As the world’s largest healthcare provider, the NHS has a huge amount of health data that scientists and researchers need to access to help find a cure or prevent disease. Scientists and researchers haven’t always had as much access to NHS patient data as they would like. But the threat of COVID-19 has prompted the NHS to make the vast repository of data available to scientists and researchers as quickly as possible, to help them find answers to questions such as why some people are more likely to die from the virus and whether the drugs a patient takes can have side effects, whether they develop serious symptoms or not. OpenSafely, a new open-source analytics platform , has made the health records of tens of millions of people in the UK who have been in the NHS available to scientists and researchers in their fight against COVID-19. Through the OpenSafely platform, scientists and researchers will be able to analyse the electronic health records of millions of people in an effort to understand COVID-19. The files contain the full care data of 24 million people, with more to be added soon. The analysis software is open for safety checks, scientific review and reuse. OpenSafely was created in just five weeks by the University of Oxford, the London School of Hygiene and Tropical Medicine and medical companies. NHS England acts as the data controller. While the idea of creating a data analysis platform like OpenSafely existed before COVID, the threat of the virus and the understanding of the value of the data held by the NHS prompted the organisations to implement the project.
Security and privacy concerns have previously hindered projects that have needed to use NHS data for research, due to the highly sensitive nature of the data. OpenSafely uses a series of hierarchical tables, each of which provides little information about individuals, and researchers do not have access to run a database query on a patient’s data at the event level. To keep NHS patient data as secure as possible, OpenSafely has moved from a trust-based model to an evidence-based one.
Scientists and researchers will only be able to analyze OpenSafely data in the data center that houses the company’s electronic health records. Instead of the usual model of exporting datasets that researchers work on locally (and thus exposing them to all the local security risks), all analysis takes place where the records are located, and only research tables can be exported by researchers. OpenSafely is also available under an open source license, with all the code published on GitHub alongside the study definition for the first study run on the data.
Projects like OpenSafely could ultimately help the scientific and research community take a more open, less proprietary stance on data and its analysis. The way the team has built OpenSafely is aimed at encouraging scientists and researchers to share what they do. When users create a list of codes – a list of people with a particular condition, for example – or an analytical script, everything is shared on GitHub.
It didn’t take long for OpenSafely to bear fruit. A study of 17 million records published last month found that people from “black and Asian backgrounds” were at higher risk of dying from COVID-19. It also identified key risk factors for dying from COVID-19, including being male, older, or having severe asthma or poorly controlled diabetes. OpenSafely was able to move from pilot to first study in a matter of weeks, using a team that included programmer-epidemiologists. It is expected that OpenSafely data will be used to help answer questions about how effective proposed treatments for COVID-19 are, risk factors for developing severe symptoms, how the virus can spread, how effective public health intervention is, and unintended health impacts caused by the virus, such as delayed cancer referrals or vaccinations. There are also plans for a second phase of the project, which will not be limited to COVID-19.
