Outcomes Risk Scores for Covid-19 Vulnerable Populations

**Background:** Outcomes Based Healthcare (OBH) is currently working on **COVID-19 'vulnerable' population analysis**, identifying cohorts 'at risk' of severe illness. This includes two key groups identified by the CMO/NHS Digital: those 'At High Risk' (c.1.5 million) on the 'Shielded Patient List', and those normally 'At Risk' of flu (c.19 million). This project builds on a programme of work undertaken for a different purpose (**'Bridges to Health' national population segmentation**), over the past 12 months, in conjunction with NHS England (NHSE), Public Health England (PHE) and Arden & GEM Commissioning Support Unit (AGEM). Using linked, longitudinal, nationally held data (less opt-outs), those 'At risk' of severe illness from COVID-19 can be identified and quantified, at local, regional and national levels, using relevant conditions in the existing segmentation model.

**Project:** OBH will evaluate relative risk of different conditions/risk factors on a) admission to hospital b) admission to ITU and c) mortality relating to COVID-19\. Development of a COVID-19 Mortality Risk Score (CMRS) and COVID-19 Admission Risk Score (CARS) will enable the risk profile of local areas to be accurately assessed according to the latest emerging evidence. This analysis would also support understanding of the effect of age on COVID-19 complications independent of comorbidities, thus providing insight into genuine risk factors versus confounding variables. Given time needed to develop a vaccine, this work will also inform interim management of, and exit from, lockdown restrictions. It may also be used to track indirect health outcomes arising from the pandemic.

**Innovation:** Rapidly repurposing the robust foundations of an existing national person-centred data model developed over a number of years, this work reflects an innovative approach to supporting the COVID-19 response, enabling rapid analysis within the bounds of OBH's existing data access arrangements and privacy safeguards. Development of COVID-19 Risk Scores would inform predictions of future patterns and trends of infection, whilst also providing insights to inform the exit strategy from the current lockdown restrictions in the UK. The flexibility of the model allows new risk factors and conditions to be incorporated in real-time as new evidence emerges. The outputs of the model can be rapidly circulated via several different methods, including the COVID-19 Health Data Research UK Slack Channel, as well as via the FutureNHS Platform. As this work is built on standard international coding terminology (ICD-10 and OPCS), this work could be translated internationally to support other digitised health systems.

**Extension for Impact Funding**: the way health systems respond to the Covid-19 pandemic is rapidly evolving, requiring OBH to adapt their support at the same pace. For instance: a) the clinical criteria for the vulnerable populations have evolved and recently have been amended by the government requiring additions to derive these from the source datasets. b) utilisation of condition information from the vaccine clinical trials became a priority so that OBH can support NHS England, as requested, with national prioritisation of administering Covid vaccines when they become available in due course c) over the summer, a group from Oxford University has developed an individual Covid risk score, based on a subset of national primary care records. The data used for this risk score is different from the data being used for this project. However, both our work and the Oxford work can be used for cross validation to refine and improve our understanding of the conditions putting people most at risk of Covid-19