Table 1.

Estimates for model fit and parameters: posterior means (standard errors)

Parameters	Model 1	Model 2	Model 3
DIC($p_D$)	5105.98	4877.68	4759.9
DIC($p_V$)	5119.56	4958.78	4864.17
${\beta }_0$	17.784 (1.258)	17.392 (1.286)	18.046 (1.410)
${\beta }_1$	0.001 (0.010)	0.014 (0.026)	−0.017 (0.022)
${\beta }_2$	–	−0.009 (0.041)	0.054 (0.039)
${\beta }_3$	–	–	−0.051 (0.050)
$C_1$	–	44.529 (1.278)	29.028 (1.186)
$C_2$	–	–	48.839 (4.168)
${\beta }_0^S$	3.790 (0.345)	3.803 (0.351)	3.808 (0.342)
${\beta }_1^S$	−0.012 (0.003)	−0.012 (0.002)	−0.012 (0.002)
$\delta$	0.397 (0.017)	0.397 (0.016)	0.399 (0.015)
${\sigma }_{b_0}^2$	42.970 (14.345)	41.289 (15.914)	40.711 (15.455)
${\sigma }_{b_1}^2$	0.002 (0.001)	0.015 (0.007)	0.007 (0.004)
${\sigma }_{b_2}^2$	–	0.035 (0.017)	0.029 (0.014)
${\sigma }_{b_3}^2$	–	–	0.042 (0.025)
${\sigma }_{b_0^S}^2$	2.543 (0.856)	2.556 (0.847)	2.572 (0.847)
${\sigma }_{\epsilon }^2$	2.146 (0.084)	1.794 (0.072)	1.646 (0.067)

DIC($p_D$): Deviance Information Criterion calculated using a penalty term for model complexity ($p_D$); DIC($p_V$): Deviance Information Criterion calculated using an estimate for the effective number of parameters in the model ($p_V$); β₀, general intercept; β₁, general change in antibiotic consumption over time; β₂, general difference in the linear trend after versus before the first change-point; β₃, general difference in the linear trend after versus before the second change-point; $C_1$, location of the first change-point; $C_2$, location of the second change-point; β₀^S, general amplitude; β₁^S, general change in amplitude over time; δ, phase shift of the sine wave; ${\sigma }_{b_0}^2$, random intercept variance; ${\sigma }_{b_1}^2$, random slope variance; ${\sigma }_{b_2}^2$, random difference (after versus before the first change-point) variance; ${\sigma }_{b_3}^2$, random difference (after versus before the second change-point) variance; ${\sigma }_{b_0^S}^2$, random amplitude change variance; ${\sigma }_{\epsilon }^2$, residual variance.