Time trends in exposure of cattle to bovine spongiform encephalopathy and cohort effect in France and Italy: value of the classical Age-Period-Cohort approach

Table 3 List of the APC models fitted and selection of the best models. Quality of fit and contribution of each new variable added to the model are presented.

		Adjustment of the models			Estimate of the effect of the covariates
No.	Model	Residual deviance	df	p-value*	Comparison with model	Difference of deviance	Difference of df	p-value**	Tested effect
	France
0	Null	1958.7	232	0.000
1	Age	991.8	201	0.000	0	966.9	31	0.000	Age
2	Age-Drift	344.3	200	0.000	1	647.5	1	0.000	Drift⁽¹⁾
3a	Age-Cohort	68.0	167	1	2	276.3	33	0.000	Non-linear cohort effect
3b	Age-Period	336.4	194	0.000	2	7.9	6	0.246	Non-linear period effect
4	Age-Cohort-Period⁽²⁾	55.5	161	1	3a	12.5	6	0.051	Period effect (non-linear + linear)
	Italy
0	Null	347.4	91	0.000
1	Age	224.9	81	0.000	0	122.5	10	0.000	Age
2	Age-Drift	101.4	80	0.053	1	123.6	1	0.000	Drift⁽¹⁾
3a	Age-Cohort	49.6	64	0.907	2	51.8	16	0.000	Non-linear cohort effect
3b	Age-Period	90.3	73	0.082	2	11.1	7	0.134	Non-linear period effect
4	Age-Cohort-Period⁽²⁾	32.5	57	0.996	3a	17.1	7	0.017	Period effect (non-linear + linear)

* goodness of fit test, ** log-likelihood ratio test (α = 5%), for p > 0.05 the effect of the variable is non-significant
df, degree of freedom
⁽¹⁾ linear effect of the period and cohort combined
⁽²⁾ period is the last covariate entered in the model

ISSN: 1746-6148