A short history of US drug regulation
Part one: safety and efficacy
I have started a series of posts sharing the research I did working on an Emergent Ventures grant in early 2020. This is the second post. The first post is here.
I said before that the world gets what the US asks for when it comes to new drugs. So what does the US ask for?
There are three key features in the drug regulatory regime and market. Firstly there is the approvals side that decides what drugs one is allowed to sell or use – in the USA this is basically the Food and Drug Administration, the FDA. Secondly, there is the patent and exclusivity regime that determines who is allowed to sell or supply them. Thirdly there is the payment side: what payers like the NHS or US insurers will pay for. This post will deal with the approvals side.
Until 1938, one could basically market any drug in the USA – including, famously, snake oil. A cough syrup disaster in 1938 killed hundreds, and the FDA started taking an interest, though it was still a pretty liberal regime. Drug firms had to carry out safety trials and submit them to the FDA. If there was no response in 60 days they could go ahead and market the drug.
In the 1940s and 1950s trials were pretty basic, typically lacking a control group or blinding. Randomised controlled trials evolved later, when Britain faced a serious antibiotics shortage. Since they were unable to give everyone drugs, researchers decided to randomise who got them to at least learn something from the experience, which was that streptomycin was a very effective treatment for tuberculosis.
Strangely enough, it was another safety scandal that led to the FDA bringing in an efficacy standard. One reviewer in the FDA team, Frances Kelsey, repeatedly refused to approve thalidomide due to a worry – later famously confirmed – that it caused birth defects.
In 1962, the FDA’s powers on safety were beefed up. Safety trials now had to be overseen by qualified investigators and patients had to give informed consent to physicians that trialled drugs on them. Drugs would no longer automatically be approved if the FDA didn’t reply in 60 days. And the FDA could now force drugmakers to withdraw their products after the fact if they believed them to be unsafe.
More importantly, drugs had to be proven to be effective. It’s not obvious why thalidomide, which was a straightforward safety issue, led to this change. In fact, thalidomide is highly effective and notwithstanding the horrible complications for pregnant mothers, is widely used for multiple applications, including against multiple myeloma. It appears on the World Health Organisation’s list of the most essential medicines for a healthcare system. Eventually the US did approve it in 1988.
Either way, from 1962 firms needed to provide ‘substantial evidence’ that drugs they wished to market would have the effect they promised. On its own this would have been a fairly weak standard. But the law was written in such a way that the FDA had latitude to decide what ‘substantial evidence’ meant. In practice his came to mean large scale double blinded randomised controlled trials. Over time, firms actually brought the agency in to help design their trials, to reduce the risk that their products would fail at the final hurdle.
The FDA also aimed to review all existing drugs on the market – the 3,400 that had only been reviewed for safety, approved between 1938 and 1962. Of these about 1,000 were removed. Other drugs were approved, and still other drugs kept being sold under a sort of brown bag compromise solution, though the agency had never officially signed them off.
In 2006 the FDA started the ‘Unapproved Drugs Initiative’ to work through this backlog, tacking 500 further drugs. This left the situation with a few quirks. Colchinine, the active ingredient in the autumn crocus, used since 1500BC to treat joint swelling and gout, and prescribed about two million times in the USA, is sold in the USA, but manufacturers are at risk of an enforcement action. This means American prices are some thirty times higher than UK prices.
Before the FDA stepped into the breach, some its functions were carried out by the American Medical Association – the professional association of doctors. The AMA had given products it believed safe and effective a ‘Seal of Acceptance’, which it discontinued in 1955 due to worries it was too influential in driving prescribing and purchasing decisions, and thus too decisive commercially for drug manufacturers. They didn’t want that sort of responsibility.
The AMA also discontinued a chemical laboratory they had run to test manufacturers were accurately describing the ingredients of their products, and another scheme it had run to regularise generic names – and even to encourage the use of generic names rather than brand names in prescribing. These responsibilities fell mostly to the FDA.
This is the basic approvals system of today. But there have been a few changes. In general, drug efficacy is about improving clinical outcomes that matter to the patient – whether they feel better or live longer. But patients may live with a condition for quite some while before this outcome comes to pass, potentially meaning extremely long trials. These would be even more expensive than the hundreds of millions typically spent.
To help with this problem, the FDA began approving drugs based on ‘surrogate endpoints’, i.e. things that show up sooner that are themselves associated with survival or other better outcomes. A classic example is approving a cancer drug based on whether a cancer grew or shrunk. Because we can see this much more quickly than cancer survival, and bigger cancers are associated with more deaths from cancer, in principle this should get us new cancer drugs quicker and more cheaply. And pharma firms love it: between 2010 and 2012, the FDA approved almost half of new drugs based on surrogates. We will see later in this series how this can have problems when combined with other features of the system.
The FDA also has four schemes intended to speed up drug approval, to address political, public, and pharmaceutical concerns the system moves too slowly and many people die while they are waiting for drugs to be approved. As we will see later, drugs often spend years between discovery and approval. In general these schemes have sped up approval a bit, and are widely used, but have not dramatically changed the system.
So the US started with almost no drug regime, brought in safety controls, and then brought in efficacy controls, which evolved into a gigantic apparatus over time, responsible for making sure all drugs on the US market are safe and effective.