Study design and sampling

We employed a qualitative exploratory study design using semi-structured interviews to explore REC members’ views on data sharing, overall experience with reviewing data-intense protocols that involved data sharing as well as their commonly faced challenges. We also explored their views on DTAs and DACs and the role that RECs play within these structures.

Data was collected from REC members from different countries in SSA–from the ground up to build concepts and theories. A grounded theory approach was used [22]. Hence, semi-structured interviews were conducted until theoretical saturation of data was reached. Flexibility, creativity and reflexivity were used in the interview process. The interviewer and interviewees were active in the construction of knowledge. There was prompting and probing during the interviews to better understand the challenges faced by REC members and how these challenges could be resolved. Flexibility was important as interviews were being conducted in different countries in SSA where data transfer laws, ethics guidelines and competence of REC members differed.

This study was conducted by a multi-disciplinary team of researchers. NC holds a Master’s degree in bioethics and has a background in Public health. SK holds a Master’s degree in Public Health and has a background in health sciences. KM is the study PI, a family physician with expertise in medical ethics, clinical research and research ethics. She has served on RECs for more than 10 years and holds a doctorate.

Our sample included representatives from 20 different SSA countries through a purposive selection of professional networks of the Stellenbosch University’s Division of Medical Ethics and Law across SSA. A snowballing technique was used for further recruitment. We also identified potential participants through a desktop search based on their profession and directly emailed REC members, inviting them to participate in their personal capacity. This study is linked to a previous descriptive cross-sectional online survey in which we invited interviewees to anonymously participate in an online survey through Research Electronic Data Capture (REDCap). At the end of the questionnaire, interviewees were asked to indicate their interest in participating in an in-depth interview. Of those who agreed and provided their email addresses (through anonymised branching logic which directed them to a Google Form), written informed consent forms were sent directly to them, signed and returned via email before the scheduled in-depth interviews. During the quantitative survey, some participants volunteered to participate in in-depth interviews for this study. Furthermore, some participants declined to participate due to time constraints.

Data collection

From 02 June 2022 to 21 February 2023, we conducted 20 semi-structured, in-depth interviews with REC members via Microsoft Teams. All interviews were conducted virtually by NC, within a private space and without the presence of a third party. All interviews were conducted online, in English and lasted an average of 45 minutes. Verbal permission was obtained to record each interview via Microsoft Teams. Interview questions were formulated and adapted from Ferretti et al [12] and through discussion with members of the team. The interview guide was pre-tested with five REC members to assess readability, ensure clarity and refine the content before study commencement. Data was collected until saturation at a continental level was achieved given that most anglophone SSA countries had similar ethics challenges with respect to data sharing.

Data analysis

Researchers transcribed verbatim the audio files and cleaned the data. The data was managed using Atlas.ti (version 22) software [23]. Within this software, the data was subjected to thematic analysis, employing an inductive approach. The data was coded and then grouped into subthemes which led to the generation of major themes. Adjustments to the final thematic map were made to improve logical cohesion through extensive deliberation among the team of involved members. The findings of this analysis are detailed in the results section. To achieve rigour in our findings and promote transparency of the coding process, intercoder reliability (ICR) was established through two researchers independently coding 20% of randomly selected transcripts followed by deliberations to reach a consensus on codes, generation of themes and any discrepancies. O’Connor and Joffe [24] recommend that 10–25% of data units should be coded to establish ICR.

Ethical aspects

Research integrity was maintained throughout the study and participation remained entirely voluntary. We approached our sample in their individual capacity and they consented in their own personal capacity. The in-depth interviews were of minimal risk and ethics approval was granted by the Faculty of Medicine and Health Sciences Health REC (Reference No: N22/03/028) at Stellenbosch University, South Africa.

Demographic information

A total of 20 REC members from across SSA were interviewed and represented 20 of the 49 SSA countries (Table 2).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Demographic characteristics of participants (n  =  20).

https://doi.org/10.1371/journal.pone.0303828.t002

Thirteen of the interviewees served in the institutional REC, while 7 served in the national REC. Eight of the 20 interviewees were female and the remaining 12 were male. The majority were between the ages of 40–49.

Frequency of and approaches used in reviewing data sharing protocols

Although a few interviewees indicated that they had not reviewed data-intensive research protocols, most reported having previously reviewed such protocols.

“…most of our studies are multi-country and involve collaborative research so, I think…about 50 per cent of protocols [that we reviewed] involve data sharing.” [Country 5]

Approaches used in reviewing data sharing protocols

Interviewees revealed that reviewing data sharing protocols involved assessing data sharing agreements, procedures and mechanisms put in place to regulate data transfer. Some respondents were able to discuss the overall reviewing of data sharing protocols while others were able to elaborate more on data sharing aspects in research protocols. This varied from country to country but mostly, data sharing aspects in research protocols were discussed. In data-intensive research countries, interviewees could discuss data sharing in greater depth. In countries where protocols are not yet related to big data, discussions revolved around data sharing aspects of regular research.

“… for the protection of the participant, we look out for the Material Transfer Agreement (MTA), the agreement between the two institutions, what has been stated in it and whether it coincides with the main proposal?” [Country 9]

“…during our review of protocols, we check what measures researchers are bringing into action to avoid any risk associated with data sharing…” [Country 5]

Practical/technical challenges faced in reviewing data sharing protocols

Interviewees reported several challenges encountered while reviewing data sharing protocols such as data access requests without ethical approval, data management issues, incompleteness of transfer agreements and the inadequacy of the contents of consent forms for data-intensive research. The concerns raised by the interviewees regarding the management of national data were primarily centred around the storage of data at a national level. In the context of conducting multi-country studies, REC members may encounter difficulties regarding the capacity off or data management in each country involved.

“before asking for data, ensure you have institutional review body clearance to use the data ethically and adhere to ethical principles… one of the biggest issues is that researchers want to access the data before even developing the protocol…” [Country 12]

“…the multi-country studies are challenging around data storage and data analysis, especially around confidence in the country’s data management capacity…” [Country 14]

Need for training

Interviewees emphasised the importance of REC members receiving training to review data sharing protocols. This is because some of the interviewees felt that their committees lacked the necessary expertise to evaluate such protocols. Some of the interviewees suggested that a structure be put in place to ensure that REC members have minimal, obligatory training before they are appointed.

“…I propose more training, more exposure and awareness of the ethical challenges and emerging issues of data sharing so that they can get more familiar with the issues and know how to address them when they are faced with such.” [Country 9]

DTA as an individual document or combined with MTA

We asked interviewees to expand on their idea of the ideal structure for data transfer documents. Participants favoured different ideas as reported below:

“I think it’s better to have them [MTA and DTA] combined because if the material that we are referring to actually gave birth to the data that is being asked for, it would be reasonable to have them together as one generated the other… .” [Country 13]

Another interviewee shared a similar view:

[MTAs and DTAs] should be combined as material and its associated data are linked. There are times when people ask for purely data transfers and there are also times when it is purely material transfers that are done…” [Country 1]

On the other hand, some interviewees reasoned that because these documents cover different matters, they should not be collated or merged.

“I think it would be good for them to be separate because they’re dealing with different things. If they were combined, then we might lose out on something.” [Country 7]

“DTAs and MTAs deal with two separate issues and should remain as two separate documents… combining them would reduce its rigour. People may cross over one aspect and focus on another.” [Country 15]

Key elements of data transfer tools

Interviewees further explained what they thought were some of the key characteristics that a data transfer agreement needed to comprise. Several ideas were noted such as the purpose of the data transfer, access control, location and duration of storage and potential for collaboration.

“…sharing the data for what purposes, the type of data being shared with people…And the last part is the publication. We need really to mention how publication should be organised in the case of data sharing.” [Country 16]

“…It should cover what the data will be used for…how long it will be stored and how it will be disposed of?” [Country 7]

“…What are the research benefits of that data transfer? [Researchers] just want to transfer data and then nothing happens, we need to see the impacts of transferring that data…” [Country 12]

Other interviewees raised some appealing considerations for a data transfer agreement suxh as accountability for data anonymisation and protection as well as mandatory monitoring and evaluation of data transfers to ensure adherence to its initial intended purposes.

“…signing a commitment form to ensure that data will be protected… anonymity will be maintained… and the agreed terms and conditions… There should be some measures to monitor. It is not just about transferring data or sharing data.” [Country 9]

Level of implementation

Interviewees were split between whether a national or an institutional level DTA would be most impactful. Interviewees in favour of a national DTA argued that having a single, standard DTA would resolve inequities and limitations that may arise from having several varying data sharing agreements, each conforming to different institutional ideals regarding data access, restrictions and regulatory frameworks etc. Interviewees added that a national DTA may reduce any reliance on the guidelines and frameworks of other countries.

“… Having one standardised data sharing agreement and one standardised material transfer agreement can resolve many challenges, inequities and inappropriateness that might exist in having many different data sharing agreements…” [Country 3]

On the other hand, a few interviewees expressed support for an institutional level DTA and justified that a national-level DTA would be limited. One interviewee mentioned:

“…[DTAs] would be best at an institutional level or even between research groups. I would be quite sceptical of a data agreement on a national level because it’s almost impossible to plug all the gaps… Data sharing at a country level will need legal advice, clarity on what types of data will be shared, how it’ll be used, by whom and for what…” [Country 19]

Key stakeholders in developing and reviewing DTA

Interviewees added that RECs, researchers, research institutions, government officials and legal experts are key role players in collectively developing a national DTA. Interviewees also highly suggested community members from which samples or data are collected from as important informants who should have their voices heard in the development of a DTA.

“I think researchers themselves should be involved…they practically share the data and are the ones that will be the most affected…” [Country 6]

“If there is a community leader or even a Community Advisory Board, I think they should be included as well or a traditional leader, depending on where the data would come from—which community or communities?”. [Country 19]

Regarding the DTA review process, once it has been developed, some respondents stated that, because of the variety of skills among their members, RECs were ideal stakeholders to review DTAs; while others sided with legal entities or DACs. On the contrary, some interviewees expressed disagreement towards RECs reviewing DTAs as they felt that they are not skilled enough to evaluate such a document. These interviewees supported the idea of DACs reviewing DTAs as they serve to regulate who has access to data.

“I would definitely support a DAC with proper legal input and with proper ethics input [as] the people that are generally on a REC will not have the skill to make sure that all these processes are in place…would be critical if we want to be research-intensive universities, we want to be more competitive. We must make data shareable, but we do that responsibly and then we need to put some effort and money and resources into such committees…can do this well…who can do it timeously…” [Country 6]

However, a few interviewees expressed unfamiliarity with the concept and role of DACs.

“Well, I’m not really sure [what DACs are], but I think there should be some guidelines [that describe their role and purpose]…” [Country 4]

Regulatory frameworks

Despite some of the favourable outcomes relating to evidence-informed policymaking, other interviewees perceived legal challenges in this regard and indicated that their countries lacked regulatory frameworks for data sharing.

“The absence of data sharing laws has placed us in a position where many people, especially outside of the country know a lot of things that they are not supposed to know, or rather unconfirmed information…” [Country 14]

“When you do not have regulatory frameworks regarding data sharing, like in my country’s case, then you are in trouble because you do not have anything of your own to use for guidance. You tend to rely on the other person who is asking for your permission to have access to your data… which is disadvantageous to those owning the data being shared…” [Country 13]

The absence of regulatory frameworks for data sharing provided leeway for researchers from the Global North to engage in “helicopter research” as reported by some interviewees. This is further explained as Global South researchers being denied or restricted in their participation and contribution to research projects, ranging from its conception to publication of findings.

“…some researchers outside the country come to collect data and while they wait for protocol approval, they include local people and local researchers, then when it comes to writing a manuscript, they publish without local names… this is really, completely unfair because if you come to my country, it is me that has to write about my country. It is me that must be a principal investigator because it is my country… People write a draft and send it to local researchers to review and even if you do not agree, they do not take it into account… this really annoys and unfortunately, we do not have laws for such issues…” [Country 17]

This interviewee further highlighted the need for data sharing to be regulated as noted below:

“…data sharing has to be really very, very controlled.” [Country 17]

Consent

Failure to achieve awareness of and prior consent regarding the sharing of data were identified as a significant issues by our interviewees. In certain cases, the research participants may lack knowledge of the fact that their data will be shared or utilised for alternative purposes. Several interviewees emphasised the necessity of informing research participants about this aspect during the process of obtaining consent.

“The participant does not [always] know that his or her data is going to be shared or is going to be used for another purpose. You cannot trace the participants again to sign a consent [form]… and I do not know how it can be solved…” [Country 3]

“…it is expected that if you want to share data, appropriate permission should be taken… when this is not done, it will also lead to a legal issue…” [Country 10]

Interviewees stressed the importance of using data only for its intended collected purpose unless further consent is obtained from the participant or data subject. One interviewee commented:

“How do you ensure that the data is being used for exactly [the same purpose] it was collected for? Anything beyond the primary objective of collection requires informing participants.” [Country 7]

“…one experience we had [in our institution] involved [a student] who took some specimens from the hospital and actually transported them outside [the country]. He wanted to use that information for his PhD studies and consent was not taken. There were concerns [about whether] the patients would want to know about their material…” [Country 7]

Confidentiality and anonymity

Interviewees agreed that maintaining confidentiality is challenging when sharing data, with one interviewee emphasising the importance of proper research study designs:

“If the original research team puts proper measures in place to protect participants through a proper study design where confidentiality is maintained from the beginning and data is shared only in de-identified format then there is no issue…” [Country 19]

One interviewee raised the issue of anonymising data, particularly when working with sensitive information even if the data has been de-identified and further mentioned issues with genomics and digital technologies:

“…We were requested to make all the transcripts available [from a study exploring the perceptions of people living with HIV] by the journal and we refused because it was not possible to completely anonymise the sensitive data. It is possible to identify them with some degree of accuracy and this would be harmful to the participants…I think a lot of us who got to learn computers later, and the Internet, you know, have an inherent distrust of, you know, who has access to data because we don’t fully understand how people can gain access to data…My fears are always for genetic information. Because, you know, we’ve seen internationally that if you have databases that can speak to one another that have lots of detailed information, then people can be linked to their genetic code… And I do fear that this data could be linked to the person that to the person, and that can have lots of implications for families, for communities” [Country 6]”

Stigmatisation and discrimination

Most interviewees cited stigmatisation and discrimination as potential social challenges to data sharing and described how sensitive data may not always be well protected.

“…If the data is not well protected, then somebody can have access to that and can be able to tell that it is [person A] who provided this information and that person can be stigmatised or discriminated [against]…” [Country 9]

“…with genetic studies, there could be findings that have social implications when disseminated to people without appropriate precautions. It could lead to stigma, labelling and social problems causing scepticism in research participation.” [Country 10]

Study limitations

The study is not without its limitations and should be acknowledged when interpreting the findings. Interviews were only conducted in English although some interviewees were from Lusophone and Francophone African countries. This may have influenced their depth of articulation, yet they were still able to proficiently convey their ideas in English. We recommend that future studies include a translator during interviews with non-native English speakers, funding and resource permitting. Future studies would also benefit from involving more than one participant per country for a broader representation and better transversality to other RECs within a particular country. Another limitation to our study is selection bias as not all RECs or all SSA countries were approached. This was a baseline study to get an overall picture of SSA REC members’ perspectives on data sharing. Despite these limitations, this study provided valuable insights into the various perspectives and experiences of REC members regarding data sharing, particularly because most studies conducted in SSA primarily focused on researchers’ views. Although the challenges faced in reviewing data-intensive-related protocols were discussed, it was unclear how RECs managed these challenges and it would be useful for future studies to further explore this phenomenon and build on these findings.