Mr. FRANK. Thank you. The panel is dismissed. We appreciate very much your testimony and we will next hear from Mr. Kronebusch.



Mr. KRONEBUSCH. Mr. Chairman, I am Karl Kronebusch, and I am employed at the Office of Technology Assessment in the health program, where I am working on an assessment of controls for occupational safety and health that will be published this summer.

With me today is Dr. Michael Gough, who is the project director for that assessment.

I have submitted written testimony and I will summarize that this morning

In our assessment, we describe the available data concerned occupational injuries and illnesses in detail. Today I will briefly discuss occupational injury data and then move on to your principal interest in illness data.

To provide a partial answer to your question about possible underreporting of injuries, I will summarize our conclusions about the comparison of data from several sources. The discrepancies that we found among these data sources could arise from different methods, and as such, may not be worrisome, or they could indicate employer underreporting or differences in how employers and employees interpret the severity of an injury. Thus, as described in my written testimony, one important issue concerns the possibility of establishing some kind of verification or quality assurance procedures for occupational injury data.

We also analyze the reported declines in occupational injury rates over the last 3 years. Although it is possible that changes in OSHA policies caused some of the decline, the effects of several other factors must be considered.

The most important of these is the effect of the recent recession, especially in construction and manufacturing. During business upturns, injury rates increase, while business downturns lead to lower rates. We believe that increased unemployment largely explains the injury rate declines from 1979 to 1982.

Compared with occupational injuries, there is substantially less quantitative information on occupational illnesses. The most commonly quoted estimates are that 390,000 cases of illness and 100,000 deaths occur annually as a result of workplace conditions. These estimates are very uncertain, but it is unclear to what extent they are biased. It is well accepted that employer records and the BLS annual survey estimates, which are based on employer records, underestimate the magnitude of the occupational disease problem.

Unfortunately, it is difficult to quantify the extent of this. For example, one study from the 1970's revealed a large number of probable occupational illnesses that were not reflected in employer records, but it is not clear to what extent this applies to other industries not included in that study.

The number of occupationally related cancer deaths has been the subject of a heated debate. In our assessment of technologies for de termining cancer risks from the environment, OTA concluded that most estimates of the fraction of cancer associated with occupation fit within a range of 5 to 15 percent. A 1981 conference produced general agreement that occupational cancer represents less than 10, and probably less than 5 percent of all cancer.

However, it is interesting to note that the BLS Supplementary Data System reported only 234 workers' compensation cases for cancers in 1980. This can be compared to the range of estimates for occupational cancer caused by asbestos alone, which is between 4,000 and 12,000 cases annually.

The debates on the proportion of disease caused by occupational factors sometimes diverts attention from the general agreement that some workplaces have very high risks. In addition, when workplace hazards are recognized, control technologies can be applied to reduce the risk of disease. Our assessment report discusses several possible changes which could improve data collection efforts, thus increasing the usefulness of collected data, ensuring the validity of the data, and facilitating epidemiology.

Two issues largely concern injuries. The first is the level of activity that OSHA, NIOSH and BLS direct toward investigating fatal and nonfatal injuries and in preparing information useful to workers, employers and professionals. A second issue concerns whether or not BLS is to conduct a new quality assurance program.

We have also identified issues that are related to information on illnesses and exposure. One concerns mortality surveys to study associations between occupations, industries and causes of death. These studies can help identify high risks associated with some types of work and indicate which occupations and industries do not present high risks.

We see two specific activities as necessary. First, ensuring accurate coding of information about industry and occupation on death certificates, and second, actually conducting these analyses. Epidemiologic investigations to explore associations between particular industries, occupations and illnesses can be made more efficient and useful by certain changes in Federal data collection efforts. These involve possible changes to the National Death Index, the use of IRS data and the linking of Federal data systems.

The National Death Index is a service of the National Center for Health Statistics and can be used to learn whether a person has died and the location of the death certificate. The National Death Index would be more useful if it could supply all the information coded on death certificates. Such a change would increase the work load of the National Center for Health Statistics and force some changes in our current vital statistics system. In particular, States would have to be reimbursed for the fees that they now receive for supplying copies of death certificates.

Epidemiologic studies frequently require that investigators interview people, but it is often difficult to locate the people being studied. Internal Revenue Service records are a reasonably complete source of addresses, but only NIOSH, other Federal agency scientists and the persons working on contract for NIOSH can obtain addresses from the IRS.

It has been suggested that a wider spectrum of researchers be permitted to obtain addresses from the IRS, but any expansion would require safeguards to ensure that the addresses from IRS records are used only for epidemiologic studies.

It has also been suggested that the record systems of the Census Bureau, Social Security Administration, Veterans Administration, OSHA and NIOSH could be linked together to provide information about medical conditions, work history and exposure and current address in a single file. Such a link could improve epidemiologic studies, but it might also increase the possibility of an invasion

of individual privacy.

Finally, based on a contract report prepared for our assessment, we believe that information collected during OSHA's health inspections may be useful for developing estimates of worker exposures. We suggest that this possible use be explored further.

Thank you.

[The prepared statement of Mr. Kronebusch follows:]

[blocks in formation]
[merged small][merged small][merged small][merged small][merged small][merged small][ocr errors]

I am Karl Kronebusch and I am employed at the Office of Technology

As sessment where I am working on an assessment of controls for occupational

safety and health.

The full assessment will be published this summer.


describes available data concerning the extent of the occupational health and

safety problem, the control technologies available to improve health and safety

in the workplace, the activities of the Federal Government, and presents a

number of options for Congressional consideration.

With me today is Dr. Michael

Gough, who is the Project Director for that assessment.

Estimates of Fatalities, Injuries, and illnesses

Fatalities and Injuries

Estimates of the number of occupational injuries can be obtained from the

Annual Survey of the Bureau of Labor Statistics (BLS), the National Safety

Council (NSC), and the National Health Interview Survey (NHIS) of the National

Center for Health Statistics.

Although subject to several limitations, the BLS

Annual Surveys are the best single source of statistical information concerning

work-related injuries.


Both the BLS and the NSC prepare estimates of the annual

number of workplace deaths. The latest BLS number is about 4,100; the NSC's is 11,100. OTA has generated an estimate of fatalities due to occupational

injuries that lies between these extremes. We adjusted the BLS Annual Survey

data by subtracting deaths reported from heart attacks and adding the number of

cases reported for Federal employees.

We then assumed that the occupational

fatality rate for employment that is not covered by the BLS survey is about the

same as the average for employment surveyed by BLS.

The total, after these

three adjustments, is about 6,000 annual deaths due to occupational injuries.

This translates into about 25 occupational fatalities each working day.


Estimates for non-fatal injuries include the NSC's estimate of

1.9 million cases of disabling injuries, BLS estimates of 2.1 million los t

workday injuries and 4.8 million "recordable" injuries, a NIOSH estimate of 3.2

million emergency room cases, and an estimate of 11.3 million injuries derived

from the National Health Interview Survey.

In part this five-fold range

reflects differences in definitions of injury, the size of the workforce

included in the estimate, and the methods used to make these estimates.

To provide a partial answer to your question about underreporting, I will

describe two analyses conducted by OTA.

The first analysis involved a

comparison of state-by-state totals of the number of injuries derived from two


the BLS Annual Survey and the BLS Supplementary Data System (SDS).

The Department of Labor requires employers to maintain logs of injuries, and BLS

requests information about injuries from about 280,000 employers to produce the

Annual Survey.

For the SDS, BLS collects Information about injuries reported by

employers to workers' compensation agencies in a number of states.

Our analysis

was designed to determine if these two data sources, both based on employer

reports, were consistent with each other for the years 1978 through 1980.

Some states report only lost-workday cases to the SDS, while other sates

report all cases involving either lost workdays or medical treatment.

For the

states that report only lost workday cases, the numbers from the two data

sources appeared to be fairly consistent after we adjusted for the minimum

« ForrigeFortsett »