PT - JOURNAL ARTICLE AU - Colin R Simpson AU - Chris Robertson AU - Steven Kerr AU - Ting Shi AU - Eleftheria Vasileiou AU - Emily Moore AU - Colin McCowan AU - Utkarsh Agrawal AU - Annemarie Docherty AU - Rachel Mulholland AU - Josie Murray AU - Lewis Duthie Ritchie AU - Jim McMenamin AU - Julia Hippisley-Cox AU - Aziz Sheikh TI - External validation of the QCovid risk prediction algorithm for risk of COVID-19 hospitalisation and mortality in adults: national validation cohort study in Scotland AID - 10.1136/thoraxjnl-2021-217580 DP - 2022 May 01 TA - Thorax PG - 497--504 VI - 77 IP - 5 4099 - http://thorax.bmj.com/content/77/5/497.short 4100 - http://thorax.bmj.com/content/77/5/497.full SO - Thorax2022 May 01; 77 AB - Background The QCovid algorithm is a risk prediction tool that can be used to stratify individuals by risk of COVID-19 hospitalisation and mortality. Version 1 of the algorithm was trained using data covering 10.5 million patients in England in the period 24 January 2020 to 30 April 2020. We carried out an external validation of version 1 of the QCovid algorithm in Scotland.Methods We established a national COVID-19 data platform using individual level data for the population of Scotland (5.4 million residents). Primary care data were linked to reverse-transcription PCR (RT-PCR) virology testing, hospitalisation and mortality data. We assessed the performance of the QCovid algorithm in predicting COVID-19 hospitalisations and deaths in our dataset for two time periods matching the original study: 1 March 2020 to 30 April 2020, and 1 May 2020 to 30 June 2020.Results Our dataset comprised 5 384 819 individuals, representing 99% of the estimated population (5 463 300) resident in Scotland in 2020. The algorithm showed good calibration in the first period, but systematic overestimation of risk in the second period, prior to temporal recalibration. Harrell’s C for deaths in females and males in the first period was 0.95 (95% CI 0.94 to 0.95) and 0.93 (95% CI 0.92 to 0.93), respectively. Harrell’s C for hospitalisations in females and males in the first period was 0.81 (95% CI 0.80 to 0.82) and 0.82 (95% CI 0.81 to 0.82), respectively.Conclusions Version 1 of the QCovid algorithm showed high levels of discrimination in predicting the risk of COVID-19 hospitalisations and deaths in adults resident in Scotland for the original two time periods studied, but is likely to need ongoing recalibration prospectively.All code, metadata and documentation for this project is publicly available at https://github.com/EAVE-II/Qcovid-validation. A data dictionary is available at https://github.com/EAVE-II/EAVE-II-data-dictionary. Most of the data that were used in this study are highly sensitive and will not be made available publicly.