The “Research Misconception” and the SUPPORT Trial: Toward Evidence-Based Consens

THE THERAPEUTIC MISCONCEPTION AND SUPPORT

Critics of SUPPORT, including Macklin and Shepherd in this issue (2013), argue that its consent forms encouraged the “therapeutic misconception.” They allege that SUPPORT investigators blurred the distinction between research and clinical care by referring to different treatment options as “standard of care.” They contend that information statements ignored or downplayed the risks of participating in the research. In this commentary we argue that these criticisms demonstrate a different error—the “research misconception” (DeMarco and Markman 2004DeMarco, J. P. and Markman, M. 2004. The research misconception. International Journal of Applied Philosophy, 18(2): 241–252.[Crossref][PubMed][Google Scholar]). Macklin and Shepherd exaggerate the risk of protocol-based research, overestimate the benefit of clinical decision making by physicians for individual patients, and make an artificial and impractical distinction between research and clinical care. Contrary to widely held assumptions, empirical evidence shows that protocol-based clinical research does not increase risk and may reduce it, while physician judgment in the absence of protocols can increase it. The best medicine is likely to be achieved if physician judgment is used within the framework of an agreed guideline or protocol. Lastly, we agree that greater transparency will be needed to maintain trust in clinical research and call for more evidence to develop consensus on how to improve the consent process and patients’ understanding of research they are invited to join. 1This commentary arose from a workshop on “Consent in Perinatal Trials” in Sydney on September 5, 2013. It reflects our perspective as present/past clinicians (DW/WTM), researchers (DW/WTM/NG), ethics committee members (DW/NG), and parents of preterm infants (MC). It raises issues from the workshop, but does not necessarily represent all participants.D. Wilkinson was supported for this work by an early career fellowship from the Australian National Health and Medical Research Council (1016641). W. Tarnow-Mordi was a principal investigator on the Australian BOOST II trial of oxygen saturation monitoring.View all notes

THE RESEARCH MISCONCEPTION

There have been a number of specific criticisms of the SUPPORT trial. One relates to the distinction between research and therapy. In their target article Macklin and Shepherd (2013Macklin, R. and Shepherd, L.2013. Informed consent and standard of care: What must be disclosed. American Journal of Bioethics, 13(12): 9–13.[Taylor & Francis Online][Web of Science ®][Google Scholar]) write:

Calling a study a “standard of care study” makes it easy to forget that what is being discussed is a research experiment. The aims of research are different from the aims of clinical care. The role of the investigator is different from the role of the clinician. Research is aimed at advancing science to benefit people in the future. Clinical care is aimed at advancing the best interests of current patients. (12)

 

In a submission to the U.S. Department of Health and Human Services, Alice Dreger and Susan Reverby (2013Dreger, A. and Reverby, S. M.2013. Testimony to HHS. Available at:http://www.alicedreger.com/support_files/DregerReverbyHHS2013.pdf(accessed September 23, 2013) [Google Scholar]) wrote, “The SUPPORT study consent documents seem at times to have actively promoted the therapeutic misconception.”

The concept of the therapeutic misconception has been used in different ways (Kimmelman 2007Kimmelman, J. 2007. The therapeutic misconception at 25: Treatment, research, and confusion. Hastings Center Report, 37(6): 36–42.[Crossref][PubMed][Web of Science ®][Google Scholar]). It was originally used to describe research participants’ failure to understand randomization, and a mistaken belief that they would receive whichever arm of the study was best suited to them. This was seen to represent a serious failure of the informed consent process. Of more relevance to the SUPPORT controversy, the therapeutic misconception has also been identified where participants overestimate the benefits of participating in research, and underestimate the harms. More broadly, it has been applied to a failure to distinguish between research and clinical care. Macklin and Shepherd identify the separation between the aims of research and the aims of clinical care as “well-known and widely acknowledged truths.” Yet this separation is by no means clear-cut. DeMarco and Markman (2004DeMarco, J. P. and Markman, M. 2004. The research misconception. International Journal of Applied Philosophy, 18(2): 241–252.[Crossref][PubMed][Google Scholar]) have characterized denial of the therapeutic value of clinical trials as a “research misconception.” This concept includes three possibilities (of which we will only discuss the second and third):

1. A mistaken belief that treatment in a clinical trial is determined only by protocols, without accounting for individual clinical need.

2. Overestimating the risks of research, and undervaluing its benefits.

3. Failure to appreciate that research and clinical care overlap.

In some situations the therapeutic misconception does represent a genuine concern. Phase 1 clinical trials of a new chemical moiety in healthy human volunteers, for example, are a major departure from normal treatment, with no prospect of individual clinical benefit. However, Phase III randomized controlled trials (RCTs) compare protocols derived from previous research with standard care. For such studies (as well as “standard care” trials like SUPPORT) the research misconception may be a more pressing issue.

RISK/BENEFIT OF INVOLVEMENT IN CLINICAL TRIALS

Among 860 Phase III RCTs in over 350,000 people, new treatments were better than standard care in just over half (Djulbegovic et al. 2013Djulbegovic, B.Kumar, A.2013. Medical research: Trial unpredictability yields predictable therapy gains. Nature, 500(7463): 395–396.[Crossref][PubMed][Google Scholar]). Further, among 86,640 patients treated in RCTs and 57,205 patients receiving the same treatment outside RCTs, there was no significant difference in outcomes (Vist et al. 2008Vist, G. E.Bryant, D. 2008. Outcomes of patients who participate in randomized controlled trials compared to similar patients receiving similar interventions who do not participate. Cochrane Database of Systematic Reviews, 3: MR000009[PubMed][Google Scholar]). These fundamentally important findings demonstrate that, in practice, the results of individual RCTs are unpredictable and it cannot be assumed that newer treatments entail greater (or less) risk.

One documented benefit of enrolment in clinical trials is that protocol-based care and intensive follow-up may themselves improve outcome for participants (McCarney et al. 2007McCarney, R.Warner, J. 2007. The Hawthorne effect: A randomised, controlled trial. BMC Medical Research Methodology, 7: 30[Crossref][PubMed][Web of Science ®][Google Scholar]). In newborn intensive care, this has been most clearly shown in the area of feeding, where a systematic review found that standardized feeding protocols reduced the risk of necrotizing enterocolitis by 87% (Patole and de Klerk 2005Patole, S. K. and de Klerk, N.2005. Impact of standardised feeding regimens on incidence of neonatal necrotising enterocolitis: A systematic review and meta-analysis of observational studies. Archives of Disease in Childhood Fetal and Neonatal Edition, 90(2): F147–F151.[Crossref][PubMed][Google Scholar]). A systematic review of RCTs in adults receiving intensive care found that strict protocols reduced duration of mechanical ventilation, weaning, and stay in the intensive care unit compared with individualized physician care (Blackwood et al. 2011Blackwood, B.Alderdice, F.2011. Use of weaning protocols for reducing duration of mechanical ventilation in critically ill adult patients: Cochrane systematic review and meta-analysis. British Medical Journal, 342: c7237[Crossref][PubMed][Web of Science ®][Google Scholar]).

With specific reference to the SUPPORT trial, Macklin and Shepherd express concern that parents of infants were not informed that being in the trial might increase their baby's risk. Other critics were concerned that participants in the trial were denied the benefits of individualized titration and adjustment of therapy that occurs in regular clinical care. However, there is no evidence that infants enrolled in SUPPORT fared worse than those receiving standard care. Participants in the trial actually had lower rates of death and serious complications than infants who were eligible but were not enrolled in the trial, albeit at least partly because fewer sicker infants were enrolled (Rich et al. 2012Rich, W.Finer, N. N. 2012. Enrollment of extremely low birth weight infants in a clinical research study may not be representative. Pediatrics, 129(3): 480–484.[Crossref][PubMed][Web of Science ®][Google Scholar]).

Second, although it might appear better at first glance for patients to have their treatment determined by physician judgment rather than chance, this is not necessarily the case. Oxygen therapy in newborn infants provides a salutary example. In the absence of clear evidence, individualized titration of oxygen, based on physician judgment, led to blindness, death, and serious disability in thousands of premature infants in the 1950s and 1960s (Silverman 2004Silverman, W. A. 2004. A cautionary tale about supplemental oxygen: the albatross of neonatal medicine. Pediatrics, 113(2): 394–396.[Crossref][PubMed][Web of Science ®][Google Scholar]).

THE OVERLAPPING NATURE OF RESEARCH AND TREATMENT

Macklin and Shepherd differentiate the aims of research from clinical care and the roles of clinicians from researchers. But these distinctions are artificial. Some elements of a research study like SUPPORT are performed solely to benefit future patients. Yet many parts of the research protocol benefit the current patient (DeMarco and Markman 2004DeMarco, J. P. and Markman, M. 2004. The research misconception. International Journal of Applied Philosophy, 18(2): 241–252.[Crossref][PubMed][Google Scholar]). These two purposes can, and often do, coexist. In SUPPORT, while the specific oxygen saturation target was randomly allocated, the aim of the protocols for administering and titrating oxygen was to achieve the best outcome for the infant. Moreover, the roles of clinician and researcher cannot be so easily distinguished. Almost all the investigators in SUPPORT and in other standard of care trials were, and are, clinicians working in intensive care. During the trial they cared for infants in the study, as well as conducting the research study. This is inevitable in clinical research, and desirable. In our experience, parents often express a preference that initial descriptions and explanations of research come from the clinicians caring for their child.

TRANSPARENCY AND THE NEED FOR RESEARCH INTO THE CONSENT PROCESS

The SUPPORT investigators were not guilty of blurring the distinction between therapy and research. That distinction is inevitably blurred in standard of care trials. They did not underestimate the risks of the research. The known risks were intrinsic to extremely preterm birth and standard treatment. We agree with David Wendler (2013Wendler, D. 2013. What should be disclosed to research participants?. American Journal of Bioethics, 13(12): 3–8.[Taylor & Francis Online][Web of Science ®][Google Scholar]) that information statements must explain any additional risks of involvement in a trial, but not risks that occur regardless of involvement.

Nevertheless, there are lessons to learn from SUPPORT. The consent process (particularly where there is only a short period to decide) can be confusing and stressful for parents. Key outcomes that will be measured in a trial should be disclosed to parents, even if no change in outcome is expected. It should be clear how involvement in the trial would change care.

How can we improve parents’ understanding? There is no evidence that longer, more detailed written statements will help (Freer et al. 2009). Rather, we should improve written and verbal communication. Concise, tailored descriptions may improve comprehension. Empirical research can help elucidate how best to communicate risks to parents of critically ill infants (Tarnow-Mordi et al. 2013Tarnow-Mordi, W.Keech, A.2013. Towards evidence-based consensus: Comment on Macklin R, Shepherd L, Dreger A, et al. The OHRP and SUPPORT—Another view. New England Journal of Medicine, 369: e3 Available at:http://www.nejm.org/doi/full/10.1056/NEJMc1308015-t=comments[Crossref][PubMed][Web of Science ®][Google Scholar]). Parents should be full partners in such research at every stage. Information leaflets should be publicly accessible alongside trial registration. Transparency is vital to maintain trust and ensure that important research, like SUPPORT, continues. If not, the ultimate losers will be babies and families.

Notes

This commentary arose from a workshop on “Consent in Perinatal Trials” in Sydney on September 5, 2013. It reflects our perspective as present/past clinicians (DW/WTM), researchers (DW/WTM/NG), ethics committee members (DW/NG), and parents of preterm infants (MC). It raises issues from the workshop, but does not necessarily represent all participants.D. Wilkinson was supported for this work by an early career fellowship from the Australian National Health and Medical Research Council (1016641). W. Tarnow-Mordi was a principal investigator on the Australian BOOST II trial of oxygen saturation monitoring.

Dominic J.C Wilkinson, Nicole Gerrand, Melinda Cruz, William Tarnow-Mordi