Applied child health informatics - our impact

In 2019 the UK Government commissioned an independent report by Dr Eric Topol to explore the impact of new technology and informatics (data) developments on healthcare. This highlighted that genomics, digital medicine and Artificial Intelligence will have a major impact on patient care in the future, but to get maximum benefit we need to invest in people as well as technology.

In that review, GOSH’s Digital Research, Informatics and Virtual Environments (GOSH DRIVE) was given as the best example of how a hospital’s data infrastructure has potential to improve healthcare.

DRIVE provides a physical space, meeting rooms and technology to enable collaborations in innovation, development and rapid use of digital medicine within the NHS. Patients, the public and researchers can be contacted, engaged and consulted in this space, developing research and feeding back as projects develop – allowing active patient involvement.

DRIVE also has a state-of-the-art digital infrastructure that is unique in the NHS. It provides a secure, so called ‘cloud-hosted’, trusted digital space where data can be collected from electronic patient records and other sources, shared and studied within GOSH. The platform can ‘de-identify’ data, to remove personally identifiable information, and make sure this process works so that, if needed, we can work with partners outside the NHS, with complete certainty that confidentiality and data governance standards are upheld. DRIVE has the very highest levels of digital security, working within the ‘five safes’ principles recommended by the national Institute for Health Data Science, HDR-UK. Namely, safe people, safe projects, safe settings, safe outputs and safe data.

The real-world patient benefits of our digital set-up were shown when GOSH rapidly published data showing the impact of COVID-19 on children (Isitt, Lancet Child Adolesc. Health, 2020), and we were the only UK site able to contribute to the global Consortium for Clinical Characterization of COVID-19 that identified and tracked different SARS-CoV-2 variants.

GOSHDRIVE has also supported over 100 collaborative projects with UCL computer science students, connecting them to NHS patients and real-life healthcare problems. The student programme is growing, supporting PhD students across the UK to help meet the need identified in the Topol report to invest in people as well as technology.

GOSH’s approach and the DRIVE infrastructure have been nationally and internationally recognised, with consultations and demonstrations of our work given across Canada, America, Australia, South East Asia and the UK.

The COVID-19 pandemic emphasised the need for high-quality, rapidly available clinical data to learn about the condition, what complications may occur and in which patients and compare approaches to treatment to improve outcomes. With our existing data infrastructure in the GOSH Digital Research Environment team (DRE- part of the wider GOSH DRIVE infrastructure), we were the only UK hospital that could contribute to the global Consortium for Clinical Characterization of COVID-19 via our Electronic Health Record.

To contribute we regularly and rapidly provided data on our COVID-19 patients, but this had to be ‘de-identified’ to remove all personal information before being shared with our collaborators. Our data structure and expertise of GOSH DRE allowed us to provide this secure data and analysis on a daily basis, playing a significant part in a global study of the effect of COVID-19 on hospitalised children1. This study collected data from 27 hospitals across the world, including 671 children (over 100 from GOSH) and almost 30,000 laboratory values, for 16 separate laboratory tests. It quickly showed the common complications (cardiac arrythmias, blood clotting abnormalities etc) and what test abnormalities are expected in children with COVID-19. These data are now used internationally to help clinicians diagnose and manage COVID-19. The collaboration illustrated the potential of using large-scale data-based approaches to analyse data across health care systems, something only possible with a data infrastructure like GOSH’s DRE.

In addition to these contributions, the GOSH DRE data platform has been crucial to several projects studying and defining serious complications of COVID-19 in children. These range from the widely reported inflammatory syndrome PIMS-TS (Paediatric Multisystem Inflammatory Syndrome, a rare condition triggered by COVID-19), to some lesser-known problems in kidney, gut and neurological function, among others. All these findings have been shared with the scientific community via high impact publications to ensure widespread awareness of these complications.

The way that we carried out this research, pooling groups of secure, de-identified data, could be easily applied to many adult and childhood conditions. In fact, as part of the International Precision Child Health Partnership, GOSH is collaborating with Boston, and Melbourne Children’s Hospitals which specialise in rare and complex childhood diseases. We will analyse genomic and clinical data from cases in these centres, using early-onset epilepsy as the first project to improve understanding and hopefully treatment of this condition.

We know taking certain drugs while pregnant can cause problems in the developing baby. These are usually very rare and to show any association requires study of many thousand pregnancies. We have shown how using data collected during routine health care visits combined with very detailed data collected at specialist centres like GOSH, can show small but significant associations.

Researchers from GOSH and GOS-ICH investigated possible links between taking antibiotics called macrolides during pregnancy and subsequent birth defects or neurological conditions in children. We analysed data from 104,605 children born between 1990 and 2016 and their mothers’ pregnancy history, one of the largest populations ever studied. We used information recorded electronically via the Clinical Practice Research Datalink as part of GP appointments, and showed that children whose mothers took macrolide antibiotics during early pregnancy were at a very small increased risk of major birth defects, especially heart abnormalities. However, it is important to note that the overall risk was extremely small.

The work was published1 and shared with the public using international media coverage, including the New York Times, CNN and ABC in America and Sky News, ITV, the Telegraph, and the Daily Mail in the UK. The coverage was so broad that the Science Media Centre in the UK collected and reported expert reactions to the findings, so journalists could widely publicise what this means for the treatment of infections in pregnancy

Our work resulted in a change in recommendations to suggest other antibiotics and not macrolides should be used where possible in pregnancy. Following publication of the study, the Commission on Human Medicines (the organisation that advises the Government on drug safety and effectiveness) advised the Medicines and Healthcare products Regulatory Agency (the agency responsible for ensuring that medicines are acceptably safe) to review more population-scale evidence on the safety of macrolide antibiotics in pregnancy. More research is now underway.

Our study shows the true power, potential national impact and benefit to patients of analysing data that is already routinely collected nationally in the UK, using both healthcare data recorded in General Practice and during specialist care. We are using this same method to see if a treatment called Palivizumab can prevent high-risk children being admitted to hospital with bronchiolitis, a common but potentially severe chest infection.