Table 2.

Latent Growth Models of Treatment and Maintenance Effects

	Treatment effect
	Intercept			Slope			Treatment maintenance
Outcome	estimate	SE	Z score	estimate	SE	Z score	Estimate	SE	Z score
Dyads (n=48)
BDI	9.99	3.04	3.29**	−0.44	0.19	−2.25*	2.73	4.15	0.66
BSS	4.07	1.58	2.57*	−0.09	0.06	−1.47	1.60	0.98	1.62
Parent (n=24)
BDI	9.83	4.85	2.03*	−0.73	0.24	−3.07**	2.62	5.71	0.46
BSS	2.07	1.81	1.14	−0.5	0.05	−1.08	0.97	1.03	0.94
Adol (n=24)
BDI	10.16	3.48	2.92**	−0.14	0.19	−0.74	1.76	3.92	0.45
BSS	1.66	1.74	1.52	0.02	0.06	0.32	2.05	1.70	1.20

All composite models test baseline (intercept) and treatment (slope) effects using latent linear growth models, with treatment regressed on intercept and slope processes as well as the 48 month maintenance outcome. Dyad analyses involve multilevel latent growth models of both parent and adolescent outcomes with dyadic membership indicated as the cluster variable.

^*p<0.05, ^**p<0.01.

BDI, Beck Depression Inventory; BSS, Beck Suicide Scale.