Bayesian hierarchical piecewise regression models: a tool to detect trajectory divergence between groups in long-term observational studies

Table 1 Analyses of the divergence in BMI trajectories between T2DM adults and non-T2DM adults: assessment of Bayesian model complexity (effective number of parameters pD), and fit (deviance information criteria DIC) for each candidate model

	Model	Females	PP p-val	Males	PP p-val
Unconditional	A	26910 (2544)	0.47	19837 (2223)	0.55
T2DMgroup (int β₀)	B	26670 (2366)	0.45	19741 (2270)	0.43
T2DMgroup (childhood slope β₁)	C	26780 (2510)	0.6	19865 (2247)	0.58
T2DMgroup (Adulthood slope β₂)	D	26701 (2401)	0.58	19828 (2242)	0.62
T2DMgroup (change point CP)	E	26076 (2777)	0.52	19762 (2213)	0.51
T2DMgroup (CP + β₂)	F	26504(2430)	0.6	19860 (2271)	0.54
T2DMgroup (CP + β₀)	G	26436 (2751)	0.55	19896 (2242)	0.45
T2DMgroup (all 4 parameters₎	H	26532 (2978)	0.52	19920 (2435)	0.51

Reported for each model are DIC (pD), and posterior predictive p-values (PP p-val). Best fitting models are indicated in bold characters

ISSN: 1471-2288