The Editorial Board of the New York Times recently condemned as unethical a study that randomized premature infants to higher levels of oxygen or slightly lower levels of oxygen. The study found that the group that received the lower levels of oxygen were at greater risk of dying than the group that received higher levels of oxygen. The ethical breech, as cited initially by the federal Office of Human Research Protections (OHRP) was “violation of the regulatory requirements for informed consent, stemming from the failure to describe the reasonably foreseeable risks of blindness, neurological damage and death.” Specifically, the consent forms used in this huge (>1300 patients enrolled) multi-institutional trial did not specifically state as a risk a higher incidence of death. The editorial calls the study “deplorable” and states “if the institutions do not offer strong reforms, the agency [Department of Health and Human Services] can suspend their ability to conduct federally financed research on human subjects.”
The story of the emergence of Bioethics as a field is shaped significantly by responses to violations of respect for human dignity in the course of conducting clinical trials. Informed consent became paramount in light of the atrocities performed by Nazi doctors, in the studies highlighted by Henry K. Beecher’s article in the New England Journal of Medicine, and in the Tuskegee study. However, when I read this editorial the venom and vitriol of the authors did not resonate with me. I think that it is probably because of the concept of clinical equipoise, and a mistaken concept that the status quo of medical practice is actually backed by statistically significant evidence.
The SUPPORT group (Surfactant, Positive Pressure, Oxygenation, Randomized Trial) was trying to answer a very important question: what level of oxygenation is ideal to avoid the Scylla and Charybdis of retinopathy of prematurity and brain damage (too much oxygen can cause blindness, too little can damage the brain). When the trial was started it was an unsettled matter as to what the best level of oxygen was. There were divided opinions and conflicting data. In fact, Institutional Review Boards take this sort of thing into consideration all the time when they are approving studies. And in the SUPPORT group the IRBs of each institution deemed it ethically permissible to randomize patients to each arm of the study–because the data to support either treatment was equivocal. Also, since the evidence for one treatment versus the other was not clear, the idea that the “standard of care” was superior to being enrolled into the trial is fallacious.
It is worth reading noted ethicist and pediatrician John Lantos response to the strong allegations that this study was unethical. He articulates this far better and at greater length than I can. He reiterates that the study was designed appropriately, that the consent forms used were appropriate, and he cites data that show that patients who were enrolled in the trial had higher survival rates than premature babies who were not enrolled in the trial. So, all the rancor about improper consent is somewhat diffused when you realize that patients not enrolled in the trial who received the “standard of care” actually were at the highest risk of death. Parents did not sign a consent form to not enroll them in the trial.
Why is this important? A misunderstanding of the actual conditions surrounding the study and a highly visible condemnation makes it even tougher for clinician-researchers to reach significant conclusions–and that is the more regrettable situation. While it is absolutely essential that human subjects are only enrolled in trials following all the ethical safeguards that were developed in the latter half of the 20th century, it looks like in this case the New York Times is fighting a straw man.