accelerated approval

Accelerated approval is a route the US Food and Drug Administration created in 1992 for treatments aimed at serious conditions that fill an unmet medical need. Proving that a treatment changes how a patient survives, feels or functions, what the agency calls clinical benefit, can take many years. Accelerated approval lets the agency license a treatment earlier, on evidence that it moves a marker instead: a laboratory measurement, an image, a physical sign or another measure that is thought to predict the benefit but is not itself a measure of it. A 2012 change to the law added measures of a therapeutic effect considered reasonably likely to predict benefit.

The FDA's own worked example is cancer. Rather than waiting to learn whether a drug extends survival, the agency may license it on evidence that it shrinks tumours, because shrinkage is reasonably likely to predict that patients live longer. The company must still run studies to confirm that the prediction holds.

That confirmation is a legal obligation, not a courtesy. The licence carries a required trial with dates fixed by the agency, and progress reports until it finishes. Where the trial verifies the benefit, the requirement is generally lifted. Where it fails to verify it, or shows a benefit much smaller or shorter than the marker suggested, the FDA may withdraw the approval or change what the treatment is licensed to be used for. Until that trial reports, an accelerated approval is a licence resting on a prediction.