FDA Approval

The FDA is the health authority of the United States. It examines whether drugs, vaccines, and medical devices may be sold. Without its green light, a manufacturer may not offer its product on the US market. This very permission is called FDA approval. In doing so, the agency decides not only whether a product is safe. It also examines whether it truly delivers what the manufacturer promises.

Why a letter from an agency moves billions

The United States is the largest healthcare market in the world. For many pharmaceutical companies, it is there that it is decided whether ten years of research have paid off. An approval can multiply a company’s revenue in one stroke. A rejection, on the other hand, can render a small biotech company practically worthless.

That is why approval dates are fixed entries on the stock market calendar. Analysts estimate the probability of success in advance, and the share price reflects this expectation. If the decision turns out differently than expected, price jumps of 30 percent or more in a single day are nothing unusual. Such headlines are frequently found in financial news.

A second reason is the signaling effect. Other countries do conduct their own reviews, but often orient themselves toward the FDA. Whoever has cleared the US hurdle usually has an easier time in Europe or Japan. FDA approval is thus a kind of international seal of quality.

The path from idea to market

It begins with laboratory and animal testing. Only after that may a substance be tested on humans, and this happens in three successive study phases. Phase one tests mainly for tolerability in a small number of healthy participants. Phase two seeks the right dose, phase three compares the substance in thousands of patients against an existing treatment or a placebo.

Afterward, the manufacturer submits all data to the agency. This package often comprises hundreds of thousands of pages. FDA experts recalculate the studies and frequently convene an independent advisory panel. In the end, there is a clear yes, a no, or a request for additional data. From the first laboratory experiment to the decision, ten to fifteen years typically pass.

An important distinction is often overlooked. Drugs go through the full review process, whereas many devices undergo only a leaner procedure. There, it is sufficient to demonstrate that a product closely resembles one that has already been approved. This very pathway is currently being used for software with artificial intelligence, that is, for programs that learn patterns from data.

AI software under the agency’s scrutiny

By now, the FDA has cleared more than a thousand products that work with learning programs. Most of these analyze images: X-rays, retinal photographs, or skin lesions. Such a system marks, for instance, a suspicious spot on a CT scan so that the doctor does not overlook it. Such approvals regularly appear in the quarterly reports of medtech companies.

For learning systems, however, there is a particular problem. Classic products remain unchanged after approval, whereas software is continuously improved. A model that decides differently after every update would, strictly speaking, be a new product. The FDA therefore requires manufacturers to describe planned changes in advance and to monitor the results.

A common misconception is that FDA approval declares a product to be perfect. It only means that, based on the available data, the benefits outweigh the risks. Approvals are also withdrawn again if new problems arise in everyday use. For investors, this means: a yes from the agency is a milestone, but not a guarantee.