Consent has become the standard way of legitimising medical interventions and protecting research subjects. Standard requirements for express informed consent are understanding, decision-making capacity and it being voluntary.1 As many patients and research participants fail to satisfy these requirements—for example, due to severe pain, disorientation or lack of consciousness—alternative forms of consent have emerged.

Biobank research has further incited the issue of alternative forms of consent because strict consent requirements can be difficult to fulfil in long-term research projects in a field where new technology, new perspectives and new research questions emerge frequently. Often it is costly and sometimes it is impossible to obtain renewed consent—for example, when donors of biological material are anonymous or dead.

How then are we to adhere to basic principles in research ethics, such as consent, without impeding important research?2^–4 One popular answer to this question is to apply “broad consent” requiring less specific information than express informed consent. This addresses one of the main challenges with consent issues in biobank research: the research person may be competent and voluntary but cannot obtain full understanding of all (future) aspects of the research project.

This article will scrutinise some of the mainstream arguments for broad consent and, by demonstrating their inadequacy, indicate some of the dangers with broad consent and also point towards a more fruitful way to address the challenge of consent in biobank research. In particular, I will address the arguments from Hanson, Helgesson and co-workers.5 6

Table 1 places broad consent in the landscape of alternatives to informed express consent, highlighting the challenges these types of consent try to address.

There is a substantial debate on whether some of these consent procedures are real consents at all—for example, blanket, broad, deferred and hypothetical consent—and whether they violate law and lead to human rights challenges.7 8

In an influential article, Hansson et al argue that broad consent and consent for future research are valid ethically and should be recommended for biobank research.5 They also argue that under given circumstances samples collected for research may be used without consent.6

The premises for applying broad consent are that:

1. personal information related to research is handled safely;

2. donors of biological samples are granted the right to withdraw consent; and

3. new research studies or changes to the legal or ethical authority of a biobank are approved by an ethics review board.5

This article will discuss whether these premises hold and whether the arguments for broad consent in biobank research are sound.

THE MAIN ARGUMENTS FOR BROAD CONSENT

The main arguments for broad (and future) consent are as follows.5 First, biobank research is very useful. Several studies have shown great potential of biobank research, especially in epidemiology, and too strict policies on informed consent may hinder progress in cancer research. Obtaining specific consent jeopardises the amount and quality of research that can be done.

The second argument is that broad consent is a genuine way of respecting people’s autonomy. We have to have good reasons in order to restrict people’s autonomy and, as long as the above mentioned three premises hold, we have no right to hinder people to participate in research even if they do not understand its full scope. If people want to give broad consent they should be allowed to do so. Consent requirements are here seen as a tool necessary to protect people against harm.

The third argument for the use of broad consent in biobank research is an argument from consistency. As current practice in many countries supports the idea that donors should be entitled to give broad consent, so should the legislation. Many countries permit research without consent when samples are not identifiable.

Yet another argument is that the risk in biobank research is low and that this justifies laxer requirements to consent.

DO THE PREMISES HOLD?

Before challenging the arguments for broad consent presented by Hansson et al, let us investigate their premises for applying broad consent. Although the premises appear quite reasonable, they tend to exclude all relevant research. If taken seriously, the premises will hamper rather than facilitate biobank research, which brings us all back to square one.

Safe handling of information

The first premise is that personal information related to research is handled safely. According to Hansson et al this means that “information is coded and handled safely, that secrecy is maintained, and that donors and families are protected from harm”.5 How can this premise be met? Is it possible to protect information through coding and smart cards and so forth? Can we predict future information safety issues in open-ended research?

Coding presupposes that the information at all times is identifiable to someone. It is because we think that the information is valuable that we use it for research and want to protect it. But how can we foresee what value it will have in the future? How are we to assess safety mechanisms for information that we do not yet foresee? It is quite clear that the main risk in biobank research is related to information,9 10 but the problem is that we do not know how the information can be harmful. This is one of the core characteristics of biobank research. All we know is that although the benefit from biobank research may be shared by patient groups and society at large, the informational risks are borne by individual research participants and their relatives. Safe handling of the information is exactly the key problem in biobank research, which has to be addressed and not evaded by criteria of little practical relevance.

Surely, the information safety that Hansson et al presuppose for broad consent can become very difficult to obtain. Even de-identified or anonymised samples can be identified.11 The emergence of various kinds of (forensic) DNA registers makes it much easier to identify a donor of cells or tissue. In a famous criminal case where the Swedish Minister for Foreign Affairs, Anna Lindh, was killed, the investigators used the biological material from a diagnostic biobank (phenylketonuria) to identify the killer (even though that was not strictly necessary for the investigation). There are few reasons to believe that this would not happen to a large-scale research biobank.

In large scale biobank research, biological material and data will be shared among many researchers in many locations making it harder to find efficient safety measures against potential future intrusion. Biobank research, more than traditional medical research, is an adventure into the unknown. Information from this research may violate the right not to know and be of a disadvantage in social relationships (stigmatising)—for example, in relation to employers and insurers.

Hence, either the criterion for information safety ignores core characteristics of biobank research, where the information and its risk is unknown, or it is overly strict and does not promote biobank research as intended.

Withdrawing consent

The second criterion is that research participants should be able to withdraw their consent. One important precondition for withdrawal is of course that you know that you are enrolled in research. For many of the population-based research programmes, such as the Mother Child Cohort in Norway, there are good reasons to believe that people do not know that they are enrolled.

There may of course be many ways to inform people about their participation in biobank research enabling them to consider whether to withdraw. However, if they are to receive detailed individual information about this, it would mean no practical difference from applying informed consent (for a new research project or for broadening or extending an existing research project). Moreover, it could breach the intentions of those participants who wanted to contribute to research and gave a broad consent because they wanted no further knowledge about the research whatsoever. On the other hand, if research participants are not properly informed about the research project and its changing course, there would be no real possibility to withdraw.

There are practical challenges with withdrawal as well. Biological material and data from analysis can be spread over many locations beyond the control of the individual researcher. This is of course enhanced by large international research collaborations and by the intentions to make biological material more available for research.12 13 It may be impossible to withdraw all information on one individual. “You can check out anytime you like, but you can never leave”, as Eric Meslin pointed out at “The Personal Genome Workshop” (Toronto, 27–29 November 2007) where he quoted the lyrics of the Eagles’ “Hotel California”.14 15

Besides, core principles in research ethics urge researchers to keep data for a reasonable time span so that it can be inspected and assessed by peers.

Review by an ethics review board

The third criterion for broad consent is that “new research studies or changes to the legal or ethical authority of a biobank are approved by an ethics review board.”5 The rationale for this is obvious: ethics review boards are supposed to compensate the protective measures lost by allowing broad consent.

Ethics review boards will assess the purpose of the suggested extended or new research and the safe handling of information. However, there appear to be many examples where the ethics review boards are not able to assess the information safety of research participants. Only to mention one recent example: a large Norwegian research project on environmental causes of type 1 diabetes, which included newborn children, analysed collected blood samples for diabetes genes.16^–19 The parents were informed whether their child had diabetes genes. Those who had the genes were to provide monthly faeces samples for 3 years and were to be followed for 15 years. The project started in 2001 and was approved of by the regional ethics committee. Due to concerns by parents, it was questioned whether the research was in accordance with the Biotechnology Act. It was revealed that the research breached with regulations on approvals of genetic tests and on genetic testing for conditions without treatment options (Norwegian Biotechnology Act §§5-3,5-7,7-1). The information safety for the research participants was considered to be controversial by the researchers themselves but not by the regional ethics committee. One reason for this may be that ethics review boards are not competent to review biobank research.10 Maybe they were too eager to promote research?

Considerable differences in research ethics committees’ practices also poses challenges20 as well as the fact that it is not obvious that the ethics review boards would be able to prevent forensic use of research biobanks—for example, although the Ethics & Governance Council of the UK biobank would not give biological information to the police, the security services or to lawyers, they admit that they would do so if they received a court order.21

Hence, it is far from obvious that ethics review boards are able to secure the information safety and provide real withdrawal options to research participants as well as being able to handle the moral challenges following from reduced control and protection through the broadening of consent.

This means then that either the criteria for broad consent are too strict, not allowing broad consent in biobank research, or the criteria do not recognise important aspects of biobank research, making them irrelevant. In both cases they fail to reach their end: facilitating biobank research. Even more, a strict application of the criteria (1–3) would not only prohibit broad consent but biobank research altogether, as well as every kind of research where information security is difficult to guarantee.

How is this possible? The criteria appear so plausible and reasonable. Part of the problem is that the authors’ criteria for broad consent appear to be wider than other comparable criteria22 and although they claim that they are “in line with comment 137 in the explanatory report of the Convention on Human Rights and Biomedicine” and “in accordance with the Council of Europe’s Recommendation on Research on Biological Materials of Human Origin from 2006” this does not seem to be the case.23

Another flaw can be found in the arguments for broad consent.

HYPE AND HOPE OF BIOBANK RESEARCH

Could it be that biobank research is not as important as argued by Hansson et al and others? Although there are examples of highly useful biobank research, it is by far obvious that the outcome of the research has had significant impact on ordinary people’s health.5 24 25 This is not the place to speculate on whether biobank research has been hyped in the same manner as other biotechnologies or to pose unfair expectations to a rather new field of research. The point here is more modestly to indicate that the power of the benefit argument hinges more on a general notion of technology optimism than on well-founded research trends or sound arguments. The benefit argument is based on what has been called “the hopeful principle”, which is as flawed as the precautionary principle.23

Giving too much weight to benefit can be challenging: the argument that “specific consent shows respect for patients and donors … would indeed be true if the process of obtaining specific consent did not jeopardize the amount and quality of research that can be done” as well as arguing that “informed consent might hinder progress in cancer research”5 can become dangerous. History should have taught us that giving primacy to benefit considerations can lead to substantial mistakes in research.

RESPECT FOR AUTONOMY

Nevertheless, the most controversial argument for broad consent is the autonomy argument. Hansson et al see no reasons to limit people’s autonomy to participate in biobank research. They see informed consent as a restriction of people’s autonomy. “Acceptance of broad consent and future consent implies a greater concern for autonomy than if such consents are prohibited.”5 As long as the criteria (1–3) are met, no restriction to autonomy is needed. Autonomy, then, is conceived as non-restriction of consent; blanket consent is the type of consent that respects autonomy the most according to Hansson et al.

This breaches with standard conceptions of autonomy as opportunities for deliberation and with standard aims of consent (protection and deliberation). Second, it confuses autonomy with liberty. Although I have the liberty to waive all kinds of consent, it does not guarantee that I will maintain my autonomy. Third, it breaches with standard criteria for autonomy, such as understanding.

Despite recommending broad consent, the authors reject blanket consent although they claim that blanket consent pays even more respect to people’s autonomy. Again they use the benefit argument: “Our argument against blanket consent is the societal importance of biobank-related biomedical research: blanket consent could lead to the consumption of important samples, potentially allowing commercial, purely technical or political applications (for example, development of new methods for criminal investigations, for refinement of paternity tests or for helping immigrant authorities to identify the ethnic origin of immigrants).” This time, the benefit argument is used to restrict the autonomy argument. It is hard to see how the same (benefit) argument is not valid for limiting broad consent as well.

CONSISTENCY

The consistency argument—that is, broad consent is consistent with current practice and therefore should become applicable, appears more plausible. For example, in the Norwegian context a “broad” consent is applied in practice, whereas the Biobank Act requires “voluntary, express, and informed consent” (Biobank Act, 2003, §12) based on information on “objective, method, risk, discomfort, and consequences”. Accordingly, the proposed new Norwegian Health Research Act also suggests broad consent (§14).26 However, the Directorate of Health and Social Affairs, who approve research biobanks, argues that the current practice is in accordance with the existing law (that is, informed consent and not broad consent).

It is clear that legislation countered to public practice is deemed to fail and may be counterproductive because it undermines the belief in the legislative system altogether. Nevertheless, it is not generally so that practice directs laws. Most people tend to break the speed limits from time to time but we still have (and believe in the justification of) speed limits. There are no convincing reasons to believe that biobank regulation is an exception.

Hence, although this pragmatic argument appears plausible at first glance, it is not necessarily so upon scrutiny.

LOW RISK

An additional argument is the argument from low risk: “However, if the risk of harm from future biomedical research is low and sufficiently well controlled, if the participant voluntarily accepts this level of risk, and if there is a mechanism for withdrawal, then there is no reason why broad consent and future consent to such studies should not be acceptable.”5 The crucial point here is how we can know and predict low risk of future research projects. Furthermore, research participants’ can be unaware of their research participation and of the risks.27

The nature of biobank research is that it is open ended. The purpose of broad consent is exactly to allow and promote research where future use of biological samples and data from analysis can be used in new and yet unforeseen ways.

Hence, the main challenge with the low risk argument is that it misses the characteristics of biobank research.28 This research does not have substantial risk in terms of bodily harm and the practical discomfort or inconveniences with participating are insignificant if apparent at all. The main risk in biobank research is related to information.9 10 The argument that a future (unexpected) increase in risk will be handled by ethics review boards, opt out mechanisms or safety handling procedures is, as has been argued, not convincing. The low risk argument is part of Hansson et al’s benefit argument, which follows the flawed hope principle, as mentioned above.

BYE-BYE TO (BROAD) CONSENT

Where does this leave us? As both the criteria for applying broad consent and the arguments for using it appear flawed, do we have to abandon broad consent and abstain from potentially fruitful research due to cumbersome and rigid consent procedures? Not necessarily! It may well be ethically defensible to apply an authorisation for research on the basis of limited knowledge of the research project and under given conditions—for example, third party assessment and withdrawal from research.29^–31 There are ways to make authorisations work.32

However, is this not a play with words? Allowing broad authorisations but banning broad consent appears ludicrous. However, as broad consent does not fulfil basic criteria for being a valid consent, the use of the term gives an impression of false justification.

Consent is the standard way to protect research subjects and to promote their wellbeing. It has become the watchword of research ethics, giving the impression that if consent is applied, all ethical issues are settled. Hence, consent has become an ethical device for making research morally acceptable. This can conceal serious challenges. As table 1 indicates, the many forms of consent are results of profound challenges. Many of them, such as blanket consent, broad consent and future consent, are not consent by standard conceptions of consent. But by using the term “consent” we tend to lend legitimacy to certain types of research and ignore the real challenges. Using more specific authorisations force us to address these challenges explicitly.

The dangers with broadening consent are illustrated by the authors’ argument that samples where donors have given consent limited to specific research—for example, to cancer research—can be used for other kinds of research without new consent.6 They argue that because the donors have not refused to participate in other types of research, the samples can be used for such purposes without new consent. However, if the authors’ consent to have £900 drawn from their bank account and donated to cancer research, can we then draw £9000 from their account because their primary donation indicates a positive attitude towards cancer research?

As stated in the Declaration of Helsinki, article 5: “In medical research on human subjects, considerations related to the well-being of the human subject should take precedence over the interests of science and society”, and in article 18, “Medical research involving human subjects should only be conducted if the importance of the objective outweighs the inherent risks and burdens to the subject.” Hansson, Helgesson and co-workers try to defend otherwise unethical research by downplaying or ignoring the risk: if the risk is zero, every potentially beneficial research is acceptable. Moreover, they ignore a basic requirement for autonomy: understanding. Combined with an over sized eagerness to promote research and recruitment,33 this combination of risk and information ignorance can become dangerous.