose asbestosis in a straightforward example of the disease. If medically trained professionals have so little success in recognizing occupational illness, a statistical assessment of the extent of the problem at this time may well be, as Dr. Norwood testified, ...an intractable problem." . In response to your last question, additional authority to the Secretary would not assist the Department in assessing the "full level of occupational illnesses." Under section 8 (c) of the Act, the Occupational Safety and Health Administration (OSHA) has the authority to prescribe and enforce regulations necessary to collect information related to occupational injuries and and illnesses, including long-latent illnesses. These are the regulations published in 29 CFR 1904. From a review of the language of the Act, the Secretary of Labor appears to have sufficient authority to collect information about the factors leading to occupational illnesses. The difficulty lies not in lack of authority or in the ability of the Secretary to issue regulations, but in the ability of employers and the general public, and ultimately the scientific community, first, to recognize an illness, and secondly, to classify it as occupationally-related. In summary, the Occupational Safety and Health Act provides sufficient authority to the Secretary to gather any available information from employers. The technical ability of the individuals at the workplace to record that information is at this time insufficient to improve current aggregate estimates about the incidences of occupational illnesses. Sincerely, Raymond J Donovan Raymond J. Donovan Enclosures |