The second BLS survey which we employ is the BLS Supplementary Data System. Under this survey, 35 states provide BLS with detailed first reports of injuries and illnesses covered by workers' compensation programs. For those injuries and illnesses which are included in this survey, the Supplementary Data System provides us with detailed information on the nature and circumstances of each case and on the demographic characteristics of each affected worker, including such information as age, race, sex, industry of employment and specific occupation. Through tablulation of these data, we are able to develop descriptive epidemiologic profiles by industry and by occupation for those categories of occupational disease which are reported to state workers' compensation programs. This survey suffers, however from variation in the extent of coverage among the state programs and from variations in the completeness of reporting.

The Social Security Administration (SSA) Disabilty Award File is a third source of data used by the Institute to monitor trends in occupational illness and injury. An estimated 400,000 disabilty awards are made annually by SSA. Data on the cause of each case of permanent disability, as well as details on each recipient's age, race, sex, industry and occupation are analyzed from this file. Although the disablity awards made by SSA include awards for many types of a disablity, in addition to those which are the result of occupational disease or injury, this file nevertheless provides a valuable means of identifying those industries and occupations in which unusual patterns of disability exits. This data base lacks specificity, since the conditions reported are not necessarily work-related.

The fourth record system which we employ is the National Vital Statistics Program, which is maintained by the nation's State health departments. This data system collects essential medical and demographic information on all births and deaths occurring in the States. Prior to 1979, there existed no uniform national system through which the states routinely record on their certificates information in the industry of employment and specific occupation of each decedent or on the industry and employment of the parents of each newborn child. Since 1979, however, with the support of NIOSH, 36 of the States have developed the capability to code occupation and industry, and 15 States now routinely provide such data to the National Center for Health Statistics (NCHS). NIOSH and NCHS are collaborating to implement a program for the coding and analysis of these data. This effort represents a significant component of NIOSH's occupational mortality surveillance program.

The fifth of the data systems which we employ is the National Electronic Injury Surveillance System (NEISS), a data base maintained by NIOSH in conjunction with the Consumer Product Safety Commission. This system collects data directly from a sample of 66 representative hospital emergency rooms. For each admission to these emergency rooms the relationship between work and injury is ascertained for each patient. Age- and sex-specific injury incidence rates are computed. Nearly 36% of all job-related injuries are treated in hospital emergency rooms. This data resource provides timely data on occupational injuries, and thus facilitates intervention through follow-up investigations of selected injury types. However, it must be noted that injuries caused by cumulative trauma, such as the carpal tunnel syndrome, are not likely to be treated in hospital emergency rooms. Also firms that refer workers to hospitals for treatment may differ from those that have other arrangements for treatment. Another weakness of this system is that the industry and occupation of an injured worker are not reported in this system.

The sixth of the data systems on occupational disease which we employ is the National Health Interview Survey, an annual survey of 40,000 households-not necessarily the same households each year-which is conducted by the National Center for Health Statistics. This survey, based on a representative sample of all U.S. households, provides valuable information on the prevalence of disability and disease and on utilization of health services. The survey also collects information on the industry and occupation of each respondent, and these data are used by NIOSH to identify unusual patterns of non-disabling illnesses among U.S. workers.

The major system used by NIOSH to monitor potential exposures to occupational hazards in the United States is the National Occupational Hazards/Exposure Survey. In each round of this survey which was first conducted by NIOSH in 197274, and which has just been completed for a second time, field investigators from NIOSH surveyed approximately 5,000 worksites in the United States to provide nationwide information on potential exposures to workplace hazards. Data obtained from this survey enabled us to estimate the numbers of workers who are at risk to various chemical and physical hazards, and in what industries and occupations these workers are found. Data from this survey are used (1) in setting priorities for NIOSH laboratory and field investigations, (2) to characterize the array of hazards

common to industries, (3) to understand the chemical constituents of trade name products observed through the survey, (4) to model the risk of exposure in particular industries or occupations by the disease, disability or mortality risks observed in these populations.

In addition to these seven systems for data collection, we are also quite active in detecting problems in the workplace by our numerous investigations in workplaces around the country. We refer to these investigations as "health hazard evaluations." We perform about 500 such field investigations each year. These investigations are carried out in response to requests for assistance from employers, employees, employee representatives, other federal agencies and State and local health agencies. Our investigations determine whether or not workers are ill, or are exposed to hazardous agents in the workplace. In our reports, we make recommendations to the parties involved regarding control procedures, improved work practices, as well as medical tests and environmental monitoring where needed.

I would very much like to tell you, Mr. Chairman, that these data collection systems and our field investigations are combined together to constitute a comprehensive reporting system producing a meaningful synthesis of the Nation's problem of occupational diseases and injuries. These activities do not provide a comprehensive epidemiologic surveillance of occupational diseases and injuries in the United States. This lack of adequate surveillance activities is one of the reasons why all participants in occupational safety and health do not share a common view as to how these problems should be attacked, and thus, one of the reasons why occupational safety and health often tends to be so adversarial, in sharp contrast to other disciplines in the public health field. I believe comprehensive information for occupational diseases and injuries is imperative if we are to protect the health and lives of workers. Unless our efforts are targeted toward comprehensive data collection and synthesis, the confusion will only grow worse.

NIOSH has several initiatives already under way within the agency with the cooperation of private employers and at the Federal, State and local levels that recognize the need for and seek to improve the quality of occupational safety and health surveillance data. I will discuss these in greater detail later in my statement.

No system exists which provides us with uniform national information on the occurrence of the major categories of occupational diseases and injuries. We therefore lack the ability to define comprehensively our problem in a way that facilitates an efficient attack on it. The limitations of the existing data bases often preclude our obtaining the data by which to follow such items as simple trends in incidence of the principal occupational diseases over a period of time. In attempting to exercise some leadership in this field and provide the Nation with syntheses of these problems, we in NIOSH are forced to look at data coming from the several disparate systems which I have described, and to attempt by inference to bridge the gaps between them. Despite these limitations, there are encouraging signs that this problem is being addressed.

Of course I am not the first person to identify this problem. Since the passage of the Occupational Safety and Health Act in 1970, at least eight government-sponsored studies have been conducted which have made recommendations designed to improve data collection systems in the field of occupational health. Indeed NIOSH established a task force on occupational health surveillance which in 1977 made a series of 10 recommendations for occupational hazard surveillance and an additional five for occupational disease surveillance. The 10 recommendations for hazard surveillance were taken into consideration and are reflected in our current hazard surveillance program.

The development of a sampling statistics decision model for occupational hazards is reflected in the design of the National Occupational Exposure Survey (NOES). Cost-benefit analyses of hazard surveillance strategies supported the decision to repeat a national survey of potential hazards in the workplace, as well as other activities geared to accessing and using data on workplace hazard; e.g. OSHA inspection data.

Incidental to the completion of the first National Occupational Hazard Survey, NIOSH developed a functioning system for resolving trade name products. This is being replicated in the course of the NOES.

Two recommendations urged NIOSH to develop mechanisms to identify and characterize American industrial facilities, as well as the worker population at risk in these facilities. Through the use of existing data available through the U.S. Bureau of the Census, as well as commercially available resources such as Dun & Bradstreet, we are able to characterize and profile hazards by industry and occupation. These characterizations may be derived at the national, State, or local level.

A number of recommendations concerned the development of mechanisms for early recognition of occupational hazards. NIOSH was encouraged to collaborate in the overall federal effort under the Toxic Substances Control Act (TSCA). We interact with EPA on a routine basis, relative to TSCA. Apart from being a source of information on potential workplace hazards, this collaboration facilitated NIOSH research on specific hazards.

Two recommendations of the Task Force involve the NIOSH health hazard evaluation (HHE). NIOSH has computerized its HHE reports, and maintains the necessary staff and equipment for the sampling of hazards in the field. NIOSH maintains the Registry of Toxic Effects to Chemical Substances (RTECS). NIOSH was encouraged to expand abstracting of reports reflected in this file. This has been done; e.g., abstracting on "target organs" in bioassays.

The five recommendations for disease surveillance were also taken into consideration and are reflected in our current disease surveillance program.

NIOSH was encouraged to assess the feasibility of adapting population-based tumor registries for occupational cancer surveillance. NIOSH supported two feasibility studies relating to this recommendation. One study examined the utility of these registries for monitoring selected occupational cohorts for cancer incidence. The other study examined the utility of these registries for indepth care studies of cancer and occupation. This past month, we approached the National Cancer Institute and encouraged their cooperation in assessing the feasibility of cancer registries routinely collecting, coding, and analyzing information on the cancer patient's occupation and employment.

NIOSH was encouraged to use available National Center for Health Statistics (NCHS) data and surveys. We use the Health Interview Survey (HIS) to charaterize the morbidity patterns of selected industrial populations. We used the National Health and Nutritional Examination Survey (NHANES) to examine the relationship of NHANES measure of blood lead with the potential for lead exposure. We were a principal supporter of the National Fetal Mortality and Natality Surveys. And currently, we are collaborating with NCHS to provide for the routine collection, coding, and analysis of decedent employment data as derived from death certificates. These efforts will result in a national data system which will facilitate the same objective-to characterize occupational and industrial characteristics of decedents in the United States, the first step toward a national occupational mortality surveillance system.

NIOSH has developed the procedural models to facilitate prospective follow-up studies. This is represented by certain uses currently made of the data from the Social Security Administration, the Internal Revenue Service, and the National Death Index (NDI). the NDI, developed under the NCHS, was supported by NIOSH, NCI, and other federal agencies in recognition of the record linkage deficiencies present in the U.S.

During the first year of my directorship of NIOSH, In reviewing progress toward the 1990 Prevention Objectives of the U.S. Public Health Service, we recommended to the Service that implementation of the occupational safety and health surveillance objectives be given high priority.

As a guide toward the prioritization of these surveillance objectives (and also toward prioritization of other research within the Institute), NIOSH has developed a suggested list of the 10 leading work-related diseases and injuries; (1) occupational lung diseases. e.g. silicosis, lung cancer, (2) musculoskeletal injuries, e.g. disorders of the back; (3) occupational cancers (other than lung), e.g. cancers of the bladder, nose and liver; (4) amputations; (5) cardiovascular diseases, e.g. hypertension, acute myocardial infarction; (6) disorders of reproduction; e.g. spontaneous abortion; (7) neurotoxic disorders, e.g. peripheral neuropathy; (8) noise induced hearing loss; (9) dermatologic conditions; and (10) psychologic disorders, e.g. alcoholism and drug dependency.

Three criteria were used to develop the list: the disease's or injury's frequency of occurrence, its severity in the individual case, and its amenability to prevention. The list is intended to be dynamic; it will be reviewed periodically for necessary updating as surveillance systems improve, knowledge increases, and as conditions change or are brought under better control. Occupational disease surveillance systems will be used to describe and monitor these ten leading work-related diseases and injuries.

In the implementation of our plans for the establishment of a uniform national system for the surveillance of occupational disease in the United States, we have worked closely with the State Health Departments across the nation. First, through a program of surveillance cooperative agreements between NIOSH and States (SCANS), NIOSH supports occupational health and safety surveillance activities

within state health departments. Six states have received NIOSH funding-Maine, New York, North Carolina, Pennsylvania, Rhode Island and Utah. These States have collaborated closely with NIOSH under this program.

Since the award of the first of these cooperative agreements in 1980, personnel within State health departments have been supported in their efforts to augment State health data bases for occupational health surveillance purposes. This support has led to the States' implementing: Periodic analyses of occupational and industrial mortality differntials; selected monitoring of birth and fetal death records to identify parental employment risk factors; collection and reporting to NIOSH of sentinel health events (occupational); and establishment and maintenance of occupational health and safety data bases.

The by-products of this program are disseminated in a variety of ways. NIOSH supports its State counterparts in presenting the results of State-initiated investigations through professional societies and technical journals. Within States, the results are communicated to the chronic disease and occupational health components of the health department. NIOSH furthers the dissemination of this State effort in two ways. First, selected reports from the States are considered for publication through the NIOSH Surveillance Report Series. Second, data tapes from the States are analyzed by NIOSH personnel. Findings are compared across States for common disease patterns. These results are communicated to other components of NIOSH for appropriate investigative follow-up, and facilitate priority setting for the Institute.

In another example of our cooperation with the State health departments in the surveillance of occupational diseases, we have worked closely with the Conference of State and Territorial Epidemiologists (CSTE) to encourage the passage by that body of a resolution endorsing the inclusion of silicosis in the list of "specified notifiable diseases" which are reported weekly to CDC. This would be the first noninfectious occupational disease ever put on that list. Specifically, CSTE in June 1982 resolved to collaborate with NIOSH in establishing a reporting system for such occupational diseases as coal worker's pneumoconiosis, silicosis, lead poisoning, asbestosis and mesothelioma. At their 1983 meeting CSTE resolved that silicosis should be the first nationally reportable, non-infectious disease. Yesterday, at their 1984 meeting, NIOSH staff presented interim definitions and plans for reporting coal workers' pneumoconiosis, silicosis, asbestosis and mesothelioma.

In another surveillance initiative, which occurred late in 1983 Dr. David D. Rutstein of the Harvard Medical School and a group of scientists from our Institute published a list of so-called "occupational sentinel health events." A "sentinel health event" is a "preventable disease, disability, or untimely death, whose occurrence serves as a warning signal that the quality of preventive and/or therapeutic medical care may need to be improved." Today, a single case of polio would be considered dramatic sentinel health event.

A "sentinel health event-occupational" is defined as a condition, work-related, which signals a breakdown in prevention in the workplace. Such breakdowns demand investigation and intervention. The published list of sentinel health events occupational now includes 51 entries. It could well serve as a framework in developing a national system for occupational disease surveillance as well as a guide to practicing physicians to assist them in identifying patients who have work-related diseases. The list is both too long-51 is too great a number to expect that all conditions will be universally reported, and too short, in that the present list contains no entries of occupational injuries.

What then should be done? One could make many recommendations and indeed a recent subcommittee (Occupation Cancer Risk Subcommittee) of the DHHS Committee to Coordinate Environmental and Related Programs, in looking at existing data systems, made 23 recommendations, most of which are quite complex. Permit me to elaborate on three high priority recommendations on this list.

(1) Complete the analysis of the data obtained in the 1983 National Occupational Exposure Survey (NOES). We are on schedule with this. The result will be an updated list of over 90,000 potentially dangerous chemicals used in the factories, shops, hospitals and offices in the United States.

(2) Explore the feasibility of using NOES data to identify high interest worksites for follow-up surveys in which measurements of exposures to workplace hazards would be made. An ad hock subcommittee of PHS scientists has been charged with the responsibility of exploring the feasibility of conducting this type of follow-up

survey.

(3) Link NOES data with environmental sampling data collected by OSHA and other agencies to develop better estimates of the extent to which workers are exposed to workplace hazards. This would be a step toward meeting two of the three

requirements for more refined risk assessment, i.e. hazard identification and assessment of workplace exposure.

I believe that every effort should be made to link, with appropriate safeguards for privacy, the various data systems existing within the Federal Government so that agencies and individuals interested in preventing disease and disability among workers can share a common data base. There exist legal barriers to those sorts of linkages. Under the leadership of the Office of Management and Budget there have been explorations of a data enclave that would permit governmental agencies to exchange individually identifiable personal records for defined purposes of statistical research, including the types of epidemiologic studies that would contribute to the identification of work-related illness. An ad hoc subcommittee of the CCERP is now exploring the steps that need to be taken for more effective sharing of data. This can be done in such a way that the privacy rights of workers are protected.

Secondly, I believe that we should do everything we can to encourage occupational physicians and nurses to share information. Sophisticated computer networks are now possible which could link and combine the experience of the thousands of dedicated physicians and nurses who are seeing occupational problems daily. Third and most important, I believe that occupational problems must become part of the mainstream of disease reporting in the United States. That means that the States must define certain occupational injuries and diseases as reportable, and ask that these be reported to the public health agency in the city or State. In turn the States should be encouraged to report this information to the Centers for Disease Control as is done for many conditions which are not occupationally-related.

In occupational safety and health, the problem, in brief, is that there are numerous systems which produce reports of death, reports of claims, reports of lost work days, reports of injuries, and other information regarding work-related diseases and injuries. None of these systems, nor indeed all of them taken together, provide a systematic national surveillance of work-related health problems.

Because surveillance of occupational disease and injury is the first step toward meeting the nation's commitment to assuring a safe and healthful workplace, we at NIOSH are attempting to improve data collection and enhance the usefulness of surveillance systems.

Mr. Chairman, thank you for the opportunity to present my views on occupational safety and health surveillance. I would be happy to try to answer any questions you or your colleagues may have.

Mr. FRANK. Thank you very much, Dr. Millar.

I, without objection, will put the entire statement in the record with all the other documentation you have and I appreciate the testimony, which was very much on point.

Let me just begin with a couple questions. You mentioned the step forward in having silicosis be reported as the first noninfectious disease. I guess I don't fully understand who makes the decision as to what diseases are reported and how we might be able to influence that.

Dr. MILLAR. Well, the States decide which conditions will be reportable within their jurisdictions.

Mr. FRANK. In each State.

Dr. MILLAR. They now have a list of approximately 35 conditions that are notifiable, reportable in every State and territory in the Union.

Mr. FRANK. Is that basically coordinated by this conference that Dr. Landrigan was in?

Dr. MILLAR. That is right.

Mr. FRANK. They decide it?

Dr. MILLAR. Yes; they would decide to add or detract from this list and then agree voluntarily to report to the Federal Government by telegram every week.

Mr. FRANK. What would happen if—and/or the reporting goes to you?