1

Introduction

Sharing research data on public health issues can promote expanded scientific inquiry and has the potential to advance improvements in public health. Although sharing data is the norm in some research fields, such as the social sciences, sharing of data in public health is not as firmly established. On March 29–31, 2015, representatives of the Wellcome Trust, Bill & Melinda Gates Foundation, National Institute on Aging, Special Programme for Research and Training in Tropical Diseases of the World Health Organization, INDEPTH network, South Africa Medical Research Council, and Academy of Sciences of South Africa organized a workshop in Stellenbosch, South Africa, to explore issues related to sharing research data to improve public health in an African context. Hosted by the South African Medical Research Council and the Academy of Sciences of South Africa, the workshop brought together public health researchers and epidemiologists primarily from the African continent, along with selected international experts, to talk about the benefits and challenges of sharing data to improve public health, and to discuss potential actions to guide future work related to public health research data sharing.

In the course of five major sessions, each characterized by a keynote presentation and ample time for panel and floor discussions, the workshop participants discussed many issues that are detailed further in this summary:

• There is growing international support for data sharing. The public health benefits of research conducted with shared data have been demonstrated in multiple settings. Funders of public health research, including members of the Public Health Research Data Forum, have begun requiring data sharing. Similarly, the United States has begun implementing a national requirement that data generated by federally funded research be shared. A long history of data sharing exists in the social sciences. Some journals are also now requiring that data be shared as a condition of publication; other journals (e.g., Ubiquity Press) are developing approaches to allow data to be published. In the course of the meeting, several examples of collaborations that fostered an environment for data sharing were offered.
• Data-sharing issues are not unique to Africa, but context matters. The issues and concerns tied to data sharing raised at the workshop (e.g., confidentiality, data quality, community relevance, cost, and ownership) are relevant to data sharing in a general public health context. However, the historical context of exploitation in Africa, the power imbalance resulting from the tendency for data collectors to be in Africa and analysts to be in the Global North, and the lack of infrastructure, combined with resource inequities that exacerbate this lack, raise a unique situation for resolving these issues in the African context. Familiarity with data sharing and data-sharing issues also appears to be limited in low- and middle-income countries, including African countries. This may contribute to nervousness among African researchers with sharing data outside of established collaborations. It also highlights the need for equity and fairness in research contracts.
• Data from Africa should benefit Africa. Many participants at the workshop conveyed a sentiment that data generated from Africa should result in a benefit to Africa—not just in terms of public health generally, but for the data subjects, African researchers, and African institutions involved. Collaborations such as H3Africa, INDEPTH, and the ALPHA network provide useful illustrations of how data sharing can be used in health policy and models to address data-sharing issues.
• Sharing has both risks and benefits. Sharing of data presents both risks and benefits, or challenges and opportunities, to the individual providing the data, to the researcher, to the institution, and to the community. Developing a framework for data sharing requires an appropriate balance of the relevant risks and benefits.
• Data sharing exists within a data cycle continuum. To have data of sufficient quality and quantity to enable them to be shared and

• productively used in secondary research requires attention to issues of data collection, standardization, curation, and management, including the associated costs. Attention to these issues requires consideration of the roles, expectations, and benefits of those involved at each stage of the research process, and for the roles, expectations, and benefits to be articulated at the start of a project.

• Data sharing is a crucial element of the research continuum. Sharing data at the end of a research project is an essential part of the process. Collecting good, standardized data that can be shared should be viewed as good scientific practice.

This page intentionally left blank.