Informed consent and phase IV non-interventional drug research

Abstract

Most of the literature on informed consent in pharmaceutical drug research works on the assumption that informed consent is something that is homogeneous and thus can be rendered procedurally universal. This may be justifiable to a certain extent owing to the fact these are all drug trials anyway. Nevertheless, in spite of this general similarity, we also know that the clinical drug development phases are characteristically different, and that phase IV is very different from the other phases because, owing to its postmarketing nature, it is much more varied in scope and in type. Thus, it is worthwhile looking into the ethical nuances relevant to the informed consent process in phase IV non-interventional drug research. We shall deal with the issues on the necessity of informed consent for this type of research and then discuss the possibilities for an opt-out system. We conclude that informed consent is necessary for non-interventional studies, and thus any form of waiving of rights of participants to informed consent must have a valid substantial justification. The distinct character of phase IV accounts for the difference in content of the informed consent document compared to that of earlier phases, and both opt-in and opt-out procedures are ethically justifiable as long as the participant’s participation remains informed and voluntary.

Key words::

• Consent forms
• Informed consent
• Non-interventional studies
• Observational studies
• Phase IV
• Research ethics

Introduction

There is a vast body of literature about informed consent in research, although most of this literature is based on the assumption that informed consent ought to be something that has uniform application. As such, when informed consent in research is talked about, it is assumed that what applies in a phase I dose tolerance study should also be applicable to a phase IV comparative effectiveness study. To a certain extent this is justified by the fact that both are phases of drug development, and thus trial participants are all taking part for the primary purpose of contributing to the furthering of knowledge about a drug that may benefit future patients. Nevertheless, in spite of this general similarity, it is also known that clinical drug development phases are characteristically different1. Phase IV is very much different from the other phases owing to its postmarketing nature – it is much more varied in scope and in type. There are many types of phase IV trial, and the comprehensive definition of the International Conference on Harmonisation of Technical Requirements for Registration of Pharmaceuticals for Human Use (ICH) refers to the enormity and heterogeneity of this phase:

‘ … all studies (other than routine surveillance) performed after drug approval and related to the approved indication … Commonly conducted studies include drug–drug interaction, dose-response or safety studies and studies designed to support use under the approved indication, e.g. mortality/morbidity studies, epidemiological studies’1.

As such, the heterogeneity, the scale, and the huge variation present in phase IV would intuitively tell us that though participants may all be ‘consenting’ to the universal goal of drug trials, the multifaceted and postmarketing nature of phase IV sets it aside as a special case. Thus, how informed consent particularly applies to phase IV may be radically different to how applies to the earlier phases, even if the original essence of it remains the same, i.e. to uphold the ‘respect for persons’2 by safeguarding their autonomy and protecting them from risks2. Or, in Mansson and O’Neill’s terms, the aims of informed consent are that that participants are not ‘abused, manipulated or undermined, or wronged’3.

Since it is presumptuous to speak about informed consent in phase IV as if it were homogeneous, it would be best to look into specific types of phase IV, and discuss them individually. In this paper, the authors look into the ethical issues and the application of informed consent in phase IV non-randomised, non-interventional, not data-only studies. The need to look into the nuances of late-phase trials is highlighted in the current European Medicines Agency Road Map to 2015 draft in relation to the goal of ‘maximising the value of information generated in the post-authorisation phase’4 as this is valuable for both conditional marketing authorisation and post-authorisation follow-ups4.

To address the main topic of concern, we shall first look into the definition of a non-interventional study. Then we shall deal with the issues on the necessity of informed consent for this type of research. Lastly, we shall proceed to the question of what sort of informed consent procedure may be ethically justifiable.

What is a non-interventional study (NIS)?

What is an NIS? Directive 2001/20/EC article 25 defines a non-interventional trial in the following manner:

A study where the medicinal product(s) is (are) prescribed in the usual manner in accordance with the terms of the marketing authorisation. The assignment of the patient to a particular therapeutic strategy is not decided in advance by a trial protocol but falls within current practice and the prescription of the medicine is clearly separated from the decision to include the patient in the study. No additional diagnostic or monitoring procedures shall be applied to the patients and epidemiological methods shall be used for the analysis of collected data.

Thus, by NIS we refer to an observational study of marketed drug/s that is/are prescribed in accordance with their marketing authorisation. This means that the trial protocol does not in principle affect the prescription practice and the diagnostic tests fall within regular clinical procedures. Questionnaires, interviews, and blood samples ‘may be considered as normal clinical practice’6 and thus normally may form part of an NIS, depending on the burden that it may require to the participants. An NIS is legally distinguished from a clinical trial and thus does not fall within Directive 2001/20/EC on the Implementation of Good Clinical Practice in the Conduct of Clinical Trials on Medicinal Products for Human Use5. That boundary between a clinical trial and an NIS when questionnaires, interviews, and blood samples are used may have grey areas which we cannot amply address here; nevertheless, keeping this grey area in mind would be useful in our discussion later in this paper.

We can further classify an NIS as data only (either anonymised, coded, or identifiable), such as registry studies and non-experimental studies that do not include randomisation but has actual subject enrolment. This article deals with the second type of NIS.

The question of necessity of informed consent in phase IV NIS

The problem of informed consent in an NIS begins with the fact that at least in Europe, ‘the regulatory governance of NIS still suffers from significant disharmony … with some countries implementing very specific and detailed legislation (e.g., Spain), while other have no legislation at all (e.g., Austria)’7. There is also no International Conference on Harmonisation (ICH) document specifically for NIS or generally for observational studies. Pharmacovigilance guidelines, such as Volume 9A6 and the ICH Pharmacovigilance Planning E2E8 document, partly guide NISs, although these are limited to NISs that are safety studies. The CIOMS International Ethical Guidelines for Epidemiological Studies9, which states that informed consent is a requirement in epidemiological studies unless a waiver is justifiable in ‘exceptional’ circumstances, is the closest international guideline that addresses informed consent in NIS. We shall go back to this point later.

This lack of unified legislation leads to ‘considerable room for inconsistencies in the approach to ethical review’ in observational studies10. The necessity of informed consent is frequently questioned because of the threat of selection bias11,12. Lemaire, pointing out the need for European legislation ‘adapted to the different categories of clinical research’13, reacted to an article by Ferrer, Artigas, Levy, et al. who said that in a particular quality improvement clinical research, ‘the need for informed consent was waived in view of the observational … nature of the study’13. What is interesting about this quote is the relation made between not needing informed consent and the observational nature of research. In one prospective, open-label study meant to ‘assess the clinical performance of a foam dressing in the management of open wounds’14, the authors stated, ‘No inclusion or exclusion criteria were provided, rendering the study exempt from institutional review board approval or informed consent restrictions according to European research requirements’14. Thus, again, we have authors who equate the observational nature of the study, plus the lack of inclusion and exclusion criteria, as a reason for informed consent ‘restrictions’ exemption. Thus the question, is informed consent necessary for NIS?

To address the question on the necessity of informed consent in NIS, we ought to look into the possible reasons stated in current literature why informed consent may not be necessary for NIS. We can categorise these reasons as follows: (1) potential benefits to society; (2) risk is minimal or none; and (3) autonomy is not meaningfully infringed.

The potential benefit to society argument

The first argument that waiving consent in NIS would greatly benefit science and society is what is often available in literature, especially in studies that demonstrate selection bias resulting from the informed consent requirement in observational research11,12,15. This argument springs from the aspiration for optimal efficiency as well as the assertion that increased scientific efficiency in research results in better health services, or in the specific case of phase IV drug research, better estimation of treatment effects in therapeutic use11. As such, better health services and increased knowledge about a drug’s effectiveness and safety would in fact benefit more and would, in the long run, reduce risk.

The main issue about this argument is that it does not touch on the very purpose of informed consent; instead, it goes straightaway into the ‘benefits’ of waiving it. Informed consent, as stated by Levine in the Encyclopedia of Bioethics, has a two-fold purpose: to minimise risk of research subjects, and to respect persons by giving them the right to choose (otherwise called the right to autonomy)2. As such, without discounting the enormous value of phase IV drug research on therapeutic use, arguing for the waiving of informed consent solely because of this value misses the point. Simply, the argument fails due to irrelevance. It is one thing to have valid and efficient research, and it is quite another to respect the basic right of human beings to choose for themselves. Therefore, the potential benefit to society argument per se does not give us reasonable grounds to think that informed consent is not necessary.

Risk as minimal argument

This second argument states that informed consent is not necessary in NIS since its observational nature that works on what is ‘safe, evidence-based, and standard procedures’16 does not impose additional risks or at the most imposes only minimal risks and burdens to the study participants. The research may continue with virtually no involvement of the researcher in terms of prescription (if the researcher is distinct from the attending physician), and diagnostic and monitoring procedures are well within clinical practice. The question is, since NIS works within clinical practice and therefore no additional risks are imposed or at most research-related risks are minimal, is this sufficient argument to say that informed consent is not necessary?

This argument is closer to the one of the purposes of informed consent, i.e. the minimisation of risk in studies. Nevertheless, it is one thing to say that informed consent protects research participants from risk, and it is quite another to say that the absence of risk justifies not asking for consent. Protection against unacceptable risk is an intended effect of informed consent by ensuring that the risks of a study are known and evaluated also by the prospective participant. Nevertheless, the fact that risks are minimal or even absent does not give the researcher the liberty or the right to include a person in a research without consent. If the right to choose, which is one of the purposes of informed consent, is a de facto basic right, then a person may exercise that right irrelevant of the external state of affairs, i.e. whether risk is present or not. Besides, wrongdoing is not restricted to physical harm, as Feinberg eloquently explains in The Moral Limits of the Criminal Law. A person can be wronged without being harmed, which Feinberg categorically calls ‘harmless wrongdoing’17,18. As such, the risk as minimal argument per se is also an insufficient argument to justify the non-necessity of informed consent.

No meaningful infringement of autonomy argument

The third argument states that informed consent is not necessary in NIS because its absence does not ‘amount to any meaningful infringement of patients’ autonomy’16 since, by alluding to the reasonable-person standard, ‘there could be no reasonable or ethical grounds for any patient to object to being included in the study without his or her consent’16. Thus, an NIS that has been judged by an ethics committee to be reasonable must by default also be consented to.

Just like any argument that utilises the reasonable-person standard argument, this argument also ‘encounters conceptual, moral, and practical difficulties’19. On the conceptual level, as Beauchamp and Childress said, the concept of a ‘reasonable person’ has not been carefully defined19. Logically, it encounters the problem of falsely equating presumed consent with actual informed consent, and in this case, the former is allowed to effectively replace the latter. This then denies the person her/his actual right to exercise consent. It is conceptually more worthwhile to make a distinction between a ‘consentable’ study as judged by an ethics committee and the actual informed consent of a participant.

On the moral level, since the no meaningful infringement of autonomy argument makes the whole exercise of consent depend upon the researcher’s (or ethics committee’s) perspective of what is reasonable and ethical, the essential moral purpose of informed consent is defeated. Since it is the generic right of an agent to make plans for herself/himself21 and to do so according to her/his own definition of purpose and fulfilment, it is immediately a violation of her/his generic right as an agent to prejudge that in this instance there are no reasonable or ethical grounds for her/him to object to consent. Simply put, as a competent agent, a person chooses what is reasonable for her/him, and her/his consent depends on herself/himself and not on a prejudgment by anybody else. This applies either to the bloodletting in NIS, or even in any other research. Therefore, the no meaningful infringement of autonomy argument also does not justify the non-necessity of consent.

Hence, we could say that informed consent is a widely accepted right that human beings have when it comes to participation in research, including NIS, and none of the arguments above show that NIS is an exception to this universal rule. Considering NIS in general, there is no compelling substantive justification for such a general infringement because experience shows that NIS can and is administered with informed consent, and thus it would be unethical to infringe a participant’s generic right to consent simply for the convenience of the researchers when it is quite possible to obtain it and there are no urgent and convincing reasons not to do so. Given thus, a more reasonable question would be, are there more compelling substantive reasons to infringe the right to consent in a particular research?

The waivability of informed consent in a particular research denotes the existence of ‘substantive justifications’20 for the waiving. Take note that when consent is waived, this means that there is a prima facie right to consent20 and this right is considered to be justifiably waivable based on the circumstances of a particular research. Thus, there are ‘exceptional’ circumstances9 when an NIS can be done without informed consent. The usual justification for the waiving of consent is the ‘impracticability’ of it in certain types of studies22, and even that justification must meet certain conditions. If a considerable grey area exists in a study such that bloodletting has reached a point of being burdensome and thus a study that is technically an NIS is beginning to feel like a clinical trial to a participant, then it is highly doubtful that impracticality is sufficient reason to waive this right to consent.

Thus, although there may be particular instances when informed consent may be waived because of compelling substantive justificatory reasons, there are no ethically justifiable reasons to universally forego informed consent in NIS; in fact, there are prima facie ethical and legal reasons for it. Plainly, and in agreement with the CIOMS Ethical Guidelines for Epidemiological Studies9, informed consent must be the standard in NIS, and waiving it must be an exception.

On the application of informed consent in NIS

Saying that informed consent is a standard requirement in NIS is different from saying that the informed consent in NIS ought to be the same as the informed consent process that happens in phases I–III. That there is a difference in immediate goals, procedure, and scope between NIS and earlier trials tell us that this may also account for a different approach in informed consent. That differences in study procedure and norms may account for differences in informed consent application has been previously suggested by Burgess23 and Hansson24. Nevertheless, the question remains, ‘how can informed consent in phase IV be different from that of earlier phases?’

In the Oxford Textbook of Clinical Research Ethics, Flory, Wendler and Emanuel25 named four factors that are relevant for informed consent in health research, namely: (1) understanding the purpose of research, (2) understanding voluntariness, (3) understanding protocol design and randomisation, and (4)understanding of risks and benefits. We shall go into these factors and see their relevance in NIS research.

First, when we speak of understanding the purpose of research, we refer to understanding the general information about the study, some specific information, the purpose of the study (which in other phases primarily refers to ‘creating generalisable knowledge’ and benefiting future patients), and with it the avoidance of therapeutic misconception. All these are the universal concerns in earlier phases, and phase IV shares the concern of making sure that participants understand what the research is about and what if any does it ask from them (like answering questionnaires). But phase IV NIS differs from other phases since therapeutic misconception ought not to be a concern since it is reasonable for participants to expect from the medical intervention which by now is considered as standard clinical practice. Also, since there ought to be virtually ‘no additional diagnostic or monitoring procedures’5 in NIS, then the risk of misperceiving research as clinical is almost nil.

Second, understanding voluntariness refers to the awareness of participants that they could withdraw without any consequence to their health care25. This concern is something that phase IV NIS shares with the other phases since it is very possible that participants would consider participation in research as a requirement for their care.

Third, as regards understanding protocol design and randomisation, the challenge in NIS is simpler compared to earlier phases since, though NIS researchers would need to make sure that participants understand that their information is important in research and that they would probably need to fill out some questionnaires or answer some interview questions, researchers need not struggle with explaining randomisation, a concept that some researchers show is difficult for participants to understand25.

Fourth, as regards understanding risks and benefits, earlier phases are concerned about too much optimism about the benefits of an investigational medicinal product, or confusion or lack of understanding about the benefits and risks of a trial drug. For NIS, this ought not to be a concern since the expectations as well as the understanding of risks and benefits of a marketed drug all fall into clinical practice and not into the study per se. This is what is meant by an NIS not having additional risks.

As such, informed consent concerns in phase IV NIS have substantial differences compared to the traditional research informed consent concerns that mostly are applicable to earlier phases of a drug trial. But then again, how does this translate to the application of informed consent in phase IV NIS?

Manner and content of informed consent in NIS

Informed consent in phase IV NIS is different from that of earlier trials in the sense that ethical concerns may revolve around the very issue of the necessity or waivability of consent and that since research-related risks are very minimal if not non-existent, the relevant factors in the informed procedure are expectedly fewer. Thus the informed consent process may be different from the other phases in at least two aspects: the content of the informed consent form and the manner of the informed consent procedure. We have already seen above how the content of the informed consent form in NIS may be different from the informed consent form used in other phases, as an NIS is less demanding to participants when it comes to the four factors for informed consent in health research compared to that of other phases. Thus, an NIS informed consent form would need to contain information about the purpose of the research, an explanation that research participation is voluntary, some information about the protocol, and some statements about risks and benefits, if any. But it need not address therapeutic misconception and randomisation since these are irrelevant, and statements about benefits and risks cannot be as scrupulous as those in earlier phases as the trial per se should not carry more than minimal risk, and ideally should impose no additional risks.

Regarding the manner of the informed consent procedure, the regular opt-in procedure remains the ideal standard9; nevertheless, there may be instances when a regular opt-in procedure is impracticable or may compromise the integrity of the research. In this case, an opt-out procedure may be considered preferable or more suitable. Some trials have demonstrated that an opt-out system could truly help in overcoming some of the problems on recruitment barriers27, as well as the problem of selection bias threat, but our concern is, could an opt-out system truly ‘inform’ a participant, or would this be an institutional informed consent that does not necessarily facilitate autonomous authorisation19?

Looking at the defunct UK opt-out system where an initial information letter from the National Health Service gives the patients the chance to opt-out by not making their contact information available to research teams, Hewison and Haines found that many do not really object to this opt-out system28. In fact, Hewison and Haines consider this system as more supportive of the aim to inform than the opt-in system since this opt-out system immediately puts the researcher in contact with the participant, thus providing the ‘support and reassurance that personal contact can provide’28. Another study by Clark and Findlay29 demonstrated how strategic information dissemination plus the ease of a system to opt-out addresses the worries of an opt-out system being deficient in terms of informing and allowing the participants to choose. As such, an opt-out system may be part of the different ways in which informed consent may be secured in a phase IV NIS, as long as it satisfactorily addresses the purposes of informed consent and the necessary factors relevant in the informed consent procedure.

Conclusion

Informed consent in phase IV NIS is a necessary factor which may be waived in exceptional situations, as long as a valid substantive justification is in place. The distinct character of phase IV NIS also gives a different slant to the manner and content of informed consent such that not all the relevant factors for informed consent in health research are applicable to NIS, and both opt-in and opt-out procedures may also be utilised, as long as a person’s participation in an NIS remains informed and voluntary.

Transparency

Declaration of funding

This study was carried out within the context of the Escher project (T6-202), a Dutch Top Institute Pharma project. R.B.’s PhD project is funded by Dutch Top Institute Pharma.

Declaration of financial/other relationships

J.R. works and holds stock in GlaxoSmithKline, the Netherlands. R.B., G.v.T, and J.v.D have no financial or other relevant relationships to declare.

No assistance in the preparation of this article is declared.

Notes

*The terms ‘study’ and ‘trial’ are interchangeable.

†Directive 2001/20/EC, which puts forth European Union regulation on good clinical practice of clinical trials, defines a clinical trial as ‘any investigation in human subjects intended to discover or verify the clinical, pharmacological and/or other pharmacodynamic effects of one or more investigational medicinal product(s), and/or to identify and adverse reactions to one or more investigational medicinal product(s) and/or to study absorption, distribution, metabolism and excretion of one or more investigational medicinal product(s) with the object of ascertaining its (their) safety and/or efficacy’5. This directive distinguishes a clinical trial from an NIS as defined above, and thus the directive on clinical trials clearly states in Article 1 (1) that NIS are excluded from the scope of the directive: ‘This Directive does not apply to non-interventional trials’5. As such, at least legally, there is a distinction between a ‘clinical trial’ and an NIS.

*Generic rights are rights of agents as agents, i.e. no agency is possible without the perfect fulfilment of such rights. ‘The generic conditions of agency consist of what vulnerable agents need, irrespective of what their purposes might be, in order to be able to act at all or in order to be able to act with general chances of success’20.

†Based on the American Common Rule, the conditions which all must be met are as follows: (a) the research involves no more than minimal risk; (b) the waiver or alteration will not adversely affect the rights and welfare of the subjects; (c) the research could not practicably be carried out without the waiver or alteration; and (d) whenever appropriate, the subjects will be provided with additional pertinent information after participation22.

*Therapeutic misconception ‘ involves a research participant’s failure to recognise how personal care (i.e., the obligation of physicians to make medical decisions solely with the patient’s interest in mind) may be compromised by research procedures…this core concept of TM could be manifest in two ways: (1) when participants express an incorrect belief that their individualised needs will determine assignment to treatment conditions or lead to modification of the treatment regimen (TM₁); or (2) when participants offer an unreasonable appraisal of the nature or likelihood of medical benefit from participation in the study, due to a misperception of the research enterprise (TM₂)26.