When federal health officials released the experimental drug DDI for early use by AIDS patients a year and a half ago, they thought they had resolved a difficult political and medical dilemma.

The drug was so new that there was not enough evidence of safety and effectiveness to warrant Food and Drug Administration approval. What little data existed, however, were so promising that thousands of AIDS patients were more than willing to take their chances and try the drug. So the agency compromised, permitting DDI to be given to some patients while formal evaluation of the drug continued.

At the time, this compromise -- known as parallel track -- was considered a bold step for the FDA, a regulatory milestone prompted by the AIDS epidemic.

But now, in the first major initiative of new FDA Commissioner David A. Kessler, an agency task force is considering extending the logic of parallel track still further, by creating a set of regulations intended to make it even easier to get promising new drugs to a wide range of patients as quickly as possible.

Under the plan, the agency would not just let a promising experimental drug such as DDI be given out to selected patients during testing. Instead, it would grant the drug temporary approval. The drug would then be sold by pharmaceutical companies and could be obtained by patients like any other drug.

The government, however, would also mandate ongoing testing of the drug. And if the drug failed to live up to its original promise, it would be withdrawn.

The proposal -- known as conditional approval -- has its critics. At the least, many health policy experts say, it will depend for its success on how cleverly the FDA structures the program and whether the agency manages to resolve fundamental questions about who will pay for experimental drugs and how they will be tested.

But if successful, it could be an important and welcome addition, experts say.

"Not all drugs are alike and not all diseases are alike. So the regulatory pathway for all drugs should not be alike," said Lawrence Horowitz, a member of the federal advisory committee on the FDA. "The FDA needs a flexible range of tools."

The advantage of conditional approval, health policy experts say, is that experimental drugs could reach far more patients than they would by being given out on a compassionate basis. Many AIDS activists, for example, have become frustrated with the availability of DDI, because it can only be given out to those AIDS patients who no longer can tolerate AZT, the only approved drug for treating AIDS.

In addition, conditional approval would allow companies to recoup their investment on experimental drugs. By contrast, the parallel track -- also known as Treatment Investigational New Drug (IND) -- only allows a minimal level of cost recovery for drugs distributed under the program, and most companies simply give their products away.

In the case of DDI, this has meant a tremendous finanical burden for the drug's manufacturer, which is distributing the drug to 10,000 patients free of charge. Some AIDS activists worry that this experience may cause other companies to shy away from the AIDS field entirely.

"The current system doesn't discourage companes, but it doesn't encourage them either," said Martin Delaney, director of the AIDS activist group Project Inform. "There's nothing in it for them."

The key to the new proposal is how insurers react. Many health insurance companies have declined to reimburse patients for costs of experimental therapies, saying that they have no obligation to pay for treatment that has not been proved fully effective. The theory is that if experimental drugs are approved -- albeit conditionally -- insurers will pick up the tab, taking the financial burden off the pharmaceutical industry.

But if the insurers balk?

"My guess is that some proportion of insurance companies are going to say this is only conditional approval, not final approval," said Wayne Roe, president of Health Technology Associates in Washington. "The more conservative companies are going to say, 'The jury is still out -- we're going to wait until the FDA really weighs in.' "

Another potential problem is deciding when a drug is promising enough to grant conditional approval. The agency would like to release drugs as early as possible. But some experts caution that after approval, further testing may be difficult.

Treatment IND, for example, is based on the principle that a company first designs and begins human trials needed to test a drug and then releases the products in such a way as not to jeopardize the results of those tests. Generally, such a drug is given only to those who could not qualify for the trials.

Even so, some scientists have expressed concern that early release of promising drugs has made it difficult to run properly designed clinical trials.

If an experimental drug were conditionally approved, however, companies would not be able to determine which patients might get it, as they might need to do to set up a clinical trial.

"If the Treatment IND has slowed down the process {of collecting data on effectiveness}, then conditional approval might bring it to a halt," said Louis Lasagna, dean of Sackler School at Tufts University and one of the nation's leading drug development experts.

"This will make it more difficult to do the restrictive kinds of trials that we have liked to do in the past," said Delaney. "We're going to do real world trials, looking at a larger number of people using a drug in a broader, less structured context."