. Author manuscript; available in PMC: 2015 Mar 12.

Published in final edited form as: Pediatr Cardiol. 2013 Feb 2;34(5):1226–1236. doi: 10.1007/s00246-013-0633-4

Table 6.

Significance of each variable included in the final model

Variable	DOF	Univariate^a,^c		Multivariate^b,^c
Variable	DOF	LRT	AIC	LRT	AIC
Sex	1	419	34,302	31	27,529
Risk category	4	5,648	29,079	3137	30,629
Age group	2	3,309	31,414	676	28,172
Time period	4	1,020	33,707	768	28,260
Surgical volume	1	505	34,216	25	27,523
Center effect	1	419	34,300	181	27,679

DOF degrees of freedom, LRT likelihood ratio test, AIC Akaike information criterion

For the univariate analyses, the LRT is used to compare the intercept-only model (AIC, 34,300) with the corresponding univariate model: lower AIC indicates that the model fits the data better

For the multivariate analyses, the LRT is used to compare the full model including all variables (AIC, 27,500) with a model omitting one variable at a time: higher AIC indicates that the removed variable fits the data better

For both models, higher LRT indicates a more significant impact of a variable