NYT had an interesting article last weekend about a promising new melanoma drug that was in clinical trials right now – profiling a pair of cousins in the trial, one who got the drug and one who didn’t get it. And couldn’t get it, despite the fact that it seemed to be helping. What an ethical pandora’s box we have here!The core dilemma:
In order to really know that a drug is both safe and effective, we have to do controlled clinical trials with a lot of patients – and to do these trials, we need a control group. Sick patients who aren’t getting the drug.

If we have a drug that shows great promise, it seems unethical to deny it to half of the patients. But at the same time, we can’t prove that it’s really a good drug without doing so, so ultimately, we get a great benefit as a population by making that “sacrifice” of some patients in the trial. It would be true if people were lab rats – but they’re not. Every patient is somebody’s kid, parent, sibling, etc.

So what can we do??

The regulatory pieces:
Besides just being good science/ arguably the best policy in terms of only spending healthcare dollars on treatments that actually work, companies do these trials because that’s what the FDA requires to approve the drug. And without FDA approval, the company can’t market the drug. If they can’t market it, they can’t make any money off it.

There is, however, an exception: a drug that shows great promise can be offered BY THE COMPANY that made it for “compassionate use,” or “treatment use,” that is, use, prior to approval. But this creates a problem: who would want to be in the study, where they might be in the control group and not get the drug, if it were available outside of the study context? Companies generally have little incentive to offer investigational drugs to non-trial patients, because they need approval to market the drug, and if they can’t market the drug, they can’t make any money, which pays for research and development, so. . . nobody would ever make new drugs.

How can we fix this:

Dad: I think a lot depends on the clinical trial design. Many have cross-over provisions if the patients don’t respond (i.e. if they get placebo). They can then get the “active” drug.   This is difficult in the case of cancer in which the time frames are long and the endpoints are generally  survival based.

Nurse: Although survival is certainly not the only benchmark for a drug’s success, especially in cancer. There’s a lot of discussion in evidence-based medicine about whether interventions increase survival, which is great, but it’s only one piece of the puzzle.

Lawyer: How does that work, though? Do they just put a check in the “didn’t work” box after the patient doesn’t respond in a fixed time-frame?

Why shouldn’t we just allow free access or marketing, as long as it’s clear that we don’t know for sure whether this works, once we’ve been through basic safety trials? In other words, if we think it’s safe, should the FDA, either directly or through the incentives they create, stand in the way of more sick patients being able to get the drug?

Nurse: Well, isn’t part of the FDA’s mandate to protect the public from totally bogus nonsense? I mean, think about supplements not regulated by the FDA and how much confusion that generates.

Dad: Safety has not been demonstrated for this drug. The Phase I trial was only 31 patients. Often, it takes thousands or more to identify serious safety concerns. This happened with Remicade [a drug Dad was instrumentally involved in. Or dare we say, invented.]

Nurse: That’s good to know/remember. This article, with its approach of profiling two specific patients, kind of mirrored (and encouraged) the public reaction to promising advances– as in, they have to have it and they have to have it now– when they may be jumping the gun. The article mentioned a promising breast cancer treatment from years ago that didn’t pan out. That happens, and I think that people may be a little misled and also may be unable to judge well if they are really starved for a crumb of hope.

Lawyer: What do they do with treatments for really rare diseases? Where the prospects of finding thousands of patients in a decent time frame isn’t so good?

Nurse: Also, as medical doctors, (even though they are researchers), shouldn’t they have some level of concern for patients as opposed to solely for data? At some point, shouldn’t we place patients’ well-being or relief of suffering above perfect data?

Dad: Then you would never find out if it really worked, and it would cost everyone a lot of money.

Nurse: Well but that’s where a crossover type design could be helpful. I think it’s inhuman to say to a patient, “we’re sorry, even though you are clearly deteriorating rapidly and your tumors are getting bigger, we refuse to try any other treatment because we might sully our results.”  And sure, it’s great if we prove that the drug increases survival, but if we have a fair amount of evidence (even if it’s not perfect evidence) that it has some beneficial effects, isn’t that good too? Why does death have to be the endpoint? That seems like a problematic element in the design.

Nurse: I wonder if there’s a way to do a retrospective study that would give us some insight– look at previous patients who received the chemo as a control. It’s not perfect, but perhaps it’s necessary from an ethical standpoint.

Lawyer: And don’t we already use lots of treatments that are FDA-approved, but really don’t work very well? They’re just the best we’ve got? I guess the question here is, are these better or worse than existing treatments? If we’re trying to improve, we need to know if we are — or not.