Skip to main content

Table 4 Effects of covariates on seizure occurrence or rate (SANAD)

From: Guide to evaluating performance of prediction models for recurrent clinical events

Variable

NB: OR (95% CI)

ZINB logit: OR (95% CI)

ZINB count: RR (95% CI)

AG: HR (95% CI)

PWP-TT: HR (95% CI)

Intercept

6.54 (5.66, 7.47)

2.23 (1.24, 3.25)

6.31 (5.583, 7.08)

N/A

N/A

Gender

Female

1.00

1.00

1.00

1.00

1.00

Male

1.15 (0.79, 1.67)

0.60 (0.41, 0.88)

1.37 (1.00, 1.88)

1.31 (0.90, 1.91)

1.09 (1.00, 1.20)

Primary relative with epilepsy

Absent

1.00

1.00

1.00

1.00

1.00

Present

2.10 (1.36, 3.37)

2.00 (1.21, 3.42)

1.83 (1.28, 2.68)

1.74 (1.07, 2.81)

1.09 (0.96, 1.24)

Treatment

Valproate

1.00

1.00

1.00

1.00

1.00

Lamotrigine

1.74 (1.13, 2.69)

1.83 (1.18, 2.85)

1.52 (1.05, 2.21)

2.01 (1.27, 3.18)

1.14 (1.02, 1.27)

Topiramate

1.55 (1.01, 2.39)

1.33 (0.87, 2.03)

1.55 (1.06, 2.26)

1.92 (1.18, 3.12)

1.17 (1.03, 1.31)

Age at randomisation (years)

0.28 (0.21, 0.37)

0.44 (0.32, 0.60)

0.34 (0.27, 0.43)

0.36 (0.24, 0.53)

0.90 (0.83, 0.98)

Annual tonic–clonic seizure rate prior to randomisation

0.86 (0.81, 0.91)

0.96 (0.89, 1.03)

0.86 (0.82, 0.90)

0.88 (0.81, 0.95)

1.02 (1.00, 1.03)

  1. NB negative binomial, ZINB zero-inflated negative binomial, AG Andersen-Gill, PWP-TT Prentice, Williams and Peterson total time, RR relative risk, OR odds ratio, HR hazard ratio