Predicting Risk of Hospital Admission in Patients With Suspected COVID-19 in a Community Setting: Protocol for Development and Validation of a Multivariate Risk Prediction Tool

Ana Belen Espinosa-gonzalez, Ana Luisa Neves, Francesca Fiorentino, Denys Prociuk, Laiba Husain, Sonny Christian Ramtale, Emma Mi, Ella Mi, Jack Macartney, Sneha N Anand, Julian Sherlock, Kavitha Saravanakumar, Erik Mayer, Simon De Lusignan, Trisha Greenhalgh, Brendan C Delaney

Research output: Contribution to journalArticlepeer-review

7 Citations (Scopus)



During the pandemic, remote consultations have become the norm for the assessment of patients with signs and symptoms of COVID-19 in order to decrease the risk of transmission. This has added to the already existing challenges experienced by primary care clinicians when assessing suspected COVID-19 patients due to the uncertainty around disease progression (e.g., risk of deterioration around the 8th day of disease) and has prompted the use of risk prediction scores, such as NEWS2, to assess severity and guide treatment. However, the risk prediction tools available have not been validated in a community setting and have not been designed to capture the idiosyncrasy of COVID-19 infection.


The objective of this study is to produce a multivariate risk prediction tool (RECAP–V1) to support primary care clinicians in the identification of those COVID-19 patients that are at higher risk of deterioration and facilitate the early escalation of their treatment with the aim of improving patient outcomes.


The study follows a prospective cohort observational design, whereby patients presenting in primary or community care with signs and symptoms suggestive of COVID-19 will be followed and their data linked with hospital outcomes (hospital admission, intensive care unit admission and death). The collection of the primary data for the model will be carried out by primary care clinicians in four arms, i.e., North West London Clinical Commissioning Groups (NWL CCG), Royal College of General Practitioners Research and Surveillance Centre (RCGP RSC), Covid Clinical Assessment Service (CCAS) and South East London CCGs (Doctaly platform), and will involve the use of an electronic template that incorporates a list of items (known as RECAP-V0) thought to be associated with worse disease outcome according to previous qualitative work.. This data will be linked to patient outcomes in highly secure environments (iCARE and ORCHID secure environments). We will then use multivariate logistic regression analyses for model development and validation.


Recruitment of participants started in October 2021. Initially, only NWL CCGs and RCGP RSC arms were active. As of 24th of March 2021, we have recruited a combined sample of 3,827 participants in these two arms. CCAS and Doctaly joined the study in February 2021, with CCAS starting recruitment process on the 15th of March 2021. The first part of the analysis (RECAP-V1 model development) is planned to start in April 2021 using the first half of the NWL CCG and RCGP RSC combined datasets. Posteriorly, the model will be validated with the rest of NWL CCG and RCGP RSC data as well as CCAS and Doctaly datasets. The study was approved by the Research Ethics Committee on the 27th of May 2020 (IRAS number 283024, REC reference number: 20/NW/0266) and badged as NIHR Urgent Public Health Study on 14th of October 2020.


We believe the validated RECAP-V1 early warning score will be a valuable tool for the assessment of suspected COVID-19 patients’ severity in the community, either in face-to-face or remote consultations, and will facilitate the timely escalation of treatment with the potential to improve patient outcomes. Clinical Trial: ISRCTN registry (ISRCTN13953727)
Original languageEnglish
Article numbere29072
JournalJMIR research protocols
Issue number5
Publication statusAccepted/In press - 1 Apr 2021


Dive into the research topics of 'Predicting Risk of Hospital Admission in Patients With Suspected COVID-19 in a Community Setting: Protocol for Development and Validation of a Multivariate Risk Prediction Tool'. Together they form a unique fingerprint.

Cite this