Another issue that we would like to cover in this segment is data sharing and informed consent. And one aspect of informed consent that may not be clear to individuals, particularly working in developed world countries, is that there may be unexpected difficulties in obtaining informed consent. Which of course is relevant in turn in sharing data from cultures where parts of the population for example cannot give consent. In particular, women without the presence of male relatives in some cultures, or in situations where the concept of the research to be undertaken is unclear. For example, in situations where the consent for genomic research is sought but words like genes or genome do not exist in the language. And this again is from the Neufield study, that we sited earlier So one area facilitating collaboration is the so-called data use agreements, also known as data transfer or data access agreements. We will use the term here data use agreement or the abbreviation DUAs. A data use agreement is helpful. And it must be in place in all cases where data are transferred from one party to another. The reason is that data use agreements help clarify and manage expectations by both parties particularly in the case of adverse to the developing world collaboration. They are, and must be seen, as part of the spectrum of conducting ethically responsible research. And this is, again, particularly relevent in low and middle income countries. Also DUAs have to be seen as an integral part of the informed consent process. And as a part of the spectrum of the informed consent which itself is an integral part of all research involving persons, in other words all clinical research. So the DUAs can not be viewed in isolation from the ethical responsibilities inherent in the responsible conduct of clinical research. Now another area where DUAs can be helpful is in describing ownership of the data to be shared among or between the parties participating in the collaboration. They are useful because they help alleviate anxiety by clearly articulating who owns the data. It must be clear within the DUA, for example, that the data are owned by the groups which produce them, and ideally the data shared should be shared for a single, well-identified and described project. If there are multiple projects, then multiple data use agreements may be, or must be in place. And data should absolutely not be used for projects for which explicit written approval has not been obtained. And I cannot understate, I cannot, rather overstate this fact enough that data should be used only for projects and should not be shared and should be used only for projects where written approval has been obtained. Now the final part of the segment is describing what some of the elements of DUAs should be. And at a minimum the DUA should have a description to which the agreement is attached. A clear description of the scope of the data use. A clear statement of the data ownership before, but as well after the completion of the project. And the description of the rights and obligations of the parties entering in the agreement. So let's speak briefly about the description of the study that should be in the data use agreement. The study that is carried out must be described briefly but clearly. Any issues with ethical approval may be cross-referenced within the DUA, but that's not always necessary, other elements also can be present, such as an outline of the role of the investigators in the two groups sharing the data. And the potential benefits to the group sharing the data from the study. And this is particularly relevant to situations in LMIC. The scope of the data sharing is also a critical part of the data use agreement, and aspects of that include whether the data are for single or multiple projects, ideally the latter should involve separate agreements. A clear statement about whether the data can be shared with anyone else. And again ideally data should not be shared with others in the absence of additional data use agreement. And the procedures for termination of the data use agreement by either party. Data ownership is obviously a critical issue as well. And the parts in the data use agreement, speaking to data ownership must be a clear statement about who owns the data, and usually this is the group or agency transmitting the data and not the group receiving the data. And also a clear description of what happens to the data after the study has been completed. For example, the Collaboration of Observational HIV Epidemiological cohorts in Europe, or COHERE, has a policy where data are to be destroyed a certain period after the publication of the study resulting from use of this shared data, that also has to be clearly articulated within the data use agreement. Other parts that can also be included in the data use agreement include the description of procedures for review of results from the studies and undertaking with the transmitted data, including administration review by government agencies, if applicable. And this is very important because in many situations data had been produced through the sponsorship, or support, or the results of government agencies, or the results of the studies from these are relevant to the policy of government agencies and other stakeholders. And these have to be included in the review of the results and also authorship agreements for manuscripts and presentations resulting to analysis which is particularly important in LMIC, which may have less of a tradition with authoring scientific publications. And this is an area where a collaboration with a developed world group may be of help. In conclusion, data sharing is complex and both whether the parties are from developing or from developed countries and cultural economic factors and the resulting real and perceived inequities and conflicts of interest must always be considered when establishing collaborations particularly between developing and developed countries. The DUAs or data use agreement are a major part of collaborative research, because they manage, clarify expectations, develop and foster a trust between all parties participating in the research. And they must be viewed within the broader context of responsible conduct of clinical research.
