Initial scientific evaluation of experimental treatments in human subjects occurs in phase 1 trials. These studies are designed to assess safety and dosing for phase 2 efficacy trials. Most phase 1 trials enrol healthy volunteers; however, when the interventions are invasive or carry risks of substantial toxicity, they are tested in patient-subjects, who have at least a small chance of medical benefits to compensate for the risks. In this article we examine the validity of consent to phase 1 oncology trials, which enrol patients who have exhausted standard treatments but retain good functional status. For many years ethical concerns have been voiced about these studies owing to the low prospect of therapeutic benefit and the quality of consent from vulnerable patients seeking cancer-fighting treatment.1–3 Considerable empirical research over the past 20 years has investigated the understanding of patient-subjects regarding the purpose of phase 1 oncology trials and the anticipated benefits and risks of participation in these trials.

Concerns about informed consent have crystallised around three dimensions: ‘therapeutic misconception,’ ‘therapeutic mis-estimation,’ and ‘unrealistic optimism.’ ‘Therapeutic misconception’ denotes the tendency of trial participants to confuse the scientific design and conduct of these studies with standard medical treatment and care.4 Substantial empirical evidence indicates that many trial participants fail to understand or appreciate how their treatment in a clinical trial defined by a scientific protocol differs from the patient-centred orientation of medical care.5 ‘Therapeutic mis-estimation’ refers to a tendency of trial participants to harbour objectively false beliefs about the potential benefits or risks offered by clinical trials.6 Most prominent among participants in phase 1 oncology trials are inflated estimates of the likelihood and magnitude of potential benefit. Unrealistic optimism is alleged to exist when trial participants’ estimates of likely personal benefit are substantially greater, or estimates of risk substantially lower, than their appraisal of benefits or risks for the average person in the trial.7

A recent interview survey of 95 phase 1 oncology trial participants found evidence suggesting defects in understanding or appreciation (application of understanding to one's personal situation) in each of these three dimensions.8 Respectively 68%, 62% and 55% of patients displayed signs of therapeutic misconception, therapeutic mis-estimation, and unrealistic optimism. Comparable results have been observed in several other surveys of phase 1 participants.9–13

These data prompt two distinct questions relating to informed consent for phase 1 trials. First, do the observed defects in understanding and appreciation call for improvements in the process of obtaining informed consent for phase 1 trials? Second, do these defects invalidate consent? Because it is often presumed that informed consent requires understanding and appreciation of basic facts about trial participation, affirmative answers to these questions may seem obvious. We argue instead that, although investigators must take steps to enhance participants’ understanding and appreciation of what phase 1 trials involve, the three types of misunderstanding concerning the purpose, methods and personal risk-benefit ratio of these trials do not necessarily render the consent of trial participants invalid.

Validity of consent to phase 1 trials

A consensus in bioethics identifies several requirements for valid consent to clinical research.14 These include the capacity of patients to make decisions about research participation; accurate disclosure of material information about the purpose of the study, the procedures involved, their potential benefits and risks and alternatives to trial participation; and voluntary choice of study enrolment free of coercion or undue influence. Although information disclosure is aimed at promoting research subjects’ understanding and appreciation of the implications of study participation, there is no consensus about what subjects must understand or appreciate in order to give ethically valid consent.

Informed consent as an ideal of rational decision-making is distinct from consent as a valid agreement between parties to engage in a particular activity, relationship or transaction. Valid consent should be understood as the minimal threshold for permissible interpersonal conduct. If a person's consent to research is invalid, then either it is unethical to enrol that person in the study or enrolment must proceed on the basis of valid surrogate authorisation. Thus it is critically important to determine whether the defects in understanding displayed by many phase I participants are sufficient to render their consent invalid.

Valid consent does not require that prospective research subjects attain an optimal understanding or appreciation of the disclosed information.15 Research on the prevalence of cognitive and affective biases suggests that bias-free rational decision-making is an unrealistic ideal, making it inappropriate as a standard for valid consent.16 While it is desirable to strive for approximation to rational decision-making, calibrating valid consent to the ideal places it outside the capacity of most people, especially those who must make research participation decisions under the pressure of life-threatening or debilitating medical conditions. This point is reinforced by focusing on the pre-eminent value underlying informed consent: personal autonomy. Within bioethics, autonomy is understood as having the opportunity to make voluntary decisions that reflect one's own preferences and values.17 Accordingly, a key question for assessing the validity of consent to a phase 1 trial is whether it reflects a personal decision of the patient-subject that is consistent with his or her preferences and values concerning pursuing treatment in the research context or contributing to scientific knowledge.

Contextual facts about study participation are central to assessing whether the understanding or appreciation of participants is consistent with valid consent. Surveys indicate that most phase 1 oncology participants are motivated primarily by the chance of therapeutic benefit.9–13 ,18 If phase 1 trials offered no prospect of therapeutic benefit, then trial participation would be contrary to the aim of patients who seek an intervention with some potential to have a positive impact on their life-threatening disease. The failure of these patients to understand the absence of any potential for therapeutic benefit in a phase 1 trial would make their consent to trial participation contrary to their preferences and values; the means they have chosen would be incompatible with their ends. Hence, such a choice would not constitute valid consent. Accordingly, it is critical to examine data relating to the benefits of phase 1 oncology trials to judge whether the evidence concerning defects in understanding and appreciation casts doubt on the validity of consent by most trial participants.

Before examining these data, a potential misconception needs to be corrected. The fact that phase 1 trials are designed to evaluate safety, not efficacy, has led some commentators to assert that these trials offer no prospect of therapeutic benefit.3 ,19 ,20 However, this is a non sequitur. The scientific purpose of clinical trials does not determine whether or not they offer a prospect of benefit.21 Instead, the prospect of therapeutic benefit is a function of the expected consequences of study treatments in modifying the disease process. For any given phase 1 study, especially those involving single experimental agents, the potential for benefit cannot be estimated ex ante. Therefore, aggregate data from previous phase 1 trials provide the best indication of prospects for benefit. A systematic review of 460 phase 1 trials sponsored by the National Cancer Institute from 1991 through 2002, encompassing 11 945 participants, found an objective tumour response rate of 10.6%.22 For trials evaluating single agents, the tumour response rate was 4.4%. While tumour response is a surrogate outcome, and not necessarily correlated with prolongation of survival or improved quality of life,21 it is a hoped-for outcome by patients seeking cancer-fighting treatment. Risk of death from toxic events in these trials was 0.49%.

The prospect of therapeutic benefit from phase 1 oncology trials, though certainly low, needs to be evaluated in light of the situation of trial participants.23 They are experiencing progressive disease despite previously undergoing standard treatment interventions. Data on the motivations of trial participants help put this low prospect of therapeutic benefit into perspective. Agrawal and colleagues interviewed 163 patients who consented to participate in phase 1 trials.18 They were asked to rank the type of information they found most useful in deciding to participate. By a large margin, the most important was that the ‘drug kills cancer cells.’ In another study, 89% of early phase trials participants cited ‘because it might help me fight my cancer’ as a major reason for research participation.24 Trial participants also report being motivated by the potential for psychological benefits. Agrawal and colleagues found that 44% of interviewed participants reported that being in the phase 1 study gave them a sense of control, 56% that it gave them hope and 78% that they valued the increased clinical attention associated with being in the study.18

Based on such data, we conclude that, considering their circumstances, phase 1 trial participation offers patients seeking cancer-fighting treatments a reasonable prospect of net benefit.21 ,23 If this conclusion is sound, then motivation to receive personal benefit does not in itself signify any compromise of informed consent.

Some may argue that the proper goals of therapy for patients eligible for phase 1 trial participation ought to be palliative care only. However, no basis exists for an objective judgment to that effect. Notably, in the survey by Agrawal and colleagues, 84% of patients consenting to phase 1 participation expressed awareness of palliative care as an alternative treatment strategy, but only 19% seriously considered this option.18 In view of the motivations of trial participants and the good performance status required for phase 1 enrolment, the lack of readiness to adopt palliative care only is hardly surprising. Patients’ continuing to pursue cancer-fighting treatment is not irrational; nor should the psychological benefits that trials participants perceive be discounted because the prospect of prolonging survival is small.

If motivation for benefit does not in itself invalidate the consent of phase 1 oncology participants, how should we assess evidence of therapeutic misconception, therapeutic mis-estimation and unrealistic optimism? Do these defects in understanding the purpose of phase 1 trials and the likely therapeutic benefits for individual participants suggest that their consent is invalid? For two reasons, we contend that they do not. First, in general, no one questions the capacity of these patients to make informed treatment decisions or the validity of their consent to treatment, either prior to eligibility for phase 1 study participation or as an alternative to research. The consent of the patient who rejects a phase I trial in favour of palliative care or of an alternative chemotherapy—even one with limited potential for benefit—is not typically called into question. Therefore, we see no reason to question patients’ decision-making capacity or the validity of their consent when they opt for anticancer interventions in the context of phase 1 trials. Second, in light of what we know about phase 1 trial participants, it is reasonable to presume that, with some possible exceptions, they would still consent to being in these trials even if they suffered no defects in understanding or appreciation relevant to informed decision-making. Thus there is no prima facie reason to believe that many patients who consent to participate in phase 1 trials would decline to enrol, were they in possession of full understanding. Therefore, it would be disrespectful to these patients, and in violation of their autonomy, to declare their consent invalid on the basis of evidence of therapeutic misconceptions, therapeutic mis-estimation or unrealistic optimism.

To be clear, we are not suggesting that hypothetical or presumed consent is what makes it ethical for patients to enrol in phase 1 oncology trials. Rather, their consent is valid, although often subject to defects in understanding.

Importantly, neither therapeutic mis-estimation nor unrealistic optimism is logically confined to patients making decisions about early phase trial participation.6 Patients with life-threatening disease seeking standard or non-validated treatments in clinical practice may overestimate the often small probability of lasting benefit or believe that they are more likely to benefit than the average patient with their condition. We are unaware of data demonstrating that those motivated to consent to phase 1 trials are more prone than other patients to manifest therapeutic mis-estimation or unrealistic optimism. Just as there is no reason why these two defects of understanding should invalidate consent for treatment decisions outside the research context, they should not be seen as rendering consent to phase 1 trial participation invalid.

Some may object that we have failed to give due attention to the distinction between research and therapy, a distinction that is pivotal to concern about the therapeutic misconception. Certainly, clinical trials differ importantly from medical care in terms of the purpose of the activities and methods employed. Standards for judging decision-making capacity and the validity of consent should be more exacting for research with respect to studies, or interventions within studies, that carry risks but no potential for therapeutic benefit.25 Yet patients considering phase 1 trials, when motivated by therapeutic benefit, are faced with making treatment decisions in the research context. Despite the low probability of lasting therapeutic benefits, failure to understand or appreciate the purpose and methods of phase 1 trials does not make joining these trials irrational or contrary to the interests of eligible patients.

The practice of informed consent serves both negative and positive goals: protecting patients from unwanted medical or research interventions, and empowering them to engage in activities that are consistent with their ends. Protecting patients from research decisions made with deficient understanding of some required elements of information disclosure must coexist with respect for their right to make voluntary choices consistent with their preferences and values. In both these respects, despite evidence of the therapeutic misconception, therapeutic mis-estimation, and unrealistic optimism, patients seeking cancer-fighting treatment can arguably give valid consent to phase 1 trials. To be sure, it is desirable for patient-subjects to be free of such defects in understanding and appreciation. Moreover, it is imperative for investigators to avoid promoting these decisional defects and to endeavour to correct their manifestations in the context of the informed consent process. But it would be disrespectful of patients to paternalistically exclude them from trial participation on the basis of these types of defects in understanding or appreciation of what trial participation involves or offers. Investigators discharge their responsibility to obtain informed consent when they give prospective subjects a fair opportunity to understand material information about trial participation and to make voluntary decisions in light of their preferences and values.15

Enhancing informed decision-making

In suggesting that therapeutic misconception, therapeutic mis-estimation, and unrealistic optimism do not necessarily invalidate consent, we do not mean to diminish the profound responsibility that investigators have to promote informed decision-making for clinical trials. Of course, investigators must accurately disclose material information about the study to prospective subjects in consent documents and conversations, and must endeavour to correct overtly manifested defects in understanding or appreciation. In addition, it may be desirable for investigators to administer simple tests of comprehension that include questions about the purpose of the research, risks and benefits of research interventions, and alternatives to trial participation. However, in administering such tests, investigators should not aim to exclude patients who despite general capacity to make treatment decisions fail to meet a specified standard of understanding of the research, but rather should use such tests to educate patients about the trial. In any case, such tests of comprehension are unlikely to offer robust indicators of understanding or appreciation, as they may merely evaluate recall of information and reflect the answers patients think they must give to enrol in the trial, rather than their genuine beliefs.

Additionally, consideration should be given to a requirement that investigators offer palliative care consultation prior to obtaining consent for phase 1 trials. Presentation of the alternative of palliative care only by clinicians with no vested interest in promoting research can help assure that decisions for trial enrolment accurately reflect the preferences and values of patients. Additionally, this consultation can address current needs for symptom management and help inform these patients’ future treatment or research decision-making. Patients should be encouraged, though not required, to undergo such a consultation prior to phase 1 trial participation. It would be reasonable, however, to require a written decline of the offer.