Flexible semiparametric joint modeling: an application to estimate individual lung function decline and risk of pulmonary exacerbations in cystic fibrosis

Table 2 Model comparison results

Model	Submodel 1	Submodel 2	DIC₁	DIC₂	\(\bar{D}\)	p _D	DIC_total
Separate model
I	Semiparametric mixed effects	Weibull	952,566	21,110.9	962,619	11,057.7	973,676
Joint model
II	Cubic mixed effects	Weibull	952,915	20,712.5	962,485	11,142.6	973,628
III	Semiparametric mixed effects	Weibull	952,409	20,605.4	961,923	11,091.4	973,015

Each of DIC₁ and DIC₂ is the sum of the posterior mean deviance (\(\bar{D}\)) and the effective number of parameters (\(p_{D}\)); DIC_total obtained by summing DIC1 and DIC2 from separate or submodels (see “Appendix 4” for details). Lower values indicate better model fit

ISSN: 1742-7622