Fig. 2 | European Journal of Epidemiology

Fig. 2

From: COVID-19 mortality in the UK Biobank cohort: revisiting and evaluating risk factors

Fig. 2

Receiver operating characteristic (ROC) curves from logistic regression models for risk of COVID-19 death. Results are presented for logistic models sequentially including age (light blue), sex (dark blue) and ethnicity (grey) (a). Results are also presented for a model sequentially including (N = 4) demographic (grey), (N = 12) social (beige), (N = 7) health risk (red), (N = 6) biological (green), (N = 10) medical (light blue), and (N = 4) environmental (olive green) factors, as well as a model including the (N = 7) factors consistently selected by logistic LASSO (selection proportion > 0.95) (purple) (b). Predictive performances were derived from a subsampling procedure (repeated independently 1000 times) of 80% of the study population as training set to produce ROC curves and corresponding AUC in the validation set (remaining 20%). The ROC curve and AUC point estimate corresponds to mean performance across 1000 subsamples, and the coloured areas (and AUC ranges) reflect the 1st and 99th percentiles of the performances yielded across the subsamples

Back to article page