gether, or "partial," if it limited the kind or amount of work performed.

Based on survey results, it was estimated that about 15 million persons age 20 to 64 nationwide were disabled from all causes. To determine the job-relatedness of a disability, the survey respondent was asked: "Was your (main condition or illness) caused by your job?" Of the resulting estimated 2.4 million job-caused disabilities, 1.7 million were attributed by the author of the study to occupational disease. (For purposes of the study, occupational diseases were defined as all cases of disability which were not caused by an accident on the job.) The study further stated: "Because of limited understanding of what diseases are occupational in nature, it is likely that the actual number of occupational disease cases is much higher" (page IV).

The study indicated that, of the 1.7 million persons disabled by occupational disease, 1.1 million or twothirds were partially disabled. Among the 0.6 million severely disabled, musculoskeletal and cardiovascular conditions accounted for almost 60 percent of all conditions reported. About 25 percent had mental and digestive ailments and 9 percent, respiratory conditions. Cancer caused by occupation was estimated among the severely disabled at little more than 1 percent, slightly over 6,000 cases.

Can we say that these are good estimates? As the author of the study points out, the survey provided a snapshot of the population at a given time." Thus, the resulting estimates measured the prevalence of occupational disease, counting as they did all existing disabilities without regard to time of onset or diagnosis. The occupational relationship was subjectively perceived by respondents and may or may not have been corroborated by objective medical evidence. Except perhaps for the musculoskeletal impairments, the other conditions were of a type for which objective evidence of occupational causality might have been difficult to obtain.

Finally, while national estimates based on survey data of self-perceived work-related disabilities suggest that an undercount in employer-based occupational illness estimates does exist and in some identifiable parameter, it appears that those diseases which have long latent periods and have yet to be diagnosed are missed in this approach as well. Many long-latent diseases or aggravating disease symptoms of possible occupational origin are recognized beyond the cutoff age of most direct surveys. When disease appears after a worker has lived at least a normal life span, other factors, related to the aging process, enter which may lessen the urgency to determine precise causes.

Disability applicant files. The Social Security disability program provides benefits to disabled adults with work experience in employments covered by the Social Securi

ty Act, and to adults disabled since childhood who are dependents of disabled or retired work beneficiaries or of deceased insured workers. To qualify, claimants for Social Security Disability Insurance must prove that they are both disabled and unable to engage in any substantially gainful work due to their medical condition. Two sample data files from disability applicant records are maintained by the Social Security Administration (SSA), the Continuous Disability History Sample (CDHS) and the Longitudinal Sample of Disability Insurance Applicants (LSDIA). CDHS contains about 25 percent of allowed claims and 10 percent of denied claims, while LSDIA is a 5-percent longitudinal sample of disability applicants.

Disability applicant records contain demographic information, such as sex, race, date of birth, education, occupation, and industry of employment, as well as important medical information, such as diagnosis (primary, secondary, and tertiary), listings of impairments, principal body system involved, and severity and duration of impairment. It is important to note, however, that in the recording and coding of disability cases, work-related illnesses are not distinguished from nonwork-related ones. That is, for adjudicative purposes, an occupational relation to the disease or disability does not have to be established. (Although occupations are associated with worker claimants, no causal relationship is required or intended.) According to a 1974 report by the U.S. Department of Health, Education, and Welfare, the leading causes of disability by diagnosis were listed as chronic ischemic heart, schizophrenia, osteoarthritis, emphysema, displacement of intervertebral disc, diabetes mellitus, rheumatoid arthritis, acute cerebrovascular disease, malignant neoplasm of trachea and lung, neuroses, pulmonary tuberculosis, and mental disorders.20 This listing also contains types of diseases for which, obviously, objective evidence of occupational causation would be hard to come by.

To qualify for disability insurance, a claimant must have a health condition sufficiently incapacitating to be unable to engage in any substantial, gainful work. Thus, a worker may have an occupational disease, but be disallowed disability benefits because pursuit of gainful employment is still possible. Therefore, disability estimates based on SSA records are not precise and comprehensive indicators of occupational impact. In addition, the occupational history of an applicant is limited to his or her longest full-time occupation in the 10-year period preceding the alleged date of onset. Because a jobcaused disability, especially one of a chronic nature, may have been due to an earlier exposure, and perhaps, to a different job, there is a potential bias in the use of these statistics for epidemiological study. While the SSA disability files are an important source of data for development of morbidity ratios which identify disease and

occupational relationships worthy of further study, they are not suitable for deriving estimates of occupational disease."

Health Interview Survey data. The Health Interview Survey of the National Center for Health Statistics is a nationwide survey of approximately 40.000 households. conducted on a continuous basis. It is designed to gather information on personal and demographic characteristics, illnesses, injuries, impairments, chronic conditions, and other health topics. Respondents are asked whether they worked in the 2 weeks prior to the interview week, and in what occupation and industry. Each year's sample includes about 120,000 persons, of whom roughly 50 percent are employed.

As data are processed and tabulated, the center publishes analytical reports on various topics. While very few reports have been published on the work force population as the primary study target, data files are available for research purposes. Like the Social Security Administration disability applicant data the center's

data can serve important epidemiological research objectives, but should not be used to derive precise estimates of occupational disease incidence."

DESPITE THEIR SHORTCOMINGS, the results of the studies and applied methodologies discussed above do, in combination, point to a larger impact of the workplace on the health of workers than is borne out in regularly published statistics, although the magnitude of the understatement remains uncertain. Continued efforts towards improved or new methods are needed to produce national estimates of greater credibility for the chronic and long latent disease component of job origin. Such efforts might include improved techniques for diagnosing occupational diseases; more sophisticated and efficient means of monitoring workers' health; education and training of doctors and workers regarding health hazards on the job; conduct of epidemiological studies representative of national experience; and establishing methodology for determining the contribution of job exposure to the origin and course of disease.

Harvey J. Hilaski, “Understanding statistics on occupational illnesses," Monthly Labor Review, March 1981, pp. 25-29.

'David P. Discher, Goldy D. Kleinman, and F. James Foster, Pilot Study for Development of an Occupational Disease Surveillance Method (Washington, U.S. Department of Health and Human Services, National Institute for Occupational Safety and Health, 1975).

'In the survey, the examining physician determined a condition to be one of five types: probable occupational disease; doubtful occupational disease; suggestive history; cannot be evaluated; probably nonoccupational. Probable occupational disease was considered to be present when "manifestations of disease are consistent with those known to result from excessive exposure to a given injurious agent; this injurious agent is present in the patient's working environment and significant contact in course of usual duties is likely." Discher and other, Pilot Study, p. 25.

• Although OSHA had a noise standard as early as 1971, environmental and audiometric testing was not formalized at the time of the pilot survey. Under the Hearing Conservation Amendment, which became effective Aug. 22, 1981, the permissible exposure level (PEL) remains at the 90 decibel (dB) level as an 8-hour time weighted average but an 85 decibel time weighted average was established as an action level which triggers the initiation of hearing conservation programs. including exposure monitoring, audiometric testing of employees. training, and some recordkeeping.

'U.S. Department of Labor, Cotton Dust: Review of Alternative Technical Standards and Control Technologies, Report to the Congress, May 1979.

U.S. Department of Labor. An Interim Report to Congress on Occupational Diseases, June 1980. Also see "Labor Month in Review," Monthly Labor Review, August 1980, p. 2.

See Occupational Injuries and Illnesses in the United States by Industry. 1980, Bulletin 2130 (Bureau of Labor Statistics, 1982). p. 32. table 7.

"This study was undertaken in the spring of 1980 as part of, the Bureau's continuing evaluation of occupational safety and health statistics.

Gilles P. Theriault, William A. Burgess, Lou J. DiBerardinis, and John M. Peters, "Dust Exposure in the Vermont Granite Sheds," Archives of Environmental Health, Vol. 28, 1974, pp. 12-17.

James A. Merchant, John C. Lumsden, and others, "Dose Re

Exposure information for all substances, except arsenic, was derived from the National Occupational Hazard Survey conducted by NIOSH during 1972-74. Exposure information for arsenic was contained in criteria for a 1975 recommended standard developed by NIOSH.

11 Richard Doll and Richard Peto, "The Causes of Cancer: Quantitative Estimates of Avoidable Risks of Cancer in the United States Today." Journal of the National Cancer Institute, June 1981, pp. 123845.

"Glen M. Shor, Occupational Disease and Compensation: An Analysis of the 1972 SSA Survey of Disabled Adults, prepared under contract for the Office of the Assistant Secretary for Policy. Evaluation and Research, U.S. Department of Labor. October 1979. "Shor. Occupational Disease. p. 2.

Social Security Disability Applicant Statistics 1970 (Washington. U.S. Department of Health, Education, and Welfare. 1974).

"For a recent study of Social Security disability data, see NIOSH Research Report, Occupational Characteristics of Disabled Workers (Washington. U.S. Department of Health and Human Services. National Institute for Occupational Safety and Health, 1980).

"For analysis of Health Interview Survey data. see NIOSH Re search Report, Industrial Characteristics of Persons Reporting Morbidity During the Health Interview Surveys Conducted in 1969-1974 (Washington, U.S. Department of Health and Human Services, National Institute for Occupational Safety and Health, 1980).