A study called SUPPORT, short for Surfactant, Positive Pressure, and Oxygenation Randomized Trial, was a major reason behind recent draft guidance published by the Office for Human Research Protections (OHRP).
The study enrolled premature infants and was widely criticized as lacking fully informed consent, the OHRP director says.
“It’s not about what did or did not happen in SUPPORT,” said Jerry Menikoff, MD, JD, OHRP director, at an Oct. 29, 2014, meeting of the Secretary’s Advisory Committee on Human Research Protections (SACHRP).
“This [guidance] is giving you examples of what generated this debate,” he says.
The SUPPORT study, conducted between 2005 and 2009 at 22 sites, involved randomly assigning about 1,300 premature infants to higher or lower levels of oxygen saturation. Both high and low levels were within standards of care.
OHRP investigated the study after receiving complaints and concluded that investigators had violated informed consent provisions by failing to inform the infants’ parents of serious risks resulting from both high and low levels of oxygen.
The American Academy of Pediatrics had recommended oxygen saturation levels of 85% to 95%, and both SUPPORT treatment groups were within that range — although at the high and low ends of the range. Infants in both treatment groups had lower rates of death than did infants who were not enrolled, and lower rate of death than historical controls.1
SUPPORT researchers had informed participating families of the risks of blindness, but drafted the consent forms according to the best available evidence that did not indicate an increased risk of death resulting from assignment to either treatment group.1
The study’s results changed this perception: The final mortality rate was 19.9% in the low-saturation arm and 16.2% in the high-saturation arm.2
SUPPORT critics point out that such a difference in mortality should have triggered a data and safety monitoring committee to stop the study and assign all of the infants to a higher oxygen saturation level.2
Another problem with the study was that it did not allow clinicians to offer the usual titrated care in which they adjust the oxygen according to how the neonate is responding, says John Noble, PhD, MA, MSW, of Georgetown, TX, a retired researcher and long-time member of IRB Forum.
The study used equipment that purposely gave inaccurate oxygen level readings to prevent clinicians from making changes, Noble explains.
“That’s hardly standard of care, and if I sign up for any research, I think I’m owed information about how I’m being held in the treatment regardless of how I am responding,” Noble adds.
“Once you stop doctors from titrating oxygen, it’s no longer standard of care,” says Brian A. Gladue, PhD, CIP, executive director for the office of research compliance at the University of North Texas Health Science Center in Fort Worth.
The more problematic aspect of SUPPORT was the inaccurate oxygen readings, says Michael Gusmano, PhD, research scholar at The Hastings Center in Garrison, NY.
“Even critics of OHRP’s proposed regulations would agree that this was problematic,” Gusmano says.
“There occasionally is a tendency for group-think in multisite protocols where, ‘All these other sites were OK with it, so we shouldn’t have any concerns either,’” Gladue adds.
Even SUPPORT’s use of a soothing acronym could be considered deception: “They mislead mothers,” Noble says.
Another debate over SUPPORT and the study’s informed consent form is whether study participation involved minimal risk, Menikoff said at the meeting.
When a researcher tells a potential participant that participation in a study involves minimal risk, the researcher is sending a message that means the following, Menikoff said:
“Nothing important to me will turn on whether or not I participate in this study.”
“There’s no need for me to think hard about whether or not to be in this study.”
None of those conclusions would be true of the type of studies OHRP addresses in its draft guidance, Menikoff says.
“We shouldn’t be providing information to prospective subjects that would lead them to these incorrect conclusions and thus cause them to do things they might not really want to do,” he added.
- Waldemar CA, Bell EF, Walsh MC. Correspondence: Oxygen-saturation targets in extremely preterm infants. NEJM. 2013;368:1949-1950.
- Merz JF, Yerramilli D. SUPPORT asked the wrong question. Am J Bioethics. 2013;13(12):25-26.
