Appendix: Choosing technology to support data sharing and access

Choosing a data repository

The Global Health Network's Data Repository Finder catalogues a wide variety of data repositories for specific types of data, including clinical trials, social sciences data, observational studies and health economics. Each type of data lends itself to a specific type of data repository. In addition, data repositories may focus on specific disease groups or geographic regions.

Key actions:

• Identify what features are required from a data repository, for example the ability to document a data dictionary, the ability to set licensing conditions, and so on.

• Review available data repositories that match these criteria for the type of data and the features required.

• When a shortlist of available data repositories has been determined, consider asking data partners and other stakeholders which data repositories they are most familiar with or would prefer to use.

Choosing a data platform

When more fully featured data sharing is required, and if resources are available, it may be necessary to build your own data platform. There are a range of options available. Commercial data governance providers such as data.world, Alation, Akvo and Atacama offer data platforms where data can be stored and made available under specific arrangements to external parties. These platforms often include a granular level of control where specific access permissions can be set for each external role, so that only specific elements of data are shared with agreed partners.

It may be necessary to build a solution from the ground up. Emerging solutions such as the Australian Lab3 infrastructure model, from Microsoft and technology partners, enables larger health providers to build a fully featured health data platform that may include components such as analysis and sandbox environments. Other technology providers offer similar tools.

Key actions:

• Identify what features are required from a data platform, for example the ability to set granular permission controls, the ability to secure data within the platform for analysis within a sandbox environment, or analytics to log who makes use of data.

• You will need to have a budget and IT involvement if you choose to host your own data platform. Work with your data manager, data systems lead and IT teams to prepare and fully document the business case and review available pre-built data platforms, or consider the implications of building a solution using components from existing technology providers.