Randomization: Quickest route to market, or a form of cruelty?


Institute’s interdisciplinary faculty point to more flexible protocols and revisiting FDA standards for evidence


In her ongoing series “Target Cancer,” New York Times journalist Amy Harmon humanizes the ethical dilemma inherent in clinical trials where a drug which has shown to be highly effective—and pose relatively mild side effects—is withheld from some patients because the study’s design requires a “control” group.


As the Sept. 18 article stated, per the dictates of a randomized control trial, half of the patients in the final phase of the study were put on a pre-market drug for melanoma to further test its effectiveness. The other half, comprising the control group, was assigned to standard therapy.


That chemotherapy is “notoriously ineffective,” Harmon noted early on, adding that melanoma is the deadliest form of skin cancer. But for the sake of preserving the most straightforward and efficient study model—the quickest route for the new drug to get to market—“PLX4032” would not be given to those randomly assigned by computer to the control group.


However, the time-honored randomization process seemed to morph from being a gold-standard practice into a cause for moral outrage in Harmon’s highly personal account of two close cousins who were both diagnosed with the specific kind of melanoma that PLX4032 was developed to treat.


Both men eventually ended up in the trial, but protocol put them on opposite sides of the fence. Photographs that accompanied the article showed the one who got the drug sitting in the yard with his family. The one assigned to the control group was pictured sitting under drab hospital lighting, with tubes inserted throughout his body.


Faculty at the Johns Hopkins Berman Institute of Bioethics discussed the article at length during a recent monthly meeting. As usual, faculty members are assigned a reading with a rich bioethical dimension prior to the gathering, with ample time set aside to share their thoughts openly and candidly.



Core faculty member Steven Goodman, a professor of oncology in the Division of Biostatistics at the Johns Hopkins Kimmel Cancer Center, moderated the Oct. 25 discussion. An authority on clinical trial designs, he began by introducing the room to a term he uses with students to describe trial protocols that raise significant questions about equipoise: “ethically rough.”


“It was the roughest,” Goodman said of the PLX4032 trial. But he also said that he was “very puzzled” as to why the study was designed the way it was.


Alternative approaches


Goodman pointed out that the drug-makers conducting the trial (Roche and Plexxicon) might have considered proposing a protocol that falls under the category of “adaptive design.”


For instance, the group “on drug” and the control arm would both start with equal numbers of patients. But as one group begins seeing significant benefits, the division veers from a 50-50 split—thereby allowing more patients to receive the treatment. Adaptive designs can also shift the randomization scheme, while simultaneously including the flexibility of stopping the trial early if those conducting it are convinced by results.


“There’s a whole, very rich class” of adaptive designs, which Goodman also described as conventional. They can reduce the ethical roughness, but not eliminate it, Goodman said. These types of trials will still have a “gap,” though: some lapse in time during which a potentially promising drug is withheld from some patients because they started off in the control group.


This gets back to a question that Goodman raised at the beginning of the discussion: What obligation do doctors have, collectively or individually, to give every patient what might be the best treatment available? When Plexxicon announced findings of the phase II trials for PLX4032 back in January, the company reported that the drug shrunk tumors in nearly all patients, with 70 percent of them experiencing at least a 30 percent decrease in tumor size.


That announcement came at the start of phase III of the trial, which has enrolled about 700 patients who were randomized one-to-one and are being monitored to determine the drug’s safety and effectiveness. “The primary endpoint of this trial is overall survival,” Plexxicon’s press release flatly states. In other words, success depends on how much longer patients on the drug live than those in the control group receiving standard treatment (dacarbazine).


Ethical discussion


However, Goodman and others members of the Berman Institute faculty quickly acknowledged that success can also be measured by how much a drug enhances the quality of a patient’s remaining days, even if the treatment doesn’t prolong survival. And that’s what made the article, as emotionally engaging as it was, also potentially explosive.


Here is a drug that’s not very toxic, and while it won’t extend a patient’s life, doctors already have data that PLX4032 will likely ensure a relatively decent quality of life for another six months or so. But the drug is withheld in the name of the “greater good”—clear results that will confirm for regulators that the drug should be approved and made available to all Americans.


From across the conference room, another member of the institute’s faculty made the point that FDA standards of evidence and trial design ought to be revisited given the new class of non-chemotherapy cancer treatments that pose lower risks of serious harm. Her mother had been in a trial similar to the one in the Times article, so the discussion hit home.


“She was not randomized, but it did stop the cancer for a short period of time,” she said, then paused before finishing her comment. “What that did for her was huge.”


While the phase III trial would put off the possibility of FDA approval by two years, the drug-makers went ahead with a control trial to secure the broadest approval possible and beat competitors to market, according to the article. Meanwhile, outrage grew among oncologists whose minds were made up about PLX4032’s benefits—and who essentially said that further testing would just cause needless suffering.


However, another member of the Berman Institute faculty raised the question of whether anyone even has the authority to tell pharmaceutical companies—which invest all the money to develop and test new drugs—to respond to the demands of patients and anguish of their physicians.


“I’m not sure when one brings it to the level of corporate interests, and their legal obligations and moral obligations,” he said, “and whether some of these other questions just become moot.”


Goodman brought up another frustration: In his experience with developing adaptive design trials for the Food and Drug Administration, the negotiation process can drag on for months longer than for a proposal for a straightforward, 50-50 randomized control trial—delaying the approval of a treatment that people desperately want.


“I really think the central issue is what is our obligation as caregivers?” Goodman said in conclusion. “That is core to any conception of ‘equipoise,’ a word that didn’t come up here.”

1 person likes this post.


Michael Pena
Steven Goodman

Tags: , , , , , , ,

Leave a Reply