Zelen Randomization: Attitudes of Parents Participating in a Neonatal Clinical Trial

https://doi.org/10.1016/S0197-2456(98)00049-XGet rights and content

Abstract

Recruitment to randomized controlled trials can be difficult for all parties involved. An alternative to the standard process has been suggested for trials in which the control group receives standard treatment or nontreatment. In this approach (the Zelen design), randomization precedes consent, which is only sought from those allocated to the experimental arm of a trial. The control group is thus unaware that randomization has taken place. As a controversial method, this approach has been often suggested but rarely used. Here we describe how 44 parents recruited to a difficult neonatal trial that used conventional randomization reacted to the idea of Zelen randomization. The arguments they gave for and against the method pertain to four areas: the giving or withholding of information, the effect on decision making, the use of data without parental knowledge, and the long-term impact for parents. The parents were evenly divided in accepting or rejecting the method. Further analysis showed that those rejecting Zelen randomization were more likely to be parents of infants allocated to the control group. This suggests that those from whom consent would not be sought, the group that this approach is primarily meant to protect, are most likely to find it unacceptable.

Introduction

The giving of informed consent is a prerequisite to participation in most randomized clinical trials (RCTs). In theory, it is appropriate and desirable that consent is given; in practice, informing a person fully about medical uncertainty and about the random allocation of treatment can be difficult and distressing for all parties involved. These inherent difficulties are further exacerbated in specialties where patients or their proxies are already experiencing a high degree of emotional strain. Discomfort with the process of consent can lead to poor rates of participation among professionals and therefore slow or inadequate rates of accrual of patients 1, 2, 3, 4, 5, 6. Fully informed consent in some circumstances has been described as “needlessly cruel,” with a statement that “attempts to gain the `informed’ participation of patients in randomized clinical trials are already doing harm in many individual cases [7].” Problems with consent therefore have serious implications for the success of RCTs, for professional morale, and also for the well-being of the participants.

Zelen addressed these difficulties by proposing a controversial departure from the conventional approach 8, 9, 10, 11, 12. Ordinarily, a clinician informs potential participants about a trial, and if consent is given, randomization follows. If consent is withheld, the standard treatment options (including no treatment) will usually be offered (Figure 1). In Zelen’s design, the clinician randomizes before asking for consent, if asking for consent at all. Consent, where sought, is for the randomly allocated treatment and for trial participation but not for randomization. This method has been termed variously randomized consent design, prerandomization design, and postrandomization design. In this article, we use the term “Zelen randomization.”

Two types of Zelen designs exist: “single consent” and “double consent.” In the single-consent design, the patients allocated to the control arm of the trial are not asked for consent—they simply receive the standard treatment without mention of the trial. Clinicians ask those allocated to the experimental arm of the trial whether they consent to receive the experimental treatment. If they decline, they may be receive the standard (or another) treatment (Figure 2). The double consent design differs in that the patients allocated to the control arm are asked for consent for the standard treatment and those among them who decline may receive the experimental (or some other) treatment. Those allocated to the experimental arm are asked for consent for experimental treatment. If they decline, they may receive standard (or some other) treatment (Figure 3).

Zelen’s designs are limited in that they are applicable only to certain trials. The single consent design is suitable only when the experimental treatment is compared with a best-known standard treatment. For a treatment to be considered experimental, it is unlikely to have been fully evaluated. So in this study, we concentrate on the situation in which the experimental treatment is only available within the trial, although the double consent design is suitable only when the experimental treatment is also available outside the trial.

Trials that use the Zelen approaches cannot be “blind” because the clinician is made aware of the treatment allocation to approach patients for consent to receive that specified treatment [13]. An assessment can, however, be made blind to allocation. Finally, trials requiring the active cooperation of the patient in a nonstandard protocol or follow-up are inappropriate when patients have not consented to participation.

Although there has been wide discussion of Zelen randomization, only a few trials are known to have used it, and most were conducted in the United States [for reviews, see 12, 14, 15]. Some use cluster randomizations [e.g., see 16], and we do not list these here. Most are in the field of cancer 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, but there are also examples from osteoarthritis [32] and neonatology 33, 34.

Zelen randomization has ethical and statistical advantages and disadvantages. Its appeal lies in the fact that in difficult circumstances a clinician need not ask for consent for the trial or for randomization, only for treatment. When stress levels are high, such simplification could ease both information giving and decision making. It might also improve rates of accrual, and therefore statistical power, in trials in which recruitment might be difficult. This might be because it leads to options that could be more acceptable to some patients and clinicians, but mainly because its design is weighted to include those who might otherwise have refused. In a conventional trial, all consenting patients are followed-up in an intention-to-treat-analysis; that is, they are analyzed in accordance with their allocation, regardless of whether they actually go on to receive the allocated treatment [35]. In Zelen’s designs, all randomized patients are followed on an intention-to-treat basis. All data are therefore appropriate for analysis as, in both of Zelen’s designs, trialists may analyze data even from those who decline the offered treatment.

Concerns about Zelen randomization arise from the fact that consent for randomization is not given in advance by any patients and that consent for treatment on protocol is not given by all. A proportion of the patients remain unaware of their role and are automatically included in the trial. The remainder have no explicit option to refuse to participate. Although it is widely accepted that patients whose data are used in epidemiologic studies are not necessarily informed of their roles in research, the acceptability of research without consent is less clear for RCTs. Of particular concern is the fact that not all patients are told that their treatment is chosen by chance 36, 37.

Moreover, there may be statistical concerns over bias. If, for instance, the experimental group of patients are required to attend for more follow-up than the control group, their relapses may be detected sooner. Lack of blinding may cause differential provision of concomitant treatments and/or differential recruitment into the two treatment arms. The largest statistical problem may be a reduction in power where an intention-to-treat analysis is used, the very concern that Zelen randomization aims to address. Power may be lost because of a dilution effect if a large proportion of patients reject their allocated treatment (termed crossover or transfer). Theoretical and empirical studies have shown that reduction in power resulting from crossover can be considerable, and the problem is compounded if the extent of dilution differs between the two trial arms 38, 39, 40. Because trialists cannot usually know the extent of crossover in advance, sample size calculations can be underestimated.

Although ethicists and statisticians have discussed the use of Zelen randomization, an important perspective is clearly missing from the debate: that of the participants themselves. Although studies have considered the views of healthy volunteers, [41] no evidence has yet demonstrated how those with experience of trials feel about the Zelen approach.

This article describes the responses to Zelen randomization from parents who had consented for their critically ill infants to participate in a conventional neonatal trial in which informed consent preceded random allocation of treatment. RCTs in such circumstances are of particular interest, because they inevitably involve persons who are extremely stressed. Parents are not only concerned for the well-being of their sick infants but are often within days or even hours of their birth and any associated traumas. They may be physically exhausted or deprived of sleep, and the mothers may be experiencing the effects of pain-killing or sedative drugs. Some have argued that consent for neonatal trials is not only difficult to obtain but may be inadequate, amounting to “an elaborate ritual” [42] or a “sham” [43]. In such a context, Zelen’s approach has been seen as appropriate [43].

Section snippets

The UK collaborative trial of extra corporeal membrane oxygenation

The UK Collaborative Trial of Extra Corporeal Membrane Oxygenation (The ECMO Trial) compared two methods of life support for mature neonates with acute but potentially reversible respiratory failure. The infants recruited to the trial were at high risk of death. At randomization they were already in a neonatal intensive care unit and receiving ventilatory support (“conventional management” or CM), one of the two methods the trial evaluated. Researchers compared CM to oxygenation of the blood

Discussion

Zelen randomization has been the subject of professional debate, but no attention has previously been paid to the views of those with actual experience of trial participation. The way in which parents in this study often engaged with and extended the issues involved in the subject was impressive and indicated the value of their particular perspective. Their interest was rooted in personal experience, and they empathized highly with others in similar situations. Despite the short time available

Conclusion

The subject of Zelen randomization proved to be a highly engaging topic for the parents in this study, who were divided in its acceptance or rejection. They saw a trade-off of protection and information, desirable to some, anathema to others. They also saw an opposition of the short-term benefits and long-term consequences. It became clear from their responses that Zelen randomization would create very different circumstances for participants depending on their allocation. Those who found this

Acknowledgements

This study was funded by the Department of Health and the Nuffield Foundation and developed from research funded by the Wellcome Trust. We thank Carole Harris from the ECMO Trial office and Sally Roberts at the Centre for Family Research for their help. Our thanks also go to the parents who kindly gave the information upon which this research is based.

References (49)

  • K.M. Taylor et al.

    Fundamental dilemmas of the randomized clinical trial processresults of a survey of the 1737 Eastern Cooperative Oncology Group investigators

    J Clin Oncol

    (1994)
  • G. Tognoni et al.

    Randomized clinical trials in general practicelessons from a failure

    BMJ

    (1991)
  • J.S. Tobias et al.

    Fully informed consent can be needlessly cruel

    BMJ

    (1993)
  • M. Zelen

    Statistical options in clinical trials

    Semin Oncol

    (1997)
  • M. Zelen

    A new design for randomized clinical trials

    N Engl J Med

    (1979)
  • Zelen M. Strategy and alternate designs in cancer clinical trials....
  • M. Zelen

    Randomized consent designs for clinical trialsan update

    Stat Med

    (1990)
  • R.I. Horwitz et al.

    Advantages and drawbacks of the Zelen design for randomized clinical trials

    J Clin Pharmacol

    (1980)
  • S.S. Ellenberg

    Randomization designs in comparative clinical trials

    N Engl J Med

    (1984)
  • C.G. Moertel et al.

    Combined 5-fluorouracil and radiation therapy as a surgical adjuvant for poor prognosis gastric carcinoma

    J Clin Oncol

    (1984)
  • B. Fisher et al.

    Five year results of a randomised clinical trial comparing total mastectomy and segmental mastectomy with or without radiation in the treatment of breast cancer

    N Engl J Med

    (1985)
  • N. Wolmark et al.

    Post operative adjuvant chemotherapy or radiation therapy for rectal cancerresults from NSAPB protocol C-01

    J Natl Cancer Inst

    (1988)
  • B. Fisher et al.

    Post operative adjuvant chemotherapy or radiation therapy for rectal cancerresults from NSAPB protocol R-01

    J Natl Cancer Inst

    (1988)
  • B. Fisher et al.

    Eight year results of a randomised clinical trial comparing total mastectomy and lumpectomy with or without irradiation in the treatment of breast cancer

    N Engl J Med

    (1989)
  • Cited by (68)

    • The adapted Zelen was a feasible design to trial exercise in myeloma survivors

      2020, Journal of Clinical Epidemiology
      Citation Excerpt :

      Qualitative data from participants in a neonatal clinic elicited mixed attitudes toward Zelen randomization, with participants evenly split between accepting and not accepting the design. Some reported that they found the Zelen design underhand yet acknowledged if you were in the control group, it was kinder than an RCT, as you were not aware you had not received a potentially desirable treatment [26]. We used this design over a standard RCT because we felt it would cause patient distress if they were not allocated to the exercise intervention and may lead to contamination and participants to drop out.

    • OARSI Clinical Trials Recommendations: Design and conduct ofimplementation trials of interventions for osteoarthritis

      2015, Osteoarthritis and Cartilage
      Citation Excerpt :

      The value of courteously providing information to patients should not be underestimated, although it is possible that this results in either an increase in goodwill, or an increase in patient concern. Failure to provide this information risks psychological harm, as subsequent discovery of inclusion in an experiment might result in a sense of violation79,80. There are no clear rules for the degree to which such information should be given; this depends on the type of intervention studied and the potential to withdraw from the intervention or data collection, and should therefore be covenanted with patient boards, the professionals and ethical committees involved.

    • Conducting clinical trials in pediatrics

      2009, Critical Care Medicine
    • Research in the Newborn Period

      2020, Emerging Topics and Controversies in Neonatology
    • The Value of Decompressive Craniectomy in Traumatic Brain Injury

      2019, Evidence for Neurosurgery: Effective Procedures and Treatment
    View all citing articles on Scopus
    View full text