TABLE 3.

Comparison of criminal justice and service use outcomes for study participants randomly assigned to forensic assertive community treatment (FACT) or enhanced treatment as usual (control)^a

	Total		FACT		Control		Per-protocol analysis		Intent-to-treat analysis
Outcome	M	SD	M	SD	M	SD	Coef.±ME	p	Coef.±ME	p
Total arrests	1.1	1.5	.8	1.3	1.3	1.7	−.53±.74	.165	−.50±.68	.159
Total convictions	.7	1.1	.4	.7	.9	1.3	−.88±.77	.028	−.86±.72	.023
N of incarcerations	1.9	2.3	2.3	2.3	1.5	2.2	.52±.62	.105	.42±.56	.151
N of incarcerations minus sanction incarcerations	1.4	1.9	1.3	1.5	1.5	2.2	−.11±.62	.967	−.16±.60	.602
Days in jail	32.5	46.7	21.5	25.9	43.5	59.2	−.62±.63	.056	−.71±.60	.025
N of emergency room visits	1.8	3.9	1.6	4.8	1.9	2.7	−.18±1.2	.778	−.19±1.1	.728
N of hospitalizationsb	.6	1.2	.3	.7	.8	1.4	−1.1±1.1	.051	−1.1±1.0	.042
Days in hospitalc	14.1	47.3	4.4	15.1	23.8	64.2	−1.69±1.4	.018	−1.68±1.4	.025
Days in treatmentd	237.5	135.9	305.5	92.1	169.4	139.6	.64±.27	<.001	.59±.29	<.001
Outpatient treatment contactse	63.0	65.3	112.0	59.3	14.1	14.5	2.1±.37	<.001	2.1±.38	<.001

^aEffect sizes and p values are from negative binomial regression of variables in the per-protocol and intent-to-treat analyses. The coefficient (coef.) is the estimated effect size (estimated regression coefficient). The margin of error (ME) is 1.96 times the estimated standard error of the regression coefficient. Variables in the per-protocol analysis were analyzed by using the period of eligibility to receive treatment as the offset. Variables in the intent-to-treat analysis were analyzed without an offset.

^bIncludes hospitalizations in both general and state psychiatric hospitals

^cIncludes days in both general and state psychiatric hospitals

^dDays between group assignment and participants’ last contact with treatment providers

^eIncludes all face-to-face contacts with FACT team members in the FACT group and all face-to-face contacts with treatment team members and with case managers in the control group