Digital environment: Harnessing the power of data
Hundreds of thousands of data points are collected every day in routine patient care. At present only a small proportion of this information is used to inform a patient’s diagnosis and treatment plans, or at scale to benefit future patients. As an Intelligent Research Hospital, GOSH is working to harness the power of data.
​
At the GOSH DRIVE unit data scientists support teams to leverage clinical and operational data using our secure, trusted Digital Research Environment (DRE). GOSH has also established a group to oversee the use of routine healthcare data for secondary purposes and provide assurance that projects meet data governance requirements. Launched in 2018, the world-leading DRE was established thanks to vital funding from GOSH Children’s Charity.
GOSH’s data lake extends over 20 years and consists of over half a billion data items. Due to integration with the
electronic patient record, data continues to be added to the data lake at a rate of approximately 20 million data points per year.
The DRE has supported over 300 projects as of 2023. These are complex data projects including provision of very detailed data, integration of data sets, collaborative digital workspaces for data processing and analytics, and building and deploying applications.
We are part of national and international collaborative projects across industry, academia, and health organisations. For example, we worked with the European Health Data Evidence Network (EHDEN) to set-up a paediatric focus group, and support data and digital projects for the European Children’s Hospitals Organisation (ECHO), of which GOSH is a founding member.
Section Title
“As the parent of a child that needed care on a paediatric intensive care unit and public member on research panels, I have been involved in discussions around the value, perceptions and ethics of data research projects.
GOSH is a Data Controller of personal data that is collected to help provide and manage care of patients. It has obligations to keep this data secure and confidential. Therefore, to approve each project, I help the panel to consider things like whether data is sufficiently anonymised and if additional consent is needed from patients to use their data for the secondary purpose of research.
It’s important that parents’ and patients’ views are considered as this can improve the quality and impact of research. I am also a Senior Lecturer at the University of St Andrews in medical ethics and healthcare policy so I am interested in how better use of existing data can help innovation in the NHS. However, it doesn’t matter what your knowledge of data science is as sharing your personal perspective is most valuable.”
- Dr Morven Shearer, parent research panel member
Making every second count to improve critical care
Tracking, Trajectory and Triggering data, known as T3 data is a continuous data stream recorded about every 5 seconds. This includes vital signs such as heart rate and blood oxygen levels. For the most seriously ill children on the Paediatric Intensive Care Unit (PICU), this data helps healthcare professionals monitor their condition and informs treatment decisions. This data can also improve our understanding of critical conditions and interventions that could help future patients.
Dr Samiran Ray, intensive care consultant said:
“On the intensive care unit, we’re always looking to predict the next step for a patient’s treatment. One of my first projects with the DRE team looked at whether we could curate T3 data and develop predictive models to support decision-making.
Often the first procedure for a child on PICU is breathing support with a ventilator. This is also often the last treatment to be stopped when they improve. Every day, clinicians like me face the decision of when to take a child off breathing support. Together with researchers from the Turing Institute we built algorithms to support clinicians with this decision and identified challenges to doing this. We found that clinical experts are vital to find hidden meaning when data are missing.
For example, when we are looking at data from ventilators, a clinician could tell us that a loss of some signals is caused by a patient being disconnected from a ventilator temporarily to help them cough. Without a clinician the loss of signal might be misinterpreted.
Another memorable project looked at the use of low molecular weight heparin that helps to prevent blood clots forming. This can be a complication after children have been on PICU and need lots of drips or lines inserted into blood vessels. Although this is rare, sometimes children will need to have injections for up to 12 weeks, twice a day, and need many blood tests to check they are on the right dose. This can keep them in hospital for longer and be upsetting. We brought this data together to better understand how we can get the dose right sooner to reduce distressing blood tests in these children. This allowed us to change guidance for clinical teams.
Without support from the team at the DRIVE unit, we would never have been able to access T3 data, which tracks vital measures as part of everyday care on intensive care units. We have data from over 15,000 admissions at GOSH over three years. We can now securely examine this data to inform even better diagnosis, treatment, and care for patients in the most critical conditions.”