Let’s Test Everyone: broader clinical trials are harder than you think
Yesterday on Facebook I linked to an article about the new class of anticoagulant medicines such as Pradaxa, and the current controversy around them. One issue is that these drugs are mainly being prescribed for older and/or sicker patients, but were probably tested on primarily young and healthy volunteers. However:
“In the case of Pradaxa trials, we know elderly people and people with renal disease were relatively under represented when compared with those who were prescribed Pradaxa when it was released onto the market. This is important because those excluded are the very people who are most likely to develop bleeding complications.”
My Facebook friend Maryana Simonovich raised the reasonable idea that “FDA should advocate less selective inclusion criteria for those trials, to get a more realistic clinical outcomes.”
That seems like a good idea, but in my experience (on the board of biotech companies as an investor) Maryana is wrong. But the reasons WHY she is wrong are non-obvious, interesting, and worth a much fuller exploration.
[edited to add: Maryana isn’t WRONG, of course. In an ideal world the broadest possible inclusion criteria would be ideal. The points below are just a look at why the potential costs of broader inclusion criteria may outweigh the benefits.]
Why not just test the drugs on the patients most likely to take the drugs?
- Clinical trials are powered to show statistically significant differences between the various arms: one group gets the drug being tested, and the other either gets a placebo or the current standard of care (such as warfarin, in the case of anti-coagulants.) The probable difference in efficacy and adverse events between the two arms is almost always small, so you need a large number of patients to see meaningful differences in the few months that the testing will occur. For something like Pradaxa, a Phase III FDA trial will have 500-2,000 patients enrolled. If the only endpoint I am looking at is “does the drug reduce clotting safely?” then I stand a good chance of having a successful trial. But once I start asking questions like “Does the drug reduce clotting safely for men and women, black and white, young and old, those with kidney troubles, those with history of hemorrhagic stroke, etc?” I need to dramatically increase the size of my trial: 10,000 patients might not be enough. Enrolling 10,000 patients (more on this in a second) will take longer, and cost much more money. And before you talk about drug companies only being interested in profits, remember that they are controlled by shareholders who do worry about the ROI for clinical trials. Depending on the disease, more than half of all Phase III trials fail, so demanding that all new drugs be tested across the full range of possible patients means that many fewer drugs (especially for rarer conditions) will be tested going forward. That’s not good.
- Drug trials have very clear rules. Not only are we looking for drugs to stop clotting, we want to make sure they are safe. Those running the trial will be looking for signs of excessive bleeding, bruising, rashes, headaches, plus the usual stuff all medicines need to worry about (sleepiness, wakefulness, nausea, diarrhea, and so on and so on.) But they will be ESPECIALLY vigilant for serious adverse events: stroke, other fatal bleeding and death. A single death (of uncertain cause) can mean that the entire trial is halted. The investigators need to study that death and figure out if the drug is killing people. It can take weeks or even months, and sometimes the entire trial needs to be restarted. In a trial of 1,000 young healthy volunteers, the odds of a subject dying unexpectedly are very low. But once we add older, sicker patients to our investigation, the laws of probability start working against us. Patient #749 dies after two weeks of the drug. They were in their 70s, and had been on dialysis for ten years now. The doctors are pretty sure the new drug has nothing to do with the death, but they can’t be sure. Once again, the cost of drug development will go much higher, and drugs will take longer to get to market. Not good.
- It’s different for pancreatic cancer. Getting 73 year olds to enroll in new and potentially risky therapies when they are already seriously ill tends not to be a problem. You can recruit quickly, as long as your disease isn’t too rare. But getting older, sicker patients to sign up for a drug trial that is NOT lifesaving is much harder. Not only are the patients worried the drug might be unsafe, but the whole clinical trial process is a big use of time and can be a pain in the ass. Getting young healthy volunteers (who are often under employed and looking for a bit of money) is tough enough. So many trials are delayed months trying to find a thousand people willing to risk their lives for a few hundred bucks. Once you start trying to study the sick and elderly as well, I can guarantee that enrollment will take at least twice as long, perhaps more. Drugs already take many years to come to market. Forcing drug-makers to broaden the inclusion criteria would unquestionably delay many drugs. Don’t get me wrong: we would know they were safer drugs, but the number of people who die waiting for the better drugs could well be larger than the number of lives saved due to safer drugs.
This post isn’t as well hyperlinked as most, but I am on vacation, so figure any blog post is better than none at all!
[As always, I am making no comments or statements about the safety or dangers of Pradaxa, or any other drug. Or its manufacturer. This is just a blog post exploring some of the issues around inclusion criteria for clinical trials, from the perspective of someone who used to work in the field. This has nothing to do with what I do at Deloitte in the tech, media and telecom areas. All opinions are strictly my own and not Deloitte’s.]