The challenge
As part of NPL’s COVID-19 response and offer of support, the RFL approached NPL with a number of issues that needing addressing to help them manage successfully during the pandemic. These included:
- Optimising hospital flow – Managing hospital flow is a long standing and important part of day to day hospital management. As a result of the COVID epidemic there was a requirement for Trusts to separate areas of the hospital in to COVID and COVID-free areas. The Trust asked if we could help with modelling the hospital flow to support optimal use of wards, rooms, beds and taking into account key requirements such as oxygen usage.
- Optimising the cancer pathway - Efficient and effective cancer pathway management is key to meeting national targets and patient safety. Cancer targets are challenging and often not met even without a pandemic. The RFL needed help to optimise their cancer pathways which have interdependencies such as Radiology and Pathology, as these pathways present limitations on operational resource.
- Understanding the cause and effect of thrombosis in COVID-19 patients - Coagulation in COVID patients is a significant problem and carries a high mortality rate. An understanding of how the coagulation changes over time would be of huge benefit in terms of a COVID treatment response. In the longer term lead to a greater understanding of the virus and its impact. The Trust asked for support with the analysis, evaluation and clustering of biomarkers that might hint at the underlying pathophysiological process.
The solution
Members from NPL’s Data Science and Medical Physics departments have worked on the following solutions:
- Optimising hospital flow - We provided a solution which allows patients to be routed through the hospital and allocated to treatment rooms whilst minimising interactions with non-COVID patients and keeping to the constraints of staffing and oxygen supply. Our solution has resulted in identifying the carrying capacity of each floor of the hospital and generated optimal partitioning of wards and other rooms.
- Optimising the cancer pathway - We set up an algorithm to minimise the number of breaches in cancer diagnosis time for patients coming into the RFL by referral. The Trust sent us data describing the diagnosis pathways for the different type of cancers, and we turned this into a visualisation where all pathways can be seen in the same diagram – giving a clearer picture of potential ‘blockages’. NPL also formulated a solution to the problem of optimising all the pathways simultaneously in order to give the Trust an idea of optimal resource allocation in order to meet diagnosis targets.
- Understanding the cause and effect of thrombosis in COVID-19 patients - We used deep learning and clustering methods to identify patterns in the data sets including the sample collection, patients’ medical history, hospital observations, imaging data, COVID history, admission data and patient outcome data (mortality). The results found several interesting trends linked to mortality and morbidity that are being reviewed and interpreted by clinical academics.
The impact
The outcomes and associated impact from each of these projects have proven to be extremely valuable to the Trust and will help to shape future COVID and associated hospital activity. Each of these projects have shown how valuable support from a national government lab to national hospitals can be in unprecedented and extreme situations such as the COVID-19 global pandemic.
Carolyn Ruston, Strategic Lead for Health and Life Sciences at NPL said: “The relationship we built between the Royal Free London and NPL has been a fantastic example of two areas of the health system coming together with a joint determination to improve outcomes for patients. We are looking forward to continuing our collaboration with the Trust with a range of follow on projects in the coming months".