The key elements of good data sharing practice
This is a guest post by
The Wellcome Trust is a leading partner in the Public Health Research Data Forum, which brings together research funders who are committed to increasing the sharing of health research data in ways that are equitable, ethical and efficient and will accelerate improvements in public health. On behalf of the Forum, the Trust funded a major international study of stakeholders’ views about best practices for sharing public health research data from low and middle income settings, which recently published its results. Dr Susan Bull and Prof Michael Parker, from The Ethox Centre, University of Oxford, discuss the key issues and findings of the study.
Data-sharing is increasingly seen as an important component of effective and efficient biomedical research – both by researchers, and research funders. At the same time, it is recognised that efforts to increase access to individual-level data raise important ethical and governance challenges, some of which may vary depending on the context in which the research takes place.
The primary argument in favour of more routine sharing of de-identified research data is its potential to generate more – and higher quality – science. This could in turn lead to improved health outcomes, and promoting the public good.
Maximising the use of data is also seen as an efficient use of resources, and a key element of demonstrating respect for research participants – by increasing the impact of their contribution to research.
While previous studies have examined the perspectives of research stakeholders in high-income settings on sharing research data, there is a dearth of evidence about the perceptions and expectations of stakeholders in low- and middle-income countries. To help address this, our study examined the views of researchers, research managers and community members about best practices in data sharing. We conducted the research in Thailand, India, Vietnam, South Africa and Kenya in order to begin to establish an evidence base for models of good data sharing practice.
The findings of the study suggested four key factors as being important considerations in judging whether any particular data sharing initiative is likely to be viewed as an example of good data sharing practice, and thus likely to command support in the development of models of data sharing practice.
- assessing the value and benefits of data sharing
- minimising risks of harm and safeguarding the privacy and confidentiality of research participants,
- promoting fairness and reciprocity,
- instilling trust and trustworthiness among participants, communities, researchers and the wider public .
In the majority of the settings we looked at, routine sharing of data from clinical and public health research was a relatively new concept, even among researchers. The majority of sharing that had taken place was with known researchers and in collaborative relationships.
Researchers noted the value of sharing within existing or novel collaborations as this helped them to minimise the potential risks of sharing data, and appropriately promote participants’ interests. Sharing data within collaborations was seen to support trust building and capacity development. In addition, collaboration could improve the quality of research by ensuring that the research context was understood and any errors addressed early. It also helped to ensure appropriate recognition of the initial researchers’ contributions to the secondary research.
Participants in the study emphasised the importance of transparent, accountable, efficient, fair, and proportionate governance processes in instances where data sets are shared with researchers where there is no ongoing collaborative relationship during secondary data analysis.
Given the relative novelty of sharing data from clinical and public health research, some participants discussed the value of reviewing different approaches to governing data sharing that are currently in place in a range of settings. They also mentioned that guidelines for the constitution of governance procedures might be useful.
Our research, together with existing evidence, suggests that whilst greater data sharing is desirable, high ethical standards of practice are going to be an essential component of sustainable, successful data driven science.
It is likely that some approaches to data sharing are more likely to promote effective and efficient research (while also maintaining public trust) than others. It is also possible that the most effective and sustainable approach may vary between settings and different forms of data or research.
As a key output of our project, we have established a website that focuses on ethics and research data sharing, with useful links to reports, articles, guidance and platforms for sharing data.
We invite you to join the discussion group and let us know about your experiences with sharing individual-level data, and look forward to your input.
Find out more about the study results and related resources in this special issue and accompanying paper, website and free online course. Please visit the Public Health Research Data Forum website to discover more about the Forum’s work.