Discussions about criteria for ethics approval of research involving tissue samples have been a feature of literature in a theoretical form for some time. The uniqueness of our study is that we have used detailed case examples and solicited the opinions of REC members about whether and under which conditions the projects should be approved. This debate can only be meaningfully advanced if the discussion takes into account real-world situations, and includes the actors directly involved in the decision making.
For these reasons, a particularly important finding of our study is that there is considerable divergence in the approaches and decisions of Swiss REC members regarding human tissue sample requests. In some cases, only a few members of certain committees express different viewpoints, or raise questions concerning the examples used. However, on a number of issues, there are significant differences in approach. Some differences may be due to cantonal or institutional regulations; while all committees in Switzerland must ensure that the same national and international laws are upheld, many research applications come from universities. These frequently differ with regards to data storage and sharing policies, and may face logistical challenges, such as lack of sample storage space or cooperation with international partners, which present challenging situations for the relevant RECs. More importantly, our results also demonstrate that members of the same cantonal committee hold opposite views. Divergence might be caused by a lack of knowledge, understanding, or clarity of laws and guidelines and would point to a need for more specific training and clarification from the inter-cantonal REC association (AGEK, Swissethics). These being the areas which can and must be improved on in order to allow optimal REC functioning, it is essential to understand their extent. Inconsistency and disagreement in RECs, both internal and between independent committees, can slow research, and can result in varying interpretations of laws and regulations which directly affect research participants [17, 18]. Our findings are an initial indication of areas which require some harmonization efforts.
The most notable finding of our study is that all three case examples, the greatest areas of disagreement concerned informed consent. Below, we discuss the points which produced significant differences, and suggest ways in which this could be minimized.
Questions related to case 1 revealed that a high level of disagreement existed between committee members on whether research involving irreversibly anonymised samples, with no accompanying clinical data, would require the informed consent of the donor. The SAMS guidelines  state, in section 4.3, that no express consent is required for such research. Irreversibly anonymised samples are widely held to pose virtually no risk to donors when used in research, and it is frequently argued that this lessens the need to obtain fully informed consent [19, 20]. Hence, requests for consent by RECs risk slowing the progress of research [12, 21]. It may be that respondents who stated that they would refuse such requests are not frequently confronted by requests pertaining to tissue samples, and apply the criteria that governs research on humans subjects.
While at the time of the survey the Law on Human Subject Research was not in effect, it is worth noting that its current wording may still be perceived as ambiguous by those tasked with upholding it. Informed consent is not required if obtaining it would be impossible or very difficult, or if the ends of the research outweigh the interests of the donor; it may be the case that specific, case-based examples would facilitate the decision-making process for REC members who are tasked with finding the appropriate balance between research and patients interests. Furthermore, the SAMS guidelines emphasise that inappropriate irreversible anonymisation should be avoided: “Both in the interests of the patients and in the interests of research, samples and data should not be irreversibly anonymised, as far as this is possible. For the patient, irreversible anonymisation means that generally he can no longer be informed of relevant results; for research, it means that the samples and the data lose in informative value” (p.6).
Regarding case 2, the most noteworthy result is that almost two-thirds of respondents are convinced that consent is necessary for samples accompanied by clinical data, regardless of whether or not they were irreversibly anonymised. When comparing case 1 to case 2 it seems that the existence of accompanying clinical data is a motivating factor in the need for obtaining consent. Whether committee members feel that this increases the risk (which will depend on the richness of the data and the procedure of anonymisation), or that it increases the chance that useful results are found (a realistic possibility, given the views on return of results) is not clear. As noted above, the SAMS guidelines, and the Declaration of Helsinki , do not make such a requirement. The Law on Human Research now makes clear that this is not always the case, which may contribute to resolving disagreement on this point.
Case 3 gives some of the most detailed explanation of respondents’ decision making, and provides insight into Swiss committee members’ approaches to the divisive field of consent in biobanking. Informed consent procedures typically receive a great deal of focus in ethics review . In Switzerland, informed consent is protected, though not explicitly, by Article 10 of the Swiss constitution  and Article 28 of the Civil Code . As well as being a crucial aspect of every international declaration on health care ethics since the Nuremberg Code, it is defended by Article 1 of the SAMS manual on research on human subjects . Switzerland therefore operates under a legal and ethical framework which recognizes the primacy of informed consent in medical research.
Of the individual consent options (which could, but must not necessarily, form part of a multiple choice form), a majority of respondents opposed the notion of a blanket consent. An even greater majority, however, were in favor of the two more specific options for consent. The more restrictive option (APC gene research) was no more favored than the slightly broader one (colorectal cancer research). Interestingly, those respondents who would accept a multiple choice form would in most cases also accept a more broad consent. These options represent a somewhat enlarged approach to consent in comparison to the traditional human subject requirements, so the correlative acceptance of both ideas is perhaps to be expected.
Our results show that only about half of the respondents approve of these more open approaches to consent, demonstrating that this remains controversial among the participating REC members. According to the Law on Human Subject Research, biological material may be reused in coded or non-coded form if the sample source has consented after having received sufficient information. The exact definition of how general or detailed the information may be in order to be considered sufficient will remain at the discretion of individual RECs. This is likely to be a key factor in the disagreements reported, as it represents an issue free of strict regulation, in which the experiences and perhaps the local context of committee members (such as the research focus on the universities they represent) take on a prominent role in decision making. It may, however, still be desirable for RECs to work towards harmonization on this matter, as any study which requires a multi-centric project to be approved by various committees risks confusion and potentially a stalled or blocked application.
Our results indicate that at present REC members prefer to err on the side of restriction as many respondents have chosen to adhere to the traditional form of consent used in clinical trials. However, there seems to be growing agreement that the specifics of the study protocol can affect the level of consent and the details of information deemed necessary. Clearly, the interpretation of the SAMS guidelines and the Law on Human Research should be clarified in further training for REC members. Indeed, the guidelines do not prohibit blanket consent, as one respondent claimed, but might require more specific consent, if the primary investigator is aware of his planned research and reuse of the samples before he takes the samples. Here again, if the sample is part of a tissue extraction undertaken for treatment purposes, the need for specific consent may not arise (for a full discussion of the intricacies and different scenarios which can occur, see Junod and Elger, (2010) ).
Finally, it should be noted that in case 3, on a number of occasions, members of a single ethics committee give different responses as to how their committee operates. This represents a noteworthy lack of agreement about the very policies and practices of the committee. If not all members of one committee agree about how protocol review should be handled, transparent criteria need to be used to resolve the inconsistencies. Our study highlights the need for committees to work closely, and together with committees from other cantons, in order to ensure a well-argued and transparent level of ethical review.
Disagreement between REC members and different RECs  occurs for three main reasons. Different committees may assess risks and benefits using particular paradigms, or may assign different worth to certain values, for example patient autonomy versus public good . This is an integral part of the ethical review process itself, allowing for consideration of local norms and standards, and detailed analysis of the need for research in that particular time and place. However, as addressed above, disagreement can also be caused by a lack of knowledge or a variable interpretation of relevant laws and guidelines (or indeed, the details of the protocol), or by an absence of such documents. In all of these cases, decisions appear to be influenced factors which have no solid basis in current law, hence differences of this kind should be minimized as far as possible. However, it is important to bear in mind that guidelines and laws as well as their interpretation may change over time, and are also sometimes left deliberately vague. Legislators may have purposely left certain parts of the law open to interpretation to allow for a degree of variation and adaptation to local culture or circumstances. The lack of specificity may, however, have the secondary effect of leaving committee members with insufficient guidance .
Our results indicate that there are some differences in the decisions made by Swiss ethics committees which cannot be explained simply by regional differences. Particularly important are the approaches to requiring informed consent for unidentifiable samples, and the various views on the level of consent necessary when tissue samples are obtained. We have suggested that this may be a result of the fact that traditional paradigms of informed consent are giving way to more relaxed recommendations in the case of tissue samples, and that certain ethics committees have yet to “catch up” with these changes.
It is essential that ethics committees in Switzerland are fully aware of the laws which will affect their assessment of study protocols, and that all members of the committee are kept up to date with new developments. A Switzerland-wide training and information programme for committee members would be one way of working towards this . Consultation with committees on the optimum way to transmit such information would also be a useful step. It has also been suggested that ethics committees could create sub-groups with specialisms in certain types of study .
Our study has several limitations. We have included only RECs from one language region in Switzerland; hence, this study is not representative of all of Switzerland, but merely serves an example of some of the issues encountered one part of country. Due to the higher number of REC members in the subcommittees in Geneva, the opinions of REC members in this canton are overrepresented. On the other hand, this represents to some extent the charge of protocols: RECs in cantons with university hospitals treat a much higher number of protocols than RECs in cantons without a university hospital. Furthermore, our aim has not been to present a quantitative overview about opinions in Switzerland but to explore agreement and disagreement among REC members of the same committees and across some other cantons. The research design, including the qualitative analysis of the comments, has enabled us to show significant controversy. However, the present data does not allow for any understanding of the reasons for disagreement or difference in interpretation between committee and their members, and we recommend this as a topic for future studies. Another limitation is the hypothetical nature of the cases. We did not evaluate how RECs would react to real protocols and answers could be influenced by social desirability. However, we ensured strict anonymity of the answers, and given their sometimes controversial tone, we have no indication for a one-sided social desirability bias.