The moral limits of medical research with children
Westra, A.E.
Citation
Westra, A. E. (2011, June 30). The moral limits of medical research with children. Retrieved from https://hdl.handle.net/1887/17752
Version: Corrected Publisher’s Version
License: Licence agreement concerning inclusion of doctoral thesis in the Institutional Repository of the University of Leiden
Downloaded from: https://hdl.handle.net/1887/17752
Note: To cite this publication please use the final published version (if applicable).
ChApter 9
General discussion
A part of this chapter has been published in Dutch:
Westra AE. Wetenschappelijk onderzoek met kinderen: maak de regels niet te ruim. Nederlands Tijdschrift voor Geneeskunde 2010;154(34):A2275
The Dutch WMO and several other ethical codes and regulations forbid review boards to approve non-therapeutic studies with children when involving more than
‘minimal’ risks and burdens. It has been suggested to liberalize this restriction, in order to enable researchers to conduct more of the studies that are necessary for improving medical care for future sick children. However, the restriction on non- therapeutic research with children was established for a good reason: it ensures that individual children will not be exposed to (relatively) high levels of risk and burden solely for research purposes. In this thesis, we therefore have examined the following question: ‘How to facilitate more of those studies that may improve medical care for children as a group, while still adequately protecting individual research subjects against research risks and burdens?’
An analysis of the approval/rejection decisions made by the Dutch Central Committee on Research with Human Subjects (CCMO) revealed that in practice, the requirement of minimal risk and burden poses three issues: 1) It is not always clear whether a study should be regarded as a non-therapeutic study and thus should comply with the requirement of minimal risk and burden; 2) Without clear definitions of ‘minimal risk’ and ‘minimal burden’, this requirement may not provide a reliable level of protection; and 3) Some data that are necessary for improving medical care for children as a group cannot be obtained if review boards are not allowed to make exceptions to the requirement of minimal risk and burden.
Below, we will discuss the three issues one by one. First, we will briefly explain the issue at stake, summarize the chapter(s) of this thesis addressing the issue and formulate a recommendation. Second, we will discuss alternative solutions that were not yet discussed in the relevant chapter(s). Of particular importance in this respect are the recommendations of a Dutch expert committee that was established by the Minister of Justice and the State Secretary of Health, Welfare and Sport in response to the CCMO’s word of warning, and which is named after its chairman: prof. J. E. Doek.3;52 Third, we will indicate which further research could strengthen our recommendation. In this general discussion we will also suggest which parts of the WMO we think should be modified to enable the implementation of our recommendations in the Netherlands, we will explore the implications of our recommendations, and we will discuss the general limitations of the thesis.
ISSUE 1: IT IS NOT ALWAyS CLEAR WHETHER A STUDy SHOULD BE CLASSIFIED AS NON-THERAPEUTIC
In the next three paragraphs we will discuss the three issues that we found to be related to the requirement of minimal risk and burden. We will start with the issue that it is not always clear whether a study should be classified as non-therapeutic and thus should comply with the requirement of minimal risk and burden. During our analysis of the approval/rejection decisions made by the CCMO we noticed that this classification-problem is not only inconvenient to review boards and
researchers, but can also have serious consequences: in case the review board, in cases of doubt, decides to apply the strict rules for non-therapeutic research, studies may be unnecessarily rejected; however, if the review board decides not to do so, research subjects may be insufficiently protected.
Solution proposed in this thesis
Chapter 3 revealed that the issue that is not always clear whether a study should be classified as being non-therapeutic, cannot simply be solved by providing a better definition of the term ‘therapeutic’. The problem is more deeply rooted:
many studies, and drug studies in particular, cannot be classified as being entirely therapeutic or non-therapeutic, because they are too complex. Some early phase drug studies (which are currently regarded as non-therapeutic drug studies) may offer a reasonable treatment option to the research subjects because there are no alternative treatment options available, and many so-called therapeutic studies include non-therapeutic procedures such as extra blood drawing, extra scans, or extra hospital visits. In addition, therapeutic procedures occasionally involve research risks and burdens that are faced solely for research purposes, because an alternative treatment option with a more favorable risk-benefit profile is already available.
Thus, to adequately identify those research risks and burdens that the subjects have to face solely for research purposes (and that therefore in principle should be minimal), we recommend distinguishing between therapeutic and non-therapeutic procedures (or clusters of procedures) rather than between therapeutic and non-therapeutic studies, and we recommend considering the therapeutic procedures in relation to the alternative treatment options for the subjects. Procedures should be classified as therapeutic in case they are administered with therapeutic (or diagnostic, or preventive) warrant and on the basis of evidence that justifies the belief that they may benefit the subjects.
We realize that this approach, which is best known by the name ‘component analysis’, may be time consuming for often already heavily overloaded review boards.13 However, it will mainly be more time consuming in complex cases, and we believe those complex cases deserve this time.
Alternative solution
The Doek Committee was established to address the issue that some of the data that are necessary for improving medical care for children as a group cannot be obtained with minimal risks and burdens. However, the Doek committee has also addressed the issue that the distinction between therapeutic and non-therapeutic studies is unclear. The committee’s solution is to stop using this distinction. The classification of medical research studies as either therapeutic or non-therapeutic, the committee argues, does not do justice to the possibility of future medical benefits to the subjects, the possibility of non-medical benefits (such as hope), or
the possibility of benefits to future patients. Abandoning the distinction between therapeutic and non-therapeutic studies would enable review boards to take such other potential benefits into consideration when making their risk-benefit assessments.52
We believe that following this proposal would not result in an improvement over the current situation, as it seems that review boards would become less rather than more likely to identify all risks and burdens that the subjects have to face solely for research purposes and that therefore, in principle, should be minimal. As we have argued in Chapter 3, the above-mentioned benefits all have a different moral status than direct benefits to the subjects and, thus, cannot in a similar way compensate for the research risks and burdens. The trouble of including potential future medical benefits to the subjects in risk-benefit assessments is that these are inherently uncertain: usually, the chance that the subjects may medically benefit from the results in the future is too low to compensate for the research risks and burdens. Hope, on the other hand, certainly may be a considerable benefit of research participation. However, in case a procedure is provided with therapeutic warrant, the requirement of minimal risk and burden is not at stake, and in case the procedure is not provided with therapeutic warrant, feelings of hope should not be stimulated. Last, regarding the possibility of benefit for future patients: such so-called ‘group-directedness’ (as we explained in Chapter 1) is important and should, we believe, always be present: after all, the objective of medical research with children is to improve medical care for children as a group.
However, it is the subjects rather than those future patients who run the risks and burdens of the study. Thus, (direct) benefits to the subjects and benefits to future patients should not be treated as interchangeable.
Suggestions for further research
We believe that additional research could strengthen our recommendation to implement component analysis. Several questions need to be addressed. For example, how to reach an overall assessment of a study on the basis of a component analysis? In addition, it would be helpful to gain more insight into the magnitude of the problem to be solved. How often do therapeutic studies include non- therapeutic components, and vice versa, how often do non-therapeutic studies include components that may benefit the subjects? How much risks and burdens are attached to these components? These questions could be addressed by a systematic analysis of all proposals for non-therapeutic (drug) research currently present in the archives of the CCMO and a random sample of all proposals for therapeutic (drug) research.
ISSUE 2: MINIMAL RISk AND MINIMAL BURDEN SHOULD BE MORE CLEARLy DEFINED
The second issue that we found to be related to the requirement of minimal risk and burden was that in the absence of clear definitions of the concepts of ‘minimal risk’ and ‘minimal burden’, it is hard for review boards to assess whether a study fulfils the requirement. Our analysis of the decisions of the CCMO revealed that in particular the complex concept of ‘minimal risk’ could benefit from a clear definition explicating that the magnitudes and likelihoods of all potential harms need to be considered. The CCMO managed to interpret the concept of ‘minimal burden’ very consistently; yet also a clear definition of this concept could be helpful, because: 1) the internal consistency we found does not guarantee that other review boards interpret the concept in the same way; and 2) in the absence of a clear definition, review boards have the opportunity to occasionally ‘stretch’
the meaning of the concept.
Solution proposed in this this thesis
In Chapter 4, we have examined the problems related to the definition of minimal risk provided by the US Federal Regulations. This revealed that it would not be recommendable to follow the US definition of minimal risk. In the second part of Chapter 4, we have developed new definitions. The definition of minimal risk is that
‘empirical data and/or expert opinions suggest that in the persons concerned, the likelihoods that the procedure(s) will cause small, moderate, and/or serious harm are < 1/100, < 1/10,000, and < 1/1,000,000, respectively’. The definition of minimal burden (which we in Chapter 4 call ‘minimal risk of discomfort’) is that ‘empirical data, expert opinions and/or the procedural characteristics suggest that at most a quarter of the persons concerned will experience considerable discomfort’. We recommend to implement these definitions and to make the decision of whether a procedure fulfils these requirements in each case as evidence-based as possible.
To provide insight into the meaning of the concept of ‘minimal burden’, Chapters 5 and 6 report on studies assessing the discomforts in children participating in a vaccine study and in children undergoing an MRI scan, respectively.
Alternative solution 1
The explanatory memorandum to the WMO provides definitions of the concepts of ‘negligible risk’ and ‘minimal drawbacks’. It states that ‘negligible risk’ means that ‘the risks may not be more than the risks of daily life or than the risks that have, in daily medical practice, appeared to be negligible.’147 ‘Minimal drawbacks’
means that ‘all together, the research procedures should not have a big impact on the person involved. The disturbance of daily life should be minimal and the pain caused by the study should not be more than the pain caused by, for example, a blood draw’.147
Perhaps, the main reason for the CCMO not (explicitly) using these definitions is that they concern the terms ‘negligible risk’ and ‘minimal drawbacks’, whereas during the review discussions the CCMO often uses the internationally accepted terminology (e.g., ‘minimal’ in stead of ‘negligible’ risk). To solve the definition issue, the CCMO and the other Dutch review boards (if not already doing so) could also opt for implementing the WMO terminology and the corresponding definitions in their practices, instead of adopting the definitions that were proposed in this thesis. However, considering the fact that the terms ‘negligible risk’ and ‘minimal drawbacks’ represent uncommon terminology, we do not think this alternative is to be preferred over the solution proposed in this thesis: in this time of research-globalization, we believe terminological confusion should be avoided where possible. Moreover, the explanatory memorandum’s definition of negligible risk partly resembles the US definition of minimal risk (i.e., as far as it concerns the risks of ‘daily life’), and thus partly involves the same problems as we discussed in Chapter 4.
Alternative solution 2
A second alternative option would be to follow the definitions of minimal risk and minimal burden that can be found in the Additional Protocol to the European Convention on Human Rights and Biomedicine (European Convention).
According to this Additional Protocol, a medical research study entails minimal risk if it ‘will result, at the most, in a very light and temporary negative impact on the health of the person concerned’, and ‘minimal burden’ ‘if it is to be expected that the discomfort will be, at the most, temporary and very slight for the person concerned.’33 However, following these definitions is not advisable, as this would block almost all research: the chance that one of the subjects will experience serious or irreversible harm, and the chance that one of the subjects will experience a level of discomfort that is higher than ‘very slight’, can almost never be ruled out entirely.
Suggestions for further research
In Chapter 4, we have recommended implementing new definitions of ‘minimal risk’ and of ‘minimal burden’, and we have recommended making the decision of whether a procedure fulfils these requirements in each case as evidence- based as possible. We believe that additional research could strengthen this recommendation. Most importantly, to enable review boards to make their decisions as evidence-based as possible, systematic overviews of the risks posed by common (and less common) research procedures should be made available, just as systematic overviews of the discomforts that these procedures may cause in various groups of children. To obtain these data, extensive literature research and empirical research will be required. We are undecided what the best method for measuring discomfort will be: considering the subjective character of the concept
of discomfort, self report (from children aged 5 and above) seems most valuable;
however, adequate questionnaires for children in the various age groups would need to be developed and standardized. Perhaps, before doing so, a conceptual analysis of the concept of discomfort could be helpful: what does it mean, for example, if children score high on only one or two out of various measures of discomfort, as was the case in our MRI study (Chapter 6)?
ISSUE 3: SOME OF THE NECESSARy DATA CANNOT BE OBTAINED WITH MINIMAL RISkS AND BURDENS
The third issue related to the requirement of minimal risk and burden was that some data that are necessary for improving medical care for children as a group cannot be obtained if exceptions to the requirement of minimal risk and burden are not allowed. Our analysis of the approval/rejection decisions made by the CCMO (Chapter 2) revealed that review boards may find ways to approve important studies that formally should be rejected, but we think that this is not an advisable solution. It may be that exceptions to the requirement of minimal risk and burden occasionally can be justified; yet, we believe, when such exceptions are made, this must be done explicitly and the reasons for doing so must be transparent and reviewable. It is relevant to note that it can be expected that even if all review boards started using component analysis as well as our proposed definitions of ‘minimal risk’ and ‘minimal burden’, some of the necessary non- therapeutic procedures still would be found to have more than minimal risks and/
or burdens.
Solution proposed in this thesis
In Chapter 7 we have shown that issue 3 cannot be solved by adapting the WMO to the European ethical framework, because in Europe there is currently no consensus. In Chapter 8 we have analyzed whether following the US Federal Regulations, which allow in certain circumstances for exceptions to the requirement of minimal risk and burden, would be advisable. To do so, we have first explored the ethical grounds for accepting only ‘minimal’ risks and burdens. From there we have deduced that making an exception to the requirement of minimal risk and burden can be regarded as acceptable in case the study is exceptionally valuable and the potential subjects are able to give their assent to participate. We have concluded that rather than following the US Federal Regulations, the Dutch and European regulatory frameworks should be revised to make exceptions in such specific cases legitimate and transparent. Thus, we recommend allowing exceptions to the requirement of minimal risk and burden in case of exceptionally valuable studies involving children who can give their assent. Studies can be regarded as ‘exceptionally valuable’, we argue, if the desired data are truly indispensable for improving medical care for children as a group (which is the
case, we believe, if insufficient treatment, diagnostic, or preventive options are available for the disease at stake, and if the desired data are likely to make a significant contribution towards developing a new option) and if the study design is ideal, as compared to all other possible designs (which is the case, we believe, if the methodology is of excellent quality, and if the subjects do not have to face more risks and burdens solely for research purposes than strictly required to obtain the desired data). Children can be considered able to give their assent to participate in the study if they are generally capable of understanding the purpose of the study, of understanding the risks and burdens they have to face solely for research purposes, and of making their own decisions based on this knowledge.
Alternative solution
The Doek Committee’s recommendation regarding the issue that some of the necessary data cannot be obtained with minimal risks and burdens is to abandon the requirement of minimal risk and burden for all intervention studies (i.e., studies that alter the condition of the subjects to evaluate the effects of that intervention) and for all non-intervention studies with children aged 12 and above. Instead, the committee proposes that the WMO should be modified to include the requirement that the research risks and burdens should be ‘minimized’ (i.e., should be as low as possible).52 The committee seems to base this far-reaching proposal on three arguments: 1) If the WMO were modified to include the requirement of minimization of risks and burdens, the levels of risk and burden would remain acceptable; 2) When abandoning the requirement of minimal risk and burden, children would get more opportunities to act altruistically; and 3) When following these recommendations, the Netherlands would follow Europe.52
We believe that all of these arguments are problematic. Regarding the first argument, we agree with the Doek Committee’s observation that the WMO should mention more clearly that risks and burdens should always be minimized.
However, for two reasons we do not believe that such a requirement would justify abandoning the requirement of research risks and burdens being minimal in the absolute sense of the word. The first reason is related to the fact that being
‘minimized’ does not equal being ‘minimal’ in the absolute sense of the word.
Risks and burdens that are as low as possible, can still be substantial. Thus, when abandoning the requirement of minimal risk and burden, one may eventually lose sight of the idea that exposing children to more than minimal risks and burdens solely for research purposes must always be regarded as a last resort. The second reason is that the absolute requirement of minimal risk and burden stimulates the creativity of the researcher to minimize risks and burdens: in our experience, when faced with the absolute threshold, risks and burdens often could be minimized a bit more.
The second argument of the Doek Committee was that children should get more opportunities to act altruistically. When maintaining the requirement of minimal risk and burden, the committee argues, there is little room for children’s own views and values. We believe that this consideration is interesting, but has insufficiently been developed. For example, the report lacks explanation of why this argument would also hold for very young children.
Finally, regarding the argument of ‘following Europe’, we believe that the Doek Committee has focused too narrowly on the Clinical Trials Directive.19 As we have shown in Chapters 1 and 7, there currently is no clear European consensus regarding the protection of children in non-therapeutic research, which implies that ‘following Europe’ cannot be a convincing argument. In the European Convention, which may soon be ratified by the Netherlands, the requirement of minimal risk and burden plays a prominent role; and a supplementary document to the Clinical Trials Directive also provides an upper limit of acceptable risk (albeit a minor increase over minimal risk).32;34 Recently, prominent members of the Dutch Health Law Association have also pointed out the current European lack of consensus regarding the protection of children in non-therapeutic research.148
Suggestions for further research
Instead of following the recommendation of the Doek Committee, we recommend (as was argued in Chapter 8 and was summarized above) to uphold the requirement of minimal risk and burden, and to allow for exceptions to this requirement in case of exceptionally valuable studies involving children who can give their assent to participate. Considering that our recommendation resulted from a theoretical analysis, further research could help operationalizing the concepts at stake. In addition, we recommend that review boards develop and publish (for example, on a website) written rationales to explain the basis for their judgments regarding potential exceptions to the requirement of minimal risk and burden. Such case decisions could be publicly discussed, could be systematically analyzed, and could provide useful examples.
THE WMO
To make the implementation of our recommendations possible in the Netherlands, several parts of the WMO should be modified. We recommend the following modifications:
1. We recommend basing Article 4 on a distinction between therapeutic and non-therapeutic research procedures, and to include the following elements:
a. The risks and burdens of all procedures must be as low as possible.
b. Therapeutic procedures must fulfill the requirement of clinical equipoise.
c. Non-therapeutic procedures, in principle, must fulfill the requirement of
minimal risk and burden. Exceptions to this requirement may be allowed in case of exceptionally valuable studies involving children who can give their assent to participate in the study at stake.
d. Each study must fulfill the requirement of group relatedness.
2. When replacing the distinction between therapeutic and non-therapeutic studies by a distinction between therapeutic and non-therapeutic proce- dures, Article 2.2.b.2 would no longer be of use. Instead, we recommend that all potential exceptions to the requirement of minimal risk and burden should be reviewed by the CCMO.
3. We recommend altering Article 6 as far as it concerns the terminology used and the cut-off age of 12 years, taking into account that children do not need to meet all requirements for informed consent (i.e., the ‘schriftelijke toestem- ming’ of a competent adult) in order to be considered capable of giving assent (i.e., the ‘schriftelijke toestemming’ of a child, which is required in ad- dition to the ‘schriftelijke toestemming’ of the parents).
IMPLICATIONS OF OUR RECOMMENDATIONS
To explore the implications of our recommendations we have translated them into two checklists with items to be considered by (researchers and) review boards, when reviewing a pediatric research proposal (See Appendix 3). Here, we will consider these lists in relation to the four example studies that can be found in Appendix 1. In case of example studies 2 and 4, we will focus on items 3 (does the study involve a therapeutic procedure?) and 4a (does clinical equipoise apply?) of Checklist 1. In case of example studies 1 and 3, which both are entirely non- therapeutic studies, we will assume that the risks and/or burdens are not minimal, and will focus on the question of whether making an exception to the requirement of minimal risk and burden can be regarded as acceptable (Checklist 2).
Example study 2 (Tolerability, safety, PK and efficacy of a new chemotherapeutic agent)
If participating in this chemotherapy study, children would be allowed to continue the study medication until progression of their disease. This aspect of the study, in combination with the lack of other treatment options available for these children, suggests that the study drug is provided with therapeutic warrant. Thus, regarding this particular procedure, the question of whether clinical equipoise applies is at stake. Following Van Agt, we suggest that such a question in cases of patients for whom no alternative therapy is available can perhaps best be answered in relation to the question of whether it would be medically defensible to administer the experimental drug at stake in a non-research setting, i.e., whether it would be medically justifiable to administer the experimental drug as a last resort.149
We would like to add that it is also important to assess the extent to which the administration of the study drug may be attuned to individual needs. In particular the dosing schedule of the drug is important in this respect: within a standard dose escalation schedule, the first subjects to be included are likely to receive a subtherapeutic dose, which means that for these subjects, administering the study-drug does not seem a defensible treatment option.
Example study 4 (Safety, PK and PD of inhaled growth hormone as compared with injections)
This is the study that was, upon reconsideration, regarded as a therapeutic study. When analyzing the various procedures, however, most appear to be non- therapeutic. To start with, the admissions to hospital, the IV cannulas, the blood draws, the lung function tests and the washout-phases all are non-therapeutic procedures. Moreover, also some of the drug-related procedures are non- therapeutic. It is important to note in this respect that that: 1) the injection of growth hormone for this group of subjects is ‘standard care’, but within this study design, the children will not receive their usual doses; 2) although inhaling something may be more comfortable than getting an injection, this potential benefit is not present during this particular study: during the study period, all subjects have to inhale growth hormone or a placebo in addition to their regular injections; and 3) the placebo-component in this case is not provided with therapeutic warrant, but to more accurately assess the safety and tolerability of growth hormone inhalation (as far as it concerns the subjective parameters).
Thus,
Week 1: Injection of the drug = therapeutic procedure Inhalation of placebo = non-therapeutic procedure Week 2: Inhalation of the drug = therapeutic procedure
Injection of placebo = non-therapeutic procedure.
Strictly speaking, the first therapeutic procedure (injection of the drug as usual, but with modified dosing schedules) seems unable to fulfill the requirement of clinical equipoise: there does not seem to be a state of honest professional disagreement as to whether the modified dosing schedules that are part of the study design, are better than the subjects’ usual dosing schedules. However, we suggest that clinical equipoise in a case like this may be understood as a state of honest professional disagreement as to whether such dose modifications, for a period of only one week out of many years of growth hormone therapy, will have any clinically relevant effects. Regarding the second therapeutic procedure (inhalation of the drug), then, the question to be answered is: does a state of honest professional disagreement exist as to whether inhalation of the drug is to be preferred over injecting the drug? Inhaling the drug probably is more comfortable. However, in this stage inhaling the drug also involves more risks: the risks of subtherapeutic dosing, and
the risk of bronchial adverse effects. We believe this is a clinical judgment that would require consultation of clinical experts.
Example studies 1 (New antiretroviral drug) and 3 (Duchenne proof of concept study)
As far as we can judge, both of these studies are indispensable for improving medical care, for HIV-infected children who are resistant to the available NNRTIs and for children with Duchenne muscular dystrophy, respectively. Assuming that both studies are excellent from a methodological point of view, and assuming that the children to be included can be considered capable of giving their assent to participate in the studies at stake, the most interesting question in these cases is: do the subjects have to face no more risks and burdens solely for research purposes than strictly required to obtain the desired data? Would it be possible to obtain these data with lower levels of risks or burden, and/or while offering direct benefits to the subjects? In case of study 1, perhaps, there is some room for improvement, for example: awaiting more safety data in adults, combining the research visits with children’s regular visits to hospital, and/or providing the drug with therapeutic warrant (to other types of patients). In case of study 3, however, this does not seem to be the case.
Conclusion
It can be expected that when following our recommendations review boards would be able to approve studies that currently have to be rejected, because:
1) many early phase drug studies would in part rather than in its entirety need to comply with the requirement of minimal risk and burden (e.g., study 2); and 2) in specific cases, exceptions to this requirement of minimal risk and burden would be permitted (e.g., in case of study 3). This increased potential for approving studies does not seem to compromise the protection of the individual research subjects. In fact, considering our recommendation to more accurately identify those risks and burdens that have to be minimal, the subjects seem to be more adequately protected than they currently are (e.g., in case of study 4).
THE LIMITATIONS OF THIS THESIS
Three general limitations of this thesis deserve consideration. The first concerns the fact that this is an ethical and not a legal thesis. We have focused on the legally codified requirement of minimal risk and burden for non-therapeutic research with children, but aimed to analyze the ethical underpinnings of this requirement and to improve practice, rather than to analyze the law. As a consequence, the thesis does not result in specific recommendations on how to adapt the law. It is also important to note in this respect that during ethical analyses, every now and then
(informed and motivated) choices and interpretations need to be made, which may have influenced our recommendations.
The second limitation concerns the scope of our inquiry. As we have focused on the requirement of minimal risk and burden, other issues, however relevant when aiming to adequately protect research subjects (e.g., the other ethical requirements for non-therapeutic research, the modified informed consent requirements for research with children, and the informed consent procedure in practice), were not addressed. Moreover, we have not specifically addressed the moral limits of research involving incompetent adults, such as people suffering from dementia. However, we believe this would be an interesting and very important topic for further research, as research in incompetent adults involves equally complex issues. Perhaps, although of course there are morally relevant differences, (parts of) our recommendations could also be applied to this other category of incompetent research subjects.
A last consideration concerns the three empirical studies (Chapters 2, 5 and 6).
As there has been a lot of debate about the question of whether empirical studies can inform ethical reasoning, some people may consider our use of empirical studies a limitation.150 However, ethics being one of the most practical branches of philosophy, we believe that information about the real world is indispensable in ethical work. We agree that the mere fact that many people believe that something is morally right, and/or the fact that it always has been done, does not make something morally right.151 Yet we believe that empirical studies can inform ethical reasoning in several ways. For example, empirical studies can identify and/
or specify ethical problems, as we showed in Chapter 2.152 They can also describe facts that are relevant to normative judgments, as we did in Chapters 5 and 6.151
CONCLUSION
In this thesis we have developed recommendations related to the question ‘How to facilitate more of those studies that may improve medical care for children as a group, while still adequately protecting individual research subjects against research risks and burdens’. These recommendations are:
To more accurately identify those research risks and burdens that have to be minimal, by distinguishing between therapeutic and non-therapeutic procedures instead of between therapeutic and non-therapeutic studies, and by considering the therapeutic procedures in relation to the alternative treatment options for the subjects.
To define minimal risk as: ‘empirical data and/or expert opinions suggest that in the persons concerned, the likelihoods that the procedure(s) will cause small, moderate, and/or serious harm are < 1/100, < 1/10,000, and < 1/1,000,000, respectively’, and to define minimal burden as: ‘empirical data, expert opinions and/or the procedural characteristics suggest that at most a quarter of the persons
concerned will experience considerable discomfort’. We recommend making the decision of whether a procedure fulfils these requirements in each case as evidence-based as possible.
To allow exceptions to the requirement of minimal risk and burden in case of exceptionally valuable studies involving children who can give their assent to participate in the study.
We expect that by following our recommendations, review boards would be able to approve studies that currently have to be rejected, because: 1) many early phase drug studies would in part rather than in its entirety need to comply with the requirement of minimal risk and burden; and 2) in specific cases, exceptions to this requirement of minimal risk and burden would be permitted. This increased potential for approving studies, we believe, does not compromise the protection of individual research subjects, because we recommend to refine rather than to liberalize the rules, resulting in tailor-made protection. In fact, considering our recommendation to more accurately identify those risks and burdens that have to be minimal, and considering our proposed definitions of ‘minimal risk’ and
‘minimal burden’, the research subjects would be more adequately protected than they currently are.
