New Oxford technology assesses cancer patient vulnerability to COVID-19

The web-based COVID risk prediction tool, QCovid, is the product of the latest research to emerge from the QResearch database co-founded by Prof Julia Hippisley-Cox based at the University of Oxford. The analysis of anonymised UK health records of more than 8 million adults using GP records, hospital records including intensive care data, mortality data, Cancer Registry data and COVID-19 testing data from late January 2020 to April 2020 allows clinicians to estimate someone’s  risk of COVID-19 infection as well as the risk of being admitted to hospital with serious illness due to the virus and the potential risk of COVID-19-related death. QCovid takes into account a range of risk factors such as age, gender ethnicity and medical conditions and enables the NHS to make evidenced based decisions when prioritising different patient groups for shielding and COVID-19 vaccination.

The work was commission by England’s Chief Medical Officer Chris Whitty, who involved the team led by clinical epidemiologist, Prof Julia Hippisley-Cox, at the University of Oxford as the group has acquired extensive experience in developing risk prediction tools for a range of diseases, including QCancer for the prediction of having  undiagnosed cancer, which are widely used in the NHS. QCovid has now been validated and published in the British Medical Journal, is accessible to the public at www.qcovid.org and has been adopted by NHS Digital as a way to assess the relative risk of COVID-19 for all members of the population, based on their medical history and other risk factors.

Based on the prediction tool, 1.7 million patients have been added to the shielding list, including many cancer patients. Those within the most ‘vulnerable’ group who are over 70 will have already been invited for vaccination and 820,000 adults between 19 and 69 years will now be prioritised for a vaccination.

CANCER & VULNERABILITY

Although previous studies from the University of Oxford have shown that blood cancer patients are at higher risk of COVID-19, until now no research or model had been published that assessed patients with different types of cancer, including their treatment history and backgrounds. The QCovid algorithm feeding into the prediction tool takes cancer factors such as diagnosis of blood, lung, oral or bone cancers and different cancer therapies into account.

Thus using the QCovid prediction tool, it has a been highlighted that those with blood (acute myeloid leukaemia, chronic myeloid leukaemia, acute lymphoblastic leukaemia, chronic lymphocytic leukaemia, Hodgkin lymphoma, non-Hodgkin lymphoma, multiple myeloma) and respiratory (lung, laryngeal, nasopharyngeal and mouth) cancers are at increased COVID-19 risk.

In addition, those undergoing therapy (including recent bone marrow or stem cell transplant, chemotherapy, radiotherapy, immunotherapy or other antibody treatments for cancer and treatments that affect the immune system such as protein kinase inhibitors or PARP inhibitors), have also been identified at higher risk.

Thanks to the QCovid algorithm cancer patients now can be appropriately categorised and prioritised based on their type of cancer, current or previous cancer treatment and other factors such as corresponding health conditions that could make them more vulnerable to COVID-19.

The development of the QCovid model was led by the University of Oxford and involved researchers from Cambridge, Edinburgh, Swansea, Leicester, Nottingham, Liverpool, the London School of Hygiene & Tropical Medicine, Queen’s University Belfast, Queen Mary University of London and University College London. It was funded by the NIHR, NIHR Oxford Biomedical Research Centre, Wellcome Trust (ISSF) and John Fell Fund and supported by EMIS GP practices and the University of Nottingham.