My first bullet was basically, clinical trials are really expensive. I should add that they take a long time and there are limited resources for doing them. Ergo, there's a selection process that needs to precede a large scale, high quality trial that might convince the FDA to approve a new drug or procedure. I'll just talk about medications for now to keep it simple.
First, obviously, you need some reason to think a compound (or a more complicated biological entity such as a monoclonal antibody) will be of use for some clinical indication. Drug companies screen all sorts of candidates, often starting in vitro to see if they have an effect on cells in culture that might translate into a useful effect in a whole organism. The next step would be to try the compounds in animals. For reasons of expense and convenience, and also because people have strong reservations against experimenting on species that are closely related to us, these are usually rodents.
Your first problem is to get mice that have a condition or disease that resembles a disease of humans. If you're testing antibiotics that isn't very complicated, you can just infect them. (In case any PETA members are reading this, yes, some people think this is unethical.) Otherwise, however, diseases in mice can't really be exactly the same as diseases in people. Gene editing has now become very precise so it's possible to create mice with a mutation which is closely analogous to a human genetic defect, but since our genomes are different from those of mice to begin with this won't be a perfect analog. You can induce cancer in lab animals by various means, but again you'll never have a cancer that's identical in either it's genetic alterations or its environment to a particular cancer in humans. And so on.
Sadly, therefore, it is most often the case that treatments that appear promising in laboratory animals don't end up being useful in humans. So you've already spent some money but now you need to spend a whole lot more to see if this can work in people. The first step is called a Phase One trial. This will start with a very small number of volunteers -- maybe just one. They'll give the initial group a very small dose, to make sure it's safe; to start to understand its "pharmacokinetics," that is how it is metabolized and how long it remains in the bloodstream, or perhaps where it goes in the body; and if it has any of the desired biological effect, however measured. This doesn't mean effectively treating or curing the disease, just whether it seems to be doing whatever it is you want it to be doing in the immediate term.
Then the investigators will try gradually increasing the dose, so they can find the window in which there aren't any immediately obvious serious adverse effects but the desired biological effects are occurring. At this point it may become apparent that the trial can't proceed, because there's no therapeutic window or the stuff just doesn't work in people after all. By the way the volunteers don't necessarily even need to have the disease for this phase, though they often do because that's who's likely to volunteer. Anyway, if it still looks promising, at this point you can go to a Phase II trial.
This is a trial with more subjects, but still not enough to be confident you'll get statistically reliable results. Here you definitely do want people who have the target condition, because you want to get a preliminary idea of what the range of responses is, and of course you're also looking at more people, perhaps for a somewhat longer time, to try to catch any adverse effects. However, your length of follow-up is going to be fairly short, so unless what you have on your hands is a miracle cure the outcomes you're looking for may well be so-called "surrogate endpoints" -- not, say, longer survival or reduced risk of a heart attack or long term symptom relief. Rather, you're looking for biological markers which you believe are associated with the desired long-term outcome.
You've already spent a lot of money, but you may or may not be able to go on to Phase III. If you do get the green light and the funding, you will need to recruit hundreds of people who meet eligibility criteria. This means having the disease, likely within a certain range of severity and/or duration; maybe excluding people with certain comorbidities, children, women of childbearing age, people who can't freely give consent. (They used to experiment on prisoners and people in mental institutions, but that's severely restricted now.) That's expensive. It's probably going to require enlisting clinical practices around the country and paying them to recruit subjects and perhaps actually conduct the trial at their site.
Even at this stage, because of the expense, follow-up may be too short to really establish that the treatment is effective -- six months is typical -- so we may still be looking at surrogate endpoints. I'll get into all the other problems and limitations later, but for now we'll just stick with the fact that this is very expensive and it takes a long time. Next, I'll discuss how that very basic problem shapes and misshapes the research enterprise.
No comments:
Post a Comment