Data Safe Havens for Health Innovation and AI Development

When it comes to developing healthcare solutions, Scotland has an abundance of high quality, population-wide, longitudinal data that can be harvested for innovation. As the potential uses of digital data in healthcare become more apparent, it is important to ensure that this data is collected, stored, and used in a secure manner.

Health Data is Captured for Research

For over a decade, the Health Informatics Centre (HIC) has been running as a Trusted Research Environment (often referred to as a ‘Safe Haven’) on behalf of the Scottish Government, as well as NHS Tayside and NHS Fife. The HIC was the first centre in Scotland to provide opportunities for the secure use of electronic healthcare data within a controlled Safe Haven environment. In Scotland, it operates as a Local Safe Haven in Dundee alongside four additional regional hubs, located in Aberdeen, Edinburgh and Glasgow, as well as the national Safe Haven at National Services Scotland.

The national Safe Haven captures eHealth data at a population-wide level and makes it available for both administrative and research purposes. The HIC and fellow regional Safe Havens collect and provide depth to this data, such as laboratory results, that cannot be captured nationally. Currently, there is great interest in trying to streamline the process of conducting cross-Safe Haven projects that benefit from both the depth and breadth of data.

Safe Havens keep Data Secure

All data within a Safe Haven is pseudonymised but there is still the potential risk of re-identification by triangulation if the data is widely available. Naturally, the public are unhappy with the idea of their data being shared generally with industry. Therefore, the use of Safe Haven environments can streamline the data access process and maintain public trust that their data is safe and secure. This allows for valuable data to be analysed and used for innovative research and development within a secure environment without the risk of sharing potentially identifiable data.

New Applications in AI

Safe Havens are traditionally used for in-centre standard observational statistical analysis on unconsented data, with only aggregate-level results being acquired for further research or publication. However, centres are actively looking to enhance these Safe Havens to become suitable for projects such as work with imaging data, genomic data, multi omic data and AI development, as opposed to just observational studies. The Safe Haven centres are actively advancing to allow for the development and use of scalable tools and scalable, population-wide information within this environment to facilitate these projects. Ongoing research has resulted in using the Amazon Cloud to make use of tools such as TensorFlow, machine learning pipelines and tools for analysing genomic data, in order to meet current demand and expand opportunities for research and collaborations within industry.

Securing the Future of Data Sharing for Research

With these developments and this huge potential on our doorstep, there are exciting and important opportunities for further realising the potential of healthcare data. For example, improving the clarity around policies concerning the acceptable, trustworthy use of Safe Haven data in industry from both Scottish and UK-wide government. This clarification will enable researchers to continue benefiting from this resource and retaining public confidence. The process of accessing good quality data must be streamlined in a way that is transparent so that the public are aware of how their data is being safely protected and used. It also must be improved for industry and academics to provide ease and clarity when completing the process, ensuring the correct paperwork and approvals are fulfilled in order to make use of this valuable resource for developing ground-breaking healthcare solutions.

Find out more about the Data Safe Havens and the Health Informatics Centre

By Abbey Hart, alumni of The Graduate Employability Masterclass and Marketing Executive, The Life Sciences Marketing Academy