Creating a new NHS England: Health Education England, NHS Digital and NHS England have merged. Learn more.
National COVID-19 Chest Imaging Database (NCCID)
The NHS AI Lab set up this national database to support better understanding of COVID-19 and develop technology enabling the best care for patients hospitalised with a severe infection.
On this page you can find out how to:
The National COVID-19 Chest Imaging Database (NCCID) is part of the AI in Imaging programme at the NHS AI Lab. It is a centralised UK database containing chest X-ray (CXR), magnetic resonance imaging (MRI) and computed tomography (CT) images from hospital patients across the country. The database was created to support a better understanding of the COVID-19 virus and develop technology which will enable the best care for patients hospitalised with a severe infection. It is a joint initiative established by the NHS Transformation Directorate at NHS England and Improvement, the British Society of Thoracic Imaging (BSTI), Royal Surrey NHS Foundation Trust and Faculty.
The benefits of collecting chest imaging data are extensive. This data has the potential to enable faster patient assessment in A&E, save radiologists’ time, increase the safety and consistency of care across the country, and ultimately save lives. It is being made available to researchers, clinicians, technology companies and all those wanting to investigate the disease and develop solutions that can support the COVID-19 patient care pathway.
The research enabled by the chest imaging database will provide information and tools that, in the context of the COVID-19 pandemic, support:
- the determination of disease severity
- clinically useful diagnosis and prognosis
- patient triage and management
- decision making.
How do I provide data?
We have worked closely with Royal Surrey NHS Foundation Trust to scale up an existing data collection process used to gather mammography data for research. This technology de-identifies the data at the point of collection, so that the centralised data warehouse will only store de-identified data.
Another advantage of this solution is that it uses an existing clinical system, the image exchange portal (IEP). This is beneficial as the processes for transferring images via IEP are well known and robust.
Get further information on the data collection process on the NCCID web portal.
If you have any questions regarding participation in the data collection process, please contact firstname.lastname@example.org.
How do I request access to the database?
Access to the database can be provided to those wanting to investigate COVID-19 and develop solutions that support patient care.
Find more information on the categories of data collected and how to access the database on the NHS GitHub website.
How patient data is used in the NCCID
Patient data is a vital part of the NHS's ability to stay up-to-date with innovations in healthcare. It is an invaluable resource for the research and development of new treatments and technologies.
- Everyone involved in the NCCID complies with strict laws around data.
- Access to pseudonymised data is managed by an NHS AI Lab committee.
- Pseudonymised data can only be accessed for COVID research.
- Access is time limited to a year.
Get more information about how patient data is used in the NCCID.
Read about the value of patient data for research and innovation.
Performance testing of AI models
A portion of the images and clinical data points in the NCCID has been set aside for the purpose of assessing the performance and fairness of AI models that have been developed in relation to COVID-19.
It is important that models are tested on previously unseen and population-representative datasets. The performance of AI models is linked to the characteristics of the data that they have been trained on, and those they encounter once put into clinical practice. This is the source of an important concern about the use of AI in healthcare: how to ensure that the claims made about a model will prove to be the case in the real world, where the data can be different from that used to develop it.
Assessing the performance of AI models on a dataset that is representative of the UK population reduces the potential for bias and provides NHS commissioners and healthcare regulators with the evidence to judge how safe and fair the technology is, whether it is effective and how well it will perform in large-scale clinical use.
Case study: Setting standards for testing AI
This blueprint assessment process is a valuable example of the quality testing that needs to take place for all AI algorithms designed for health and care.
The NHS AI Lab has been working closely with NHS commissioners, regulators and end-users to better define a process for ensuring the fair and ethical use of AI. Read more about our work on AI Ethics.