of care, to accommodate parents’ personal value judgments about the acceptability of different harms. The argument that we do not inform parents of these trade-offs as part of usual care merely highlights the inadequacy of clinical informed consent.12 In SUPPORT, parents should have been informed about the study’s purpose, the experimental procedures (including the altered pulse oximeter), the reasonably foreseeable risk of ROP in the higher SpO2 group, and the unforeseeable but potential harms of NDI and death in the lower SpO2 group. Disclaimer

The opinions expressed in this commentary are solely those of the author and do not represent the policies of the Food and Drug Administration. 1. J. D. Lantos and C. Feudtner, “SUPPORT and the Ethics of Study Implementation: Lessons for Comparative Effectiveness Research from the Trial of Oxygen Therapy for Premature Babies,” Hastings Center Report 45, no. 1 (2015): 30-40. 2. National Institute of Child Health and Human Development Neonatal Research Network, “The Surfactant Positive Airway Pressure and Pulse Oximetry Trial in Extremely Low Birth Weight Infants: The SUPPORT Trial,” August 28, 2004, last modified March 28, 2005, accessed November 29, 2014, http://www.nih.gov/icd/ od/foia/library/Protocol.pdf; see, especially, p. 9. 3. C. H. Cole et al., “Resolving Our Uncertainty about Oxygen Therapy,” Pediatrics 112, no. 6, part 1 (2003): 1415-19. 4. National Institute of Child Health and Human Development Neonatal Research Network, “The Surfactant Positive Airway Pressure and Pulse Oximetry Trial in Extremely Low Birth Weight Infants.”

Other Voices

The Controversy over SUPPORT Continues and the Hyperbole Increases by ALAN R. FLEISCHMAN

T

wo articles in this issue of the Hastings Center Report address the continuing controversy over the Surfactant, Positive Pressure, and Oxygenation Randomized Trial (SUPPORT).1 This controversy is part of a larger discussion about the appropriate regulatory framework for protecting human research participants in comparative effectiveness research (CER), a group of studies that aims to compare two “usual” or “standard” treatments in order to provide evidence of which treatment is most effective.2 Henry Silverman and Didier Dreyfuss argue that randomizing subjects to restrictive interventions at the outer Alan R. Fleischman, “The Controversy over SUPPORT Continues and Hyperbole Increases,” Hastings Center Report 45, no. 1 (2015): 42-44. DOI:

10.1002/hast.416 42 HASTI N G S C E N T E R R E P ORT

5. I. Chalmers et al., “Data Sharing among Data Monitoring Committees and Responsibilities to Patients and Science,” Trials 14, no. 102 (2013), doi 10.1186/1745-6215-14-102; L. M. Askie et al., “NeOProM: Neonatal Oxygenation Prospective Meta-analysis Collaboration Study Protocol,” BMC Pediatrics 11, no. 6 (2011): doi 10.1186/1471-2431-11-6. 6. L. M. Askie, D. J. Henderson-Smart, and H. Ko, “Restricted versus Liberal Oxygen Exposure for Preventing Morbidity and Mortality in Preterm or Low Birth Weight Infants,” Cochrane Database of Systematic Reviews 1 (2009): CD001077. 7. Cole et al., “Resolving Our Uncertainty about Oxygen Therapy.” 8. H. J. Silverman and D. Dreyfuss, “Were There ‘Additional Foreseeable Risks’ in the SUPPORT Study? Lessons Not Learned from the ARDSnet Clinical Trials,” Hastings Center Report 45, no. 1 (2015): 21-29. 9. B. J. Stenson et al., “Oxygen Saturation and Outcomes in Preterm Infants,” New England Journal of Medicine 368 (2013): 2094-2104. 10. B. Schmidt et al., “Effects of Targeting Higher vs Lower Arterial Oxygen Saturations on Death or Disability in Extremely Preterm Infants: A Randomized Clinical Trial,” Journal of the American Medical Association 309 (2013): 2111-20. 11. M. S. Carome, S. Wolfe, and R. Macklin, “Analysis of the Complete Protocol and Consent Form for the SUPPORT Study: Lack of Informed Consent and a Failure to Ensure That Risks Were Minimized,” May 8, 2013, http://www.citizen.org/documents/2124.pdf; see p. 11 for a discussion of the factorial protocol design. 12. A. M. Capron, “The Real Problem Is Consent for Treatment, Not Consent for Research,” American Journal of Bioethics 13, no. 12 (2013): p. 27-29.

limits of a standard-of-care practice, as was the case in SUPPORT, differs in important ways from treatment available outside of the proposed study and imposes additional reasonably foreseeable risks on the subjects. They conclude that the study not only required review by institutional review boards and an enhanced informed consent process revealing all of the reasonably foreseeable risks but also that local IRBs did not have the authority to approve the study because it contained significant risks that could not be justified by any compensating benefits to the subjects. They believe that the study required review and approval by the U. S. secretary of Health and Human Services. As they explain, federal regulations create four categories of permissible research involving children, and research in the fourth category, in which participants face significant risk without compensating benefit, requires the January-February 2015

approval of the secretary of Health and Human Services.3 The SUPPORT investigators and many commentators have argued that the study was approvable under the first category, in which there is no more than minimal risk to participants. Others have suggested that the risks were more significant but that the possible benefit of decreasing retinopathy of prematurity allowed an IRB to approve the study under the second category, in which there is some risk but a compensating prospect of direct benefit. Silverman and Dreyfuss hold that the study falls into the fourth category because it posed significant additional foreseeable risks to the subjects, as compared to alternative standard therapies outside the trial, without any increase in the anticipated benefits to the subjects. This is just not correct. Virtually every critic of the study, including the Office for Human Research Protections, agreed that the study was an important clinical trial designed to generate knowledge that could be of help in the treatment of premature infants. All agreed that there was clinical equipoise for the trial, since clinicians were uncertain about the appropriate level of oxygen required, and that for at least some of the subjects there was the potential to reduce the incidence of severe retinopathy, as compared to the current standard practice. Many critics argued that the informed consent process needed to have described the correlative risks and benefits of the study, but no one denied that there was a potential benefit to study participants. John Lantos and Chris Feudtner make several very different arguments, beginning with the assertion that the SUPPORT study “is a prime example of CER,” so that analysis of this case can be helpful in setting the regulatory approaches to standard-of-care research. As lawyers often argue, bad cases make bad laws. SUPPORT was not representative of CER and is not a good case upon which to make guidelines about best practices in regulating standard-of-care research. The controversy over the appropriate regulatory review and informed consent process of this study may end up impeding or even preventing important research that examines the efficacy and risk of more routine questions in CER. Although SUPPORT’s comparison of two levels of oxygen has been emphasized, the study actually was far more complicated. Eligible infants were first stratified into younger and older preterm infants, all within the range of twenty-four to twenty-eight weeks of gestation, and then were randomly assigned to receive one of two types of resuscitation, either standard resuscitation with intubation and administration of a surfactant to enhance lung development or constant positive airway pressure without intubation. If they required supplemental oxygen during their stay in the nursery, they were assigned to either the high- or low-oxygen group. The two oxygen groups were at the two extremes of the standard range for oxygen saturation. Each extreme was being recommended by a few clinicians, but most clinicians were choosing the January-February 2015

broader range of 85 to 95 percent. “Prime” examples of CER might compare two different but widely accepted treatments of a condition, where a great deal is already known about the treatment. The research would seek to determine which of the two treatments is most effective, least risky, and least costly. Lantos and Feudtner offer a brief, straightforward, and concise informed consent paragraph that they suggest might have been sufficient for SUPPORT. While their attempt to simplify the consent form is laudable and in general quite good, they are a bit disingenuous when they state, “But we don’t know which group will have better outcomes. (If we knew, we wouldn’t be doing the study.)” In fact, the purpose of the study was to validate the hypothesis that the infants in the low-oxygen group would have a lower incidence of severe retinopathy, implying that the infants in the high-oxygen group might have a higher incidence of that problem. The consent document should have been able to disclose this information to the parents without unduly frightening them or significantly decreasing enrollment in the study. It might have explained, “Virtually all small premature babies will need supplemental oxygen in order to maintain body functions and prevent brain damage. But too much oxygen may cause problems in vision by damaging the retina in the eye.” And later, it might have continued, “While we don’t know which group will have a better outcome, we believe that babies in the lower oxygen group may have less severe vision problems than babies in the higher oxygen group.” Neither of the articles deals adequately with the question of randomization in CER. Many commentators have argued that since subjects in CER will be receiving standard treatments that they might have received in clinical care, being randomly assigned to treatments does not increase risk and therefore does not require informed consent. In most cases, however, randomization at the level of the patient should require informed consent, even when two FDA-approved medicines or two standard health care delivery interventions are being compared. Randomization may or may not increase risk, depending on the level of risk in each of the arms of the study. If there are differences in the types or severity of risk in the arms of the study, then the risk to an individual subject is altered by participation in the trial. Subjects ought to be informed of the differential risk that is a result of being in a study. In addition, randomization requires informed consent whether or not the level of risk is different, based on the principle of respect for persons. Consent should be sought because patients and their surrogates may differ in their preferences and beliefs and respect for their personhood requires that they be allowed to make a fully informed decision about participation. If OHRP develops guidance for IRBs, it must provide examples of CER protocols with varying degrees of risk as well as explanations of how current guidelines can be interpreted to offer the right level of oversight. H AS TI N GS C EN TE R RE P O RT

43

1. H. J. Silverman and D. Dreyfuss, “Were There Additional Foreseeable Risks in the SUPPORT Study?,” Hastings Center Report 44, no. 6 (2014): 21-29; J. Lantos and C. Feudtner, “SUPPORT and the Ethics of Study Implementation: Lessons for Comparative Effectiveness Research for the Trial of Oxygen Therapy for Premature Babies,” Hastings Center Report 44, no. 6 (2014): 30-40.

Other Voices

2. A. R. Fleischman and M. Z. Solomon, “Comparative Effectiveness Research: Ethical and Regulatory Guidance,” JAMA Pediatrics 168, no. 12 (2014): 1089-1090. 3. Department of Health and Human Services, “Final Regulations concerning Protection of Human Subjects,” subpart D, “Children,” 45CFR46.401-409.

SUPPORT and Comparative Effectiveness Trials: What’s at Stake? by lois shepherd

D

efenders and critics of the Surfactant, Positive Pressure, and Oxygenation Randomized Trial tend to see the relationship between physicians and research subjects in fundamentally different ways. John Lantos and Chris Feudtner’s article in this issue points out that the critics tend to include more lawyers or Ph.D. bioethicists among their ranks than the group defending SUPPORT and its consent forms.1 They are not the first to draw attention to differences between the views of clinicians and nonclinicians on SUPPORT. Jon Tyson, one of the SUPPORT investigators, does so in a recent article in Pediatrics, writing that “[m]any nonclinical ethicists have sided with OHRP” (the Office for Human Research Protections) while defenders of SUPPORT “include clinical ethicists, clinical investigators, clinicians, the editor of the New England Journal of Medicine, and the director of the National Institutes of Health.”2 And it is certainly true that if one compares the signatories of the “dueling bioethicists” letters published in the summer of 2013 in the NEJM, more M.D.s signed the first letter, which defended the SUPPORT consent forms, and more J.D.s and Ph.Ds signed the counter-letter.3 Unlike Lantos, Feudtner, and Tyson, I do not think this disagreement stems from a difference in understanding over facts. Nor does this disagreement signify that those associated with the more “legalistic” camp believe that clinical researchers do not care enough about their patients. When critics of SUPPORT explain why the consent forms were inadequate, they are not, as Lantos has elsewhere suggested, accusing the investigators of being immoral or uncaring people or mad scientists.4 They are pointing to the requirement of an informed consent process that attempts to address, in part, the long-recognized conflict of interest experienced by physicians combining research and therapeutic aims in the treatment of their patients.

Lois Shepherd, “SUPPORT and Comparative Effectiveness Trials: What’s at Stake?,” Hastings Center Report 45, no. 1 (2015): 44-45. DOI: 10.1002/

Lawyers see conflicts of interest as structural problems requiring structural solutions rather than as ethical lapses of morally deficient actors.5 When an investigator undertakes clinical research, his or her legal and ethical duties to patients change. Because a physician cannot be expected in clinical research to pursue only the best interests of the patient (although he or she can be expected to protect patient safety and meet other obligations), research protections are essential. These protections rely on people and institutions other than the physician and can be achieved only if it is acknowledged that what is taking place is in fact research. This understanding of the physician-researcher conflict of interest is hardly radical. In fact, many of the research ethics norms developed over the last fifty-plus years have been built upon it. The question that has arisen recently is whether comparative effectiveness research, which “compares outcomes between patients who receive two different treatments that are both in widespread use,”6 calls for a new paradigm. Is CER research or clinical care? Should it be regulated as research or left largely to individual physician discretion? Lantos and Feudtner argue that SUPPORT was an instance of CER and that CER differs from research involving unproven, experimental therapies because it exposes research subjects to the same risks patients regularly face in clinical practice. Like many defenders of SUPPORT, they formally acknowledge the study as research but want it to be thought of as clinical care. There is debate about whether SUPPORT qualifies as CER. But I agree that how we view SUPPORT can guide us in how we should think about comparative effectiveness trials. Even accepting all the defenders’ assertions that make it appear to be CER, the true nature of SUPPORT as research—accompanied by substantial risks of harm and requiring all of the usual protections for research subjects—is clear. When we set aside disagreements over facts, we see that the gulf between defenders and critics is a disagreement over the fundamental questions of research

hast.417

44 HASTI N G S C E N T E R R E P ORT

January-February 2015

The Controversy over SUPPORT Continues and the Hyperbole Increases.

The Controversy over SUPPORT Continues and the Hyperbole Increases. - PDF Download Free
218KB Sizes 36 Downloads 3 Views