Citations, please. Last time I asked for a citation of a study on safety and effectiveness of stem cells, you linked a basic-science study, which does not prove anything about safety or effectiveness. I’d like to see a human clinical-outcomes study that backs up the claims you are making.
I’m not saying they don’t exist, but I am saying that thus far you have not shown us any.
Let’s talk about this a little more.
Every treatment has side effects. These are biologically active compounds, after all. The purpose of research is to quantify just how big that “deluge” of side effects can be, and make a value judgement on whether the risk profile of the treatment is worse than the disease that the drug is designed to treat, and if the drug’s impact on the disease is favorable enough to justify the risk.
I happen to agree that in selected areas we probably over-treat and end up doing more harm than good; the research process is a constant tug-of-war between getting effective therapies to market while screening out ineffective and/or harmful therapies. Every study is a complex balance of a number of factors - the number of patients feasible to recruit and enroll, the prevalence of the outcome(s) of interest, and sometimes a study can be “wrong” by random chance. Therefore, sometimes a drug will be green-lighted and, in light of post-market research, later be pulled because it either wasn’t as effective as the original research appeared or the side effects were more severe than initially anticipated. The “research” capacity isn’t infinite - at some point we have to look at the existing data and “approve” the drug - we can study a drug through a rigorous process of phase 0 (pharmacokinetics), 1 (dose-escalating), 2 (safety), 3 (safety/efficacy), and 4 (post-marketing). If a drug makes it through phases 0-3, it can be approved while acknowledging that it will still be monitored via post-market research to remain vigilant for potential safety issues.
As a statistician, one of my primary jobs is helping to design these studies. Since we cannot recruit an “infinite” number of patients, we have to make a judgement on the anticipated benefit to the patient, the anticipated rate of events or outcomes, and many other factors to guesstimate how many patients we’ll recruit for each phase.
Here’s an example. I’m proposing a new drug that’s a one-time injection before undergoing cardiac surgery. I believe that my magic drug will work through the Blood Gnomes Pathway and reduce the risk of bleeding.
We will start with a Phase 0 study (usually 10 or fewer patients, sometimes as few as 3) which confirms that the agent is biologically active (i.e. that the Blood Gnomes Pathway actually changes in a human when you inject my new drug).
Next, we will do a Phase 1 study (usually in the range of 15-20 patients, although sometimes as many as 30) for dose-escalation to expose patients to gradually larger doses of the drug and find the MTD, or maximally tolerated dose. At this stage we still aren’t really assessing effectiveness for the desired clinical purpose, only safety and determining what dose range(s) are appropriate to test in a larger population.
Next, we will do a Phase 2 study (50 - 200 ish patients, depending on the specifics) which gives a slightly longer-term look at the safety and effectiveness. This is the first time clinical-outcomes data are considered, although we don’t usually have a large enough sample-size at this stage to make an accurate assessment of the risk reduction from the new therapy. But we usually get enough from this stage to know “this drug has no chance” or “this drug looks promising, let’s proceed.”
Next, we will do a Phase 3 study (or in some cases several concurrent Phase 3 studies, in different patient populations and different conditions) with hundreds or thousands of patients to get a longer-term picture of safety and efficacy. Here’s where the drug is really on the line, and the company will have to spend millions and millions of dollars at this stage because that’s where we collect sufficient data to get an accurate picture of whether the drug is effective at impacting clinical outcomes instead of just jiggering around with the cells in our blood.
Here’s the catch: at each level, there is a small risk of a “false positive” - passing on a drug that does not actually work, but by random chance happened to produce results that made it LOOK like it worked. This risk is controlled by our sample-size calculations to be very small at each step, but it cannot be reduced to zero unless we use extremely large studies that are untenable for several reasons (ethics of assigning that many patients to a relatively un-studied compound, sheer volume of recruitment, etc). We design the study to strike a balance between the risk of “false positive” (passing on a drug that does not work) and “false negative” (killing a drug that DOES actually work) while also minimizing the number of patients that will be exposed to the therapy (since it is not yet known to be effective or safe).
So, what all of this adds up to is that the FDA requires a very rigorous process, but it cannot be made 100% bulletproof. A small percentage of the time, bad drugs that don’t work or cause harm will get through. Most of the time, drugs that do actually work will eventually get approved, and while this process takes a long time, the “research rigor” required is not the problem. Paperwork and bureaucracy is the problem, but even that’s explainable: the paperwork is required because, without it, people will engage in all sorts of fuckery with their data, so extensive documentation of everything is a necessity to minimize misconduct.
This has reached mind-numbingly stupid levels.
