. Author manuscript; available in PMC: 2014 May 5.

Published in final edited form as: Adm Policy Ment Health. 2012 Mar;39(0):133–145. doi: 10.1007/s10488-012-0415-y

Table 3.

Results of Multilevel Multinomial Logit Models (Estimates of Intercepts and Slopes) of Session Content Addressed or Focused on by SRF Topic and Informant

SRF Topic	Informant¹	N	Addressed Topic		Focused on Topic
SRF Topic	Informant¹	N	Est. Intercept²	Est. Slope³	Est. Intercept²	Est. Slope³
Emotional issues	Cl	1297	0.95^*	0.10	−0.91	0.17^*
	Y	1089	1.21^*	0.08	−0.52	0.16^*
	Cg	716	1.12^*	0.05	−0.35	0.08
Behavioral issues	Cl	1303	0.88^*	0.12^*	−1.25^*	0.28^*
	Y	1091	1.23^*	0.03	−0.20	0.07
	Cg	718	1.31^*	0.05	−0.11	0.13^*
Family issues	Cl	1296	1.38^*	0.21	−0.13	0.53^*
	Y	1088	1.42^*	0.14	−0.12	0.39^*
	Cg	714	1.71^*	−0.04	0.56	0.15
School/work issues	Cl	1293	0.72^*	0.05	−1.03^*	0.38
	Y	1087	0.89^*	−0.07	−0.55	0.05
	Cg	710	0.57^*	0.14	−0.73	0.23
Friends/peer issues	Cl	1295	0.43^*	0.10	−1.64^*	0.21^*
	Y	1087	0.63^*	0.06	−1.45^*	0.20^*
	Cg	707	0.41	0.09	−1.50^*	0.12
Problems w/delinquent behavior	Cl	1275	−2.09^*	0.38^*	−3.73^*	0.58^*
	Y	1075	−1.83^*	0.09	−3.74^*	0.40
	Cg	704	−1.69^*	0.15	−3.67^*	0.48
Alcohol/substance use	Cl	1273	−2.22^*	1.09^*	−3.97^*	1.41^*
	Y	1074	−1.98^*	0.68^*	−3.55^*	1.29
	Cg	697	−2.02^*	0.87^*	−4.10^*	1.29
Harm to self or others	Cl	1272	−2.00^*	1.09^*	−4.20^*	1.61^*
Client hope for future	Y	305	0.25	0.23	−1.45^*	0.19
Client motivation for treatment	Y	261	0.33	−0.15	−1.78^*	0.05
Therapeutic relationship w/client	Y	1114	−1.17^*	0.15	−4.09^*	0.31
Client perceptions on counseling impact	Y	467	−0.34	0.10	−2.76^*	0.07
Client satisfaction w/services	Y	298	−0.62	0.20	−3.14^*	0.35
Caregiver strain	Cg	194	−0.47	0.19	−2.69^*	0.37
Caregiver satisfaction w/life	Cg	234	−0.80	0.09	−2.97^*	0.21
Therapeutic relationship w/caregiver	Cg	690	−1.26^*	−0.08	−3.62^*	0.00

Note: Y=Youth; Cg=Caregiver; Cl=Clinician.

Intercepts give the estimated log-odds of addressing or focusing on a content area relative to the reference group (not addressed) when informants do not report any problem alerts (problem alerts =0).

Slopes are the multinomial logit estimate for one unit increase in the number of problem alerts reported by the informants on clinician-reported content addressed or focused on in sessions compared to the reference group (not addressed).

p < 0.001.