With success that has surprised even the most radical among them, AIDS patients and their relentless advocates have in less than three years transformed the way experimental drugs are tested in the United States. Enraged by what they saw as the glacial pace of federal AIDS research, activists demanded a revolution -- and they got one. Articulate and unyielding, gay leaders who were once openly regarded by federal scientists as renegades now participate in nearly every key government decision involving AIDS. "In the beginning, those people had a blanket disgust with us," said Anthony S. Fauci, director of AIDS research for the National Institutes of Health. "And it was mutual. Scientists said all trials should be restricted, rigid and slow. The gay groups said we were killing people with red tape. When the smoke cleared we realized that much of their criticism was absolutely valid." Over the past two years, AIDS activists have forced a fundamental shift in the way drugs for fatal illnesses are tested. Before AIDS, experimental drugs could only be given to patients in formal trials where their effects were compared with those of a placebo or of another drug. Only after such trials had established the safety and efficacy of a drug beyond doubt did the FDA permit others to use it. All that has changed. Doctors of AIDS patients routinely help them obtain drugs long before they have been proven effective. A major network of community programs has for the first time established alternative drug trials that guarantee greater access to new medicines. And the FDA itself has halved the time it takes to approve most new drugs for fatal illnesses. Previously, federal health officials held as a basic principle that it was irresponsible to release drugs before they had been proven effective, in large part because scientists feared easy access would discourage patients from joining a study where they might not get the experimental drug. Since the drug AZT was approved for AIDS treatment in 1987, however, government regulators have come to realize that carefully broadening the group of people who can receive promising experimental drugs may not deter people from entering controlled studies in which they risk getting an old drug instead of the new one. Similar shifts in policy have begun to apply to drugs for cancer and may be extended to Alzheimer's disease and other fatal illnesses. Last week, Fauci said he would favor a system where experimental drugs in the early stage of clinical study could also be given to those outside the study who might benefit while the study continues. But he stressed that the plan would have to be applied on a case by case basis. "Its been astonishing to see the magnitude of the changes," said Jim Eigo, an AIDS activist who has devoted himself to mastering the complexities of drug development. Eigo is a member of AIDS Coalition to Unleash Power (ACT UP), a political AIDS organization that has used everything from rage to reason in attacking the federal AIDS research program and vilifying its leaders. "When they finally opened their eyes," Eigo said, "they began to realize what we were saying is right." Because AIDS has afflicted gay men disproportionately, gay groups have been at the forefront of efforts to deal with the epidemic. The far-reaching changes have not come easily. AIDS activists often say that drugs are now much more readily available in spite of federal efforts rather than because of them. And scientists trained to require rock-hard proof that drugs work say premature release will often do more harm than good. Despite several initiatives by the FDA and NIH, many activists remain suspicious of the federal AIDS program. And among researchers there is a great fear that if drugs are too widely distributed before much is known about them, scientists will have lost their only real chance to figure out how -- or, more importantly, whether -- they work. To the surprise of many, Fauci, along with officials from the FDA and other AIDS researchers, have begun at times to sound almost as if they were shadow spokesmen for groups like ACT UP. In return, many ACT UP leaders have gone from handing out "Fauci=Death" posters at AIDS meetings to calling for well-designed experiments that help answer essential questions about new drugs. Their most consistent point is that as soon as a drug appears to be safe -- but before it has proven effective -- it should be distributed more widely to people with no other medical hope. "It makes life a lot more complicated," FDA Commission Frank E. Young said of granting wider access to experimental treatments. "But the system isn't supposed to be for us. It's supposed to be for patients. And I think if we are very careful, we can often get the results we need while distributing drugs to a wider group." At meetings with community activists, public health officials and politicians, many ranking federal AIDS specialists are delivering their new message: We will do everything we can to get drugs to more people and to get them out more quickly than we ever thought possible. "The activists are skillful and smart, and they know how to get attention," said Louis Lasagna, professor of medicine at Tufts University School of Medicine and chairman of a influential national committee now holding hearings on drug development. "But we are still playing with fire. Because if you make drugs available without the necessary information, you do something we will clearly regret." Striking the proper balance between speeding access to powerful new drugs and carefully assessing their value has been the most troubling issue for many people involved in the AIDS epidemic. For years, there has been constant tension between the need to get treatments to desperate patients and the reality that many drugs that first appear promising, even miraculous, turn out to be worthless or even dangerous when they are more fully tested over time. Almost every month, some new drug sweeps underground networks with great fanfare. Dextran sulfate, ribiviran and AL-721 have all at one time had huge cult followings among AIDS patients eager for any drop of good news. But none of them has proven effective, and most researchers feel confident that at best they are useless and expensive. Last week, when two men died in San Francisco while participating in a clandestine trial of a new underground drug called "Compound Q," the conflict became more painfully apparent. The drug, derived from proteins in a Chinese cucumber, is highly toxic and completely unproven as an AIDS treatment, although it is now being tested at San Francisco General Hospital. In early experiments, it appeared to kill immune system cells infected with HIV, the virus that causes AIDS, while leaving other cells alone. But it will be months or years before the official trials yield conclusive results. If the drug works in people as well as it does in the test tube, it could be a very effective therapy. But as has been the case since the epidemic began, dying people are seldom willing to wait for trials to be completed before they use the drug. So Martin Delaney, head of Project Inform, the nation's leading non-governmental AIDS treatment information organization, decided to take matters into his own hands and organize a secret Compound Q study involving 42 patients. He says he did it with elaborate care and great precautions. Most scientists and many AIDS activists disagree. "When you don't have any idea in the world whether a drug is going to work or kill you, it's outrageous to put people on it," said Jere Goyan, a former FDA commissioner and currently dean of the school of pharmacy at the University of California at San Francisco. Goyan has argued strongly throughout the epidemic that federal regulators need to become more flexible and alter many of the scientific requirements of the drug-approval process. But he said that without requiring some evidence to justify wider use, chaos would develop. "If that's the approach you are going to take, we might as well just hand out every possible drug to whoever comes by," he said. "We run the very real risk of mortgaging our future by throwing away the ability to judge these drugs." At times, there is so much confusion over when to permit wider use of experimental drugs that even the separate agencies of the Public Health Service are out of tune. On June 16, for example, a task force led by NIH researchers proposed in a publication from the Centers for Disease Control that some people infected with HIV should receive a drug, aerosol pentamidine, for a deadly pneumonia as a preventative even if they were well. But the FDA had not approved the drug for that purpose yet, although it clearly intended to, and the recommendation set the agency on a mad scramble to approve pentamidine lest it appear it was behind the pace of NIH and the CDC. "If the consequence of their failures weren't so deadly, you could laugh," said Delaney of Project Inform. "Unfortunately, we don't see anything funny going on."