mortality analyses cannot be representative of the Nation as a whole and lack the statistical power that would be present in a nationwide analysis. Measures to Facilitate Epidemiologic Studies. OTA has identified three issues concerning measures that might facilitate the conduct of epidemiologic studies. These involve possible changes to the National Death Index, use of IRS data, and the linking of Federal data systems. The National Death Index. Information on death certificates is essential to any epidemiologic study investigating causes of death. When supplied with the name and date of birth or social security number of an individual, the National Death Index (NDI), a service of the NCHS, can be used to learn whether that person has died and the location of the death certificate. Until NDI's establishment, epidemiologists had to make inquiries of every department of vital statistics to locate the holder of a death certificate. Quite simply, the NDI reduces the number of inquiries necessary to locate a death certificate from more than 50 to 1, but each death certificate must still be obtained from the state that holds it. The NDI would be more useful if it could supply all the information encoded upon death certificates. Were it to be modified to do that, epidemiologists could obtain all vital information for mortality studies from a single inquiry. The benefits of such a change would be to speed up studies and to reduce the costs of the flow of information. The costs of such a change would be an increase in the workload of the NCHS in preparing the NDI and some adjustments in the relationship between the Federal and State participants in the vital statistics system. In particular, one area of concern is that the revenue from the small fee that state departments of vital statistics charge for each death certificate is often an important part of their financing. Addresses from Internal Revenue Service Records for epidemiologic studies. Epidemiologic studies frequently require that investigators interview subjects of the study or their families. An impediment to the studies is the difficulty of locating people. Internal Revenue Service (IRS) records are a reasonably complete source of recent addresses, but only NIOSH and other Federal agency scientists and persons working on contract to NIOSH can obtain addresses from IRS. This has been termed the "NIOSH window." In order to facilitate the conduct of epidemiologic studies, it has been suggested that a wider spectrum of researchers be permitted to obtain addresses from IRS. Any expansion of the window would require safeguards so that the addresses from IRS records are used only for epidemiologic studies. Linking Federal data systems to facilitate epidemiologic studies. It has also been suggested that the records systems of the Census Bureau, Social Security Administration, Veterans Administration, and OSHA and NIOSH could be linked together to provide information about medical conditions, work history and exposures, and current address in a single file. Such a link could improve epidemiologic studies, but it might also increase the possibility of invasion of an individual's privacy. Although epidemiologists are convinced of the value of linking together data systems, few efforts to do so have been approved. It is a tricky proposition to balance more efficient epidemiology against possible invasions of privacy. The recent report of the Department of Health and Human Services Committee to Coordinate Environmental and Related Programs "On Occupational Cancer Estimation" suggests some options for linking data systems. Injury Rate (per 100 employees) Injury Rate. Unemployment Rate 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 Mr. FRANK. Thank you very much, Mr. Kronebusch. We appreciate particularly the specific recommendations. Without objection your entire statement will be entered in the record. I think this is one incidence when we are going to have as a subcommittee report, and possibly a piece of legislation, some very specific points. Let me just make one point that comes to me out of all of this with regard to illness is that workplace data which pretends to measure the incidence of illness is virtually useless. It seems to me we ought to be clear that it is going to drastically underreport. It seems to me that what we are going to be wanting in the workplace data is what were the conditions that people were working under, what were they exposed to, et cetera. But I get the sense that trying to measure the incidence of occupational illness from actual workplace data is just going to be so woefully underreporting that it is just going to distort things. Mr. KRONEBUSCH. Yes, we believe you are correct on that. Diseases with a latent period that don't become manifest for years after exposure, diseases that develop over time will always be a problem to capture in the employer-maintained records, which tend to be oriented toward writing down injury information, things that happen instantaneously in the workplace. Moreover, many occupational exposures cause diseases that are relatively common in the population as a whole and are caused by factors other than workplace factors. I think, a very clear example of that is lung cancer, where a large number of-most cases-of lung cancer are due to smoking, but it is also accepted that certain workplace exposures can cause lung cancer. When a particular worker comes down with lung cancer, it is very difficult to determine whether it is due to an occupational exposure or other factors. Mr. FRANK. Hearing loss also appears to be one of those that— there is a natural hearing loss with aging, but there is also an exacerbation of that in some places and-for instance, the workplace data is just so misleading that they ought to be put aside. Financially, how much-you have given us some specific recommendations-what is the order of magnitude of increased funding that you would be talking about? Mr. KRONEBUSCH. We don't have any specific estimates of funding. Some of our suggestions would, of course, be simpler than other suggestions. We suspect that, for example, conducting mortality surveys, as we suggest, would be relatively inexpensive. It is just a matter of putting some money in the right places, ensuring accurate coding on death certificates and then collecting the information. Much of this information is already coded into computers in the present system, and it is a matter of extracting the appropriate information and conducting an analysis. Mr. FRANK. The approach that NIOSH was taking-yes, go ahead. Better identify yourself for the record, please. Dr. GOUGH. I can answer specifically- Mr. FRANK. Could you identify yourself, please. Dr. GOUGH. Yes, I am Michael Gough. I think the effort between NIOSH and NCHS. Mr. FRANK. NCHS is the National▬▬ |