Clinical trial participants had clear views about priority topics they wanted information about when broad consent to data sharing is sought. However, views about how much data should be provided to research participants, and how best to promote understanding, varied substantially, and were informed by the contexts from which interviewees were drawn. These are discussed in turn below. The results are presented according to four themes. In each theme, we start with perspectives from clinical trial participants, followed by those from researchers and community members.
What is it important to know about data sharing?
The majority of clinical trial participants interviewed were of the opinion that as long as data sharing had potential benefits to society and they would not be harmed or inconvenienced by such sharing then they were happy to consent to it.
‘If it could benefit others in the future I’m OK with sharing data to others.’ (P7, clinical trial participant, female, SSI, Bangkok).
‘It’s the data that is already collected, so it think it’s OK to share it. It’s something that doesn’t have a negative impact to me anyway.’ (P6, clinical trial participant, male, SSI, Bangkok).
More specifically, the most commonly occurring topics arising in questions clinical trial participants asked during interviews and their responses to interview prompts, suggest that there are four priority areas that they would like to receive information about when consenting to data sharing. These are: that sharing research data has potential benefits; that data will be de-identified; that no foreseeable individual or group harms will come to the research participants; and that there will not be any foreseeable additional inconvenience or burdens incurred by agreeing to data sharing.
When interviewed, the majority of clinical trial participants spontaneously sought to confirm if data would be de-identified before sharing. When they were asked if they were willing to allow their data to be shared more widely, all agreed, provided that they could not be identified. Many also implied that they knew that de-identification was standard in clinical trials, which was unsurprising given that they are familiar with the activities of the MORU-affiliated health volunteer ward, and de-identification of samples and data is routinely discussed during trial recruitment.
‘Yes, I am willing to share that but only blood result, not my name, my first name. The rest are fine, including blood sample or blood test result.’ (P2, clinical trial participant, male, SSI, Bangkok).
‘She asked whether I am willing to participate or is it okay for my data to be shared. She told me that the data that will be shared contains no names or any of my identification. So, I told her it is ok. I will give my consent.’ (P3, clinical trial participant, female, SSI, Bangkok).
In addition to ensuring that names were kept confidential, some clinical trial participants noted that they would not want their location, addresses and phone numbers to be shared. In this context, where healthy volunteers were taking part in pharmacokinetic research, participants did not express concerns about potential harms of being identified as taking part in a specific study. Instead, they expressed concerns about their identifiable information being used by third parties, such as telemarketers and insurance agencies, for purposes unrelated to health research.
Clinical trial participants also raised concerns about potential additional responsibilities arising from consenting to data sharing. For example, a participant asked, ‘So the consent to data sharing in this sense means that I have to come back here again or just only this time? (P7, clinical trial participant, female, SSI, Bangkok). She mentioned that she did not mind being re-contacted by phone but she would be reluctant to attend extra meetings about data sharing as she would not want to take leave from her work. This suggests that research participants should be informed while consenting to data sharing that such sharing will not result in any additional burdens to them and how any possible recontact will be managed.
In contrast, researchers and community members noted the importance of de-identification in minimising harms to research participants, and raised concerns that de-identification at the individual level is not adequate to prevent community or group-based harms. They explained that this was because some groups, such as migrant workers on the Thai-Myanmar border, may face increased stigmatization if they are identified as being a source of infectious diseases like malaria. To mitigate these ethical challenges, one researcher suggested that it is necessary for participants to be assured in the consent process that data would be shared in a way that would minimize any harm to them and their communities (R6, researcher, male, SSI, Mae Sot).
The following sections review perspectives about how much information should be provided on these priority topics and how best to explain complex aspects of data sharing.
How much information should be provided about data sharing?
Clinical trial participants did not have strong opinions about how much information about data sharing should be provided.
However, there was substantial variation in researchers and community members’ views about how much information should be provided about data sharing when broad consent was sought. Some believed that detailed information should be given, while others thought that it was too burdensome to explain data sharing in detail.
Some researchers commented on the importance of tailoring consent to context, and suggested that as with the primary healthcare study, the amount of information to be provided about data sharing should be responsive to research participants’ health condition and state of mind at the time of enrolment in a biomedical study, their interest, level of literacy, background knowledge about research and their culture. When research participants are unwell at the time of recruitment, a two-stage process may be appropriate with specific consent for the primary study being sought at recruitment, with broad consent being sought when the patient has recovered enough to understand the details of data sharing.
‘It depends on, I think, it may be because of the clinical condition and also the another thing is their education level…we have to make a balance in our ill patients so that sometime the patient they just understand only main information, the objective of the study if they are going to involve the benefit and the advantage of the study, they know only those things. And then when they are getting better we have to explain again what we are taking. For example, like after a week of the study and then we have to explain again so at that time they feel much better and then they understand.’ (R8, researcher, female, SSI, Mae Sot).
In the quotes below a long serving CAB member stressed the importance of a more detailed explanation, while a participant from the Health Research Interest Group suggested that it was not necessary to go into too much detail about data sharing as that could be confusing and alarming for research participants.
‘We have to take time to explain until they (the clinical trial participants) understand. The bulls will not pull the cart and take off right after you connect the cart to them. As for us, we have been committee members for many years so once you explain, we understand immediately. New people will not understand it…Carefully explain one by one step by step.’ (CAB2, CAB member, male, FGD, Mae Sot).
‘…only difference between these two (without data sharing and with data sharing) is one will be used one time, that’s it…. that the data will be used by several different departments or different groups and isn’t that enough?’ (HRIG2, HRIG member, female, FGD, Bangkok).
Understandings of data sharing
We asked clinical trial participants about whether and how data sharing was explained during the consent process for the clinical trial. They said that they had been provided with enough information, but we found that some did not clearly understand data sharing or had difficulty recalling the information provided about data sharing. This finding is congruent with an earlier empirical study on understandings of consent in our setting, where many respondents commented that they could not remember or understand everything that was explained to them about research [15]. We also found that the words “data”, “sharing” and “data sharing”, as we understand them in this context, and as they were translated into Thai, were not immediately understood by the clinical trial participants, and as a consequence they exhibited a wide range of interpretations of data sharing.
It also became apparent that some researchers and community members did not feel fully confident that they understood the nuances of data sharing. This could be because they had not been involved in data sharing decisions for studies from which data was requested by secondary users. The MORU data governance process involves a review by the DAC in consultation with only senior investigators of the studies from which data are being shared. Field workers, junior staff and CAB members may not have been fully aware of the data sharing policy or the governance process. Consequently some researchers also assumed that data sharing was always via open access mechanisms, meaning that shared research data would be available online for anyone to use. Many researchers were familiar with open access as they had made their data available online as supplementary files to a journal article. Others had uploaded their datasets onto external repositories where datasets can be accessed by registered users who have agreed to the repository’s terms and conditions of use. A contrasting assumption was that data are only shared with collaborators working on the same topic: ‘there may be other researchers in (research unit) or outside (research unit) who are interested in the same topic and so we may share that health information with them’ (R9, researcher, female, SSI, Mae Sot). The importance of ensuring that researchers have a clear understanding of institutional policies and processes for data sharing, so that these can be accurately discussed with research participants, is considered below in the discussion section.
Suggestions for promoting understanding
Due to the varying understandings of data sharing, interviewees were prompted to describe such sharing in their own words. When seeking to describe data sharing, clinical trial participants from Bangkok, where the majority of the population are familiar with the internet and social media, drew analogies and likened data sharing to sharing information on social media: ‘Share can be many things like sharing on the Internet etc.’ (P3, clinical trial participant, female, SSI, Bangkok). However, others recognized ways in which privacy was protected during research data sharing and described it as the opposite of Facebook, ‘It’s like the opposite...you can know somebody’s name but you know nothing else until you are friends with them.’ (I2, research intern, female, FGD, Bangkok). This could be because the word for “data” in Thai, “khor moon” is the same word as “information”, and that one common usage of the word “share” refers to hitting the share button to openly share information on social media.
Researchers in the healthy volunteer ward in Bangkok thought that it could be helpful to use the illustration when consenting clinical trial participants and have it displayed in poster format on the walls of the waiting area of the healthy volunteer ward. In contrast, respondents in Mae Sot had mixed feelings about the value of the illustration, saying that it might be too complicated for patients with less education. One researcher explained that most of her patients are illiterate and are not familiar with the concept of clinical research and said ‘I am a hundred percent sure that ninety-nine percent of our patients would be totally bewildered by this [illustration]’ (R9, researcher, female, SSI, Mae Sot). CAB members in Mae Sot were in favour of using the illustration, saying that ‘you need to spend more time to explain that and then using a tool like visualize or photos is better because it is easier to understand than text’ (CAB2, CAB member, male, FGD, Mae Sot).
Given that data sharing is difficult to explain, we also asked respondents for suggestions about how to improve the wording about it in the information sheet and consent form. The information sheet for this study contained typical language used in MORU studies, approved by the Faculty of Tropical Medicine Ethics Committee, Mahidol University. During consent processes, this information is supplemented by verbal information. The information sheet stated “De-identified data from this study may be shared with other groups of researchers. All applications for data sharing will be reviewed by a Data Access Committee. All researchers accessing data need to adhere to a set of terms and conditions that aim to protect the interests of research participants and other relevant stakeholders”, and in the consent form, “By signing this page, I agree that de-identified data from this study may be shared with other groups of researchers”.
Some researchers and community members thought that the wording was clear, while others suggested a range of potential improvements. Some recommendations focused on simplifying the language to promote comprehension, for example ‘data will be used by several different departments or different groups’ (HRIG2, HRIG member, female, FGD, Bangkok). Other suggestions focused on clarifying and explaining concepts such as de-identification and secondary data users. “De-identification” was considered potentially confusing because the word “identity” is not usually used in daily conversation. One useful way proposed to describe de-identification was “data without your personal information for example your name, address, phone numbers etc” (khor moon thee mai rabu tuaton in Thai) (N1, researcher, female, FGD, Bangkok). Researchers also discussed the importance of clear descriptions of who the data may be shared with.
‘…we should explain …we are not the only organization to use the data. So, after taking a biological sample and the patient information, we will share with other countries. You know, it (countries) is quite understandable for the patient, I think. Other organizations or other they don’t know, NGOs or government or…? In our own words, it is called the organization is not very clear, police, army? All are organizations, you know…and in Myanmar we have research centre of course, but they never heard about it.’ (R7, researcher, male, SSI, Mae Sot).