Development and validation of a multivariable prediction model for infection-related complications in patients with common infections in UK primary care and the extent of risk-based prescribing of antibiotics

Table 5 Performance metrics for the prediction models fitted to the derivation cohort (CPRD GOLD)

	LRTI	URTI	UTI
C-statistic (area under curve)	0.719 (se = 0.007)	0.71 (se = 0.005)	0.821 (se = 0.018)
R²	0.006 (max possible = 0.217)	0.001 (max possible = 0.036)	0.003 (max possible = 0.048)
Likelihood ratio test	1032 on 29 df, p = 0	2571 on 29 df, p = 0	366.2 on 28 df, p = 0
Wald test	1023 on 29 df, p = 0	3005 on 29 df, p = 0	258.3 on 28 df, p = 0
Score (log-rank) test	1218 on 29 df, p = 0	3926 on 29 df, p = 0	436.9 on 28 df, p = 0
Internal bootstrap concordance (C-statistic)	0.719	0.710	0.821