Federal agencies know the names of hundreds of thousands of people who have been exposed to cancer-causing chemicals on their jobs, but have made no effort to tell them about the risk to their health.

Nor has any effort been made to name or notify 21 million workers -- one in every four -- who are known to have been exposed to hazardous materials regulated by the Occupational Safety and Health Administration (OSHA).

The National Institute of Occupational Safety and Health (NIOSH), a primary holder of this information, has used the lists of names only for research purposes, waiting for the people on them to die. Then the names are followed up, and if the causes of death can be learned, they become part of the tally the government uses to set up control standards for dangerous substances.

This week, a joint notification pilot project of NIOSH and the Workers' Institute for Safety and Health, an arm of the AFL-CIO, is getting off the ground for 1,100 chemical plant workers in Augusta, Ga., and another pilot program is under way for asbestos workers in Port Allegany, Pa.

For the Georgia chemical workers, the notification program comes nine years late: NIOSH learned of their exposure to a known carcinogen in 1972. But notification for thousands of other workers in the NIOSH files is still several years and billions of dollars away, if it comes at all.

The lists of names are on file at NIOSH and the National Cancer Institute, as well as at labor unions and in university research centers across the nation. They are the microfilmed records of current and former employes from job sites where researchers studied substances suspected of causing cancer. NIOSH officials estimate that their lists alone, which detail only places later found to involve carcinogens, include about 200,000 names.

While the labor unions who hold the information have made efforts, through publications or meetings, to get the word to their members, a majority of the workers involved are not unionized. In other cases, the exposed workers have since left their jobs.

The debate over what a government agency should do when it is reasonably certain that former workers have been exposed to a dangerous substance is not a new one. The uneasy conclusion has been that workers have the right to know about risks to their health, but that it would cost too much for the government to find everyone, notify them and then provide follow-up services.

Many bureaucrats think it is immoral to object on grounds of cost, but they never have agreed on who should have the responsibility of finding and notifying everyone, who should pay for it and the best way to do it without scaring the workers into filing millions of lawsuits. As a result, nothing has been done.

"You can't just send out a letter saying, 'Hey, you might have been exposed.' You have to have some kind of mechanism to take care of them once they know," said Philip Bierbaum, deputy director of NIOSH's health hazards division. "That could cost hundreds of millions of dollars, and who's going to pay for it?"

NIOSH estimated at a 1977 Senate hearing that it would cost up to $40 million to find and notify all of the estimated 21 million workers who have been exposed to a hazardous substance, and that giving them all medical surveillance -- not treatment -- afterward, just to spot developing illness, would cost a staggering $54 billion. The figures have risen since then but the estimate has not been updated, according to NIOSH executive officer Dr. Ron Coene.

"We see our role as research, not as a public health agency," Coene said. "We publish our results in technical reports not geared to workers but for our peers and the regulators. The worker is not our primary audience."

But workers are NIOSH's primary subject. Armed by Congress with the right to enter any workplace -- a right confirmed by the U.S. Supreme Court -- and the right to subpoena records for research purposes, NIOSH does two kinds of studies.

One, called a health hazard evaluation, can be sparked by a complaint from three workers, an employer or a labor group and includes air and water sampling in a workplace, questioning employes and listing the materials they are handling. There may be follow-up physical examinations, including blood and urine sampling for the employes, and the findings of these tests are always reported to the workers.

The overall findings of the inquiry, however, are not.

If the overall finding is that some danger to the worker may exist, NIOSH notifies the company officers and any union at the plant, relying on them to spread the word. The plant employment records are microfilmed as far back as they go, becoming part of what NIOSH calls a "cohort," the group of people all presumably exposed to that substance. These lists are filed for research purposes.

Nonunion workers, who comprise 70 percent of the labor force, and former employes, even those who quit the day before NIOSH arrived at the scene, may never know about their exposure.

"We do a minimum notification," said Coene. "We don't make the individuals aware of the results of the study they were a part of, the collective decision on what's going on there." Why not? "It just never has been done," he said.

The second kind of NIOSH study looks into a substance or a process on an industry-wide basis. These probes, which are launched either by the findings of a health hazard evaluation or by outside research results, generally check a few selected plants typical of the industry or the process. The files are microfilmed and the names added to the "cohort."

In some cases NIOSH writes to a selection of names from the cohort, asking for health information for a study of the substance involved. "It says if you have questions call this number," a NIOSH worker said, "but it doesn't mention the reason for the study."

In most cases, however, NIOSH has no personal contact with anyone in the cohort. It only wants to know which people on the list have died.

"We contact Social Security, the Internal Revenue Service, the Post Office, the bureaus of motor vehicles, credit unions," said Philip Bierbaum, deputy director of NIOSH's industrywide studies group in Cincinnati. "It's a records study. We seldom contact individuals; maybe occasionally a next of kin."

IRS law has a specific exemption to provide NIOSH with taxpayers' addresses. NIOSH also uses military, veterans', labor union, insurance firm and state demographic records.

When the lists of the dead are obtained, NIOSH then tries to get death certificates, which it codes for cause of death and compiles for statistics. Once NIOSH decides those statistics show an increased risk of cancer or some other illness for exposed workers, it proposes to the Department of Labor that worker exposure be regulated. NIOSH has sent in 105 such recommendations since it was set up in 1970, but OSHA has only established regulations for 10 of those, according to NIOSH.

More than 500 substances, including 22 carcinogens, are regulated as hazardous by OSHA, the Environmental Protection Agency and other agencies, but workers may not know they are being exposed to them. Legislation requiring labels of industrial products to carry health warnings has repeatedly failed to make it through Congress.

NIOSH's research technique is well established. By finding, for example, that a higher number of union workers exposed to asbestos had died of lung cancer than the proportion in the general population, Dr. Irving Selikoff of the Mt. Sinai School of Medicine in New York established an apparent link between asbestos and cancer in 1964.

Selikoff made his findings known to the union and like many other researchers, thinks that is sufficient. "We've always worked with groups, and the groups can tell their members better than we can," he said. Asbestos is an example of the questions that notification raises. An estimated 13.2 million workers in construction, power plants and shipyards were exposed to it, Selikoff said. "Obviously it's too many people to send letters to," said one NIOSH worker. "But maybe it is feasible to send letters to 80 former workers at some small plant. Then where do you draw the line?"

In 1978, Joseph A. Califano Jr., then secretary of health, education and welfare, set up a task force to try to notify just the 5 million people who had worked with asbestos in World War II shipyards. Dr. David Rall, now head of the National Institute for Environmental Health Sciences, was co-chairman of that group.

"We used little TV spots," Rall recalled. "The surgeon general mailed a letter to every physician in the country. We tried to convince people to stop smoking" because both smoking and asbestos exposure seemed to increase the lung-cancer risk. The National Cancer Institute set up information programs for doctors and veterans' groups. The six-month effort cost $450,000 and reached millions of people.

Dr. Diane Fink, who co-chaired the program for the NCI, said smaller publicity campaigns later sought to reach the estimated 4 million people whose mothers had taken the cancer-causing fertility drug DES, and an untold number who had received head X-rays for skin problems and were in danger of thyroid cancer.

"We worried that there might be panic" among the affected people, Fink said, "but we decided it was more important that they know." She said she knew of no panic reactions triggered by the campaigns, even though there was no effort to provide follow-up services.

Labor leaders agree, however, that the follow-up is essential. Workmen's compensation laws, which date from the early 1900s, often require that claims be filed within short periods after the injury or exposure occurs. Knut Ringen of the AFL-CIO's Worker's Institute for Safety and Health said only 5 percent of workers are now covered under workmen's comp for illnesses like cancer that develop over a long period.

The AFL-CIO has launched what it calls a major new drive to deal with worker health problems. "There has to be exposure notification," said Joe Velasquez, director of the Workers' Institute. "Society has the responsibility to take care of these people. It's blaming the victim to say the worker has to take care of himself."