Louise (not her real name), a scientist and avid runner in her 50s, first noticed the strange pangs under her right ribcage a few weeks before we met. Her CAT scan revealed a liver so riddled with islands of tumor that the radiologist called them "too numerous to count." None of the standard surgical, radiologic or chemotherapy treatments would help, and her doctor told her that she had two, maybe three, months to live.
Louise came to us for a second opinion. I swallowed hard before going into her exam room to summarize my findings. It was, and still is, not easy for me to tell patients they are dying.
But her reaction surprised me. Instead of a grim silence or even tears, she simply smiled. "I want to become a grandmother," she said, looking over at her pregnant daughter-in-law, who was just beginning to show. Louise knew she was going to die from the cancer; she wanted only to live long enough to see her grandchild born.
For the rest of the afternoon, Louise's plea played over and over in my mind. I wanted to offer something — anything — that might increase her chance of survival by a few months. No standard therapy would work, so I wondered if a clinical trial, a study of a new drug or treatment, might hold some promise.
I also knew that every research trial maintained strict criteria for enrollment and that Louise was hardly the ideal candidate and might not qualify. While there were trials that accepted patients with diminished liver function, her organ was on the precipice of all-out failure. If she took part in a trial and her liver failed, she could muddy the data, perhaps even alter the trial's outcome. The investigators would assume that liver failure was a side effect of the experimental drug, a complication potentially so significant it could prevent future use of the drug. But the reality was that her liver failure would have probably had nothing or very little to do with the drug; it would have been the result of not having had enough normal liver to begin with.
Nonetheless, an experimental drug seemed to me to be my patient's only hope. I called up a medical oncology colleague and was ready to consider violating the entry rules of a clinical trial. "She just wants to see that child born," I said to my colleague. "I wonder if we could finagle things so she could be in a trial?"
It turns out that I am not the only doctor who has considered bending the admissions criteria of a clinical trial for a patient.
There is an essential conflict for doctors involved in clinical research. As collaborators in research, they want to obtain information that is valid and able to be generalized for all future patients; but as doctors working within a patient-doctor relationship, they need to focus only on improving the condition of the patient before them. A patient in a chemotherapy clinical trial, for example, might complain of severe nausea after receiving a dose of an experimental drug. While the doctor as researcher might place a high priority on limiting the types of antinausea medications because of fears they could interact with and affect the experimental drug's effects, that same doctor as patient advocate might choose to ignore the list of "approved" antinausea medications to use one that has previously worked.
Until recently, bioethicists who have studied the tension between these two competing commitments for doctors involved in clinical trials have primarily written about how patients might be particularly vulnerable. Some bioethicists have expressed concern that doctors might tolerate or promote research that could compromise a patient's well-being; others have worried that doctors might not fully inform patients of the potential risks involved.
Now, a recently published survey reveals that it may not be the patients who are most vulnerable, but the trials themselves.
In the current issue of the bioethics journal IRB: Ethics & Human Research, investigators from four different institutions surveyed over 700 clinicians involved in clinical trials and found that 90 percent believed that ignoring certain entry criteria was acceptable if a patient could, in their estimation, benefit from the trial. In addition, over 60 percent of those surveyed also believed that researchers should deviate from study rules if doing so might improve a patient's care.
While bioethicists and researchers have long suspected that doctors and other clinicians might be committing an occasional protocol infraction, few if any studies have looked at the extent to which such violations occur and how they might compromise research results.
"What struck me," said Charles W. Lidz, research professor of psychiatry at the University of Massachusetts Medical School in Worcester and lead author of the study, "was the number of clinicians who described circumstances where they would do things that violated protocol in order to help their patient."
Such violations could include altering medical records in order to get a patient into an H.I.V. treatment trial, downplaying a substance abuse history in order to help a patient enroll in a trial on depression, or artificially improving an otherwise poor kidney function test by having a patient drink a gallon of water the night before the study's blood draw.
While each of these violations was made in the best interest of individual patients, they also had the potential to compromise, and even negate, a trial's findings. "Not everyone had an account of how they would go around a protocol in order to help their patients, but a significant number did," Dr. Lidz said. "And they told their story proudly because it was a reflection of their commitment to patients."
Clinicians are motivated to "bend" protocol rules because they view the experimental therapy as promising. "There's a pervasive idea among clinicians and patients that a new drug or device is going to make things better," Dr. Lidz noted. But statistically, while many experimental treatments are as good as standard therapy, few actually end up being superior, and some are worse.
Nonetheless, Dr. Lidz maintains that participating in a clinical trial remains the right choice for most patients if their current treatment is not working. Patients in experimental protocols tend to receive more attention than those who are not enrolled, and clinical trials are central to determining if a new treatment really works. "If people don't join them," he remarked, "we are all in trouble. Patients just need to realize that they are doing this for a larger cause, not necessarily their own interest. And that larger cause is really important."
Despite the survey results, Dr. Lidz believes it is still possible to support the human connection and sense of obligation between doctors and their current patients while advancing the clinical research that may help future patients. Researchers devising clinical trials will need to make it clearer to doctors how important it is to follow the entry criteria and treatment protocol. And they will need to understand that a poorly designed clinical trial can force clinicians to choose between obligations to the research and their commitment to a patient.
"It's really easy to assume there are 'good guys' and 'bad guys' and that someone is not taking certain aspects of clinical research seriously," Dr. Lidz said. "But, in fact, there is something investigators are not adequately taking into account when designing trials. They need to recognize that the doctor-patient relationship is important to both the doctor and the patient and that one cannot design a protocol in such a way that ignores the feelings of the doctor and patient about how important it is to provide good care."
Louise, my patient with liver cancer, did end up receiving a new drug, but not as part of a clinical trial. Despite my suggestion that we try to squeeze her into a trial, my colleague found a way to give her the new drug off protocol. And whether it was because of that drug or, more likely, her extraordinary will, she survived to become a grandmother. She died just days after her grandson was born.
"In a certain sense," Dr. Lidz mused, "these survey results suggest that the doctor-patient relationship is in better shape than many people have suggested.
"If you are willing to go around the clinical trial design in order to provide better care for your patient, that suggests to me you are pretty committed to your patient. And in a sense, I think that is sort of reassuring."