The effectiveness of a Housing First adaptation for ethnic minority groups: findings of a pragmatic randomized controlled trial

Table 3 Odds ratios from unadjusted and adjusted logistic regression models examining the effect of baseline measures on percent of time spent stably housed during 24 months of study (< 50 % of time vs. ≥ 50 % of time)

Baseline variables	Unadjusted^a		Adjusted
	OR (95 % CI)	P	OR (95 % CI)	P
Treatment group	7.17 (3.83 to 13.42)	< 0.001	7.86 (4.06 to 15.23)	< 0.001
Age (years)	1.01 (0.98 to 1.04)	0.47
Gender	0.57 (0.29 to 1.12)	0.10	0.58 (0.29 to 1.16)	0.12
Total length of homelessness (years)	0.96 (0.91 to 1.02)	0.24	-	-
High School or higher education	0.73 (0.39 to 1.37)	0.33	-	-
Immigrant Status	0.92 (0.46 to 1.84)	0.82	-	-
Length of time in Canada	0.99 (0.96 to 1.02)	0.36	-	-
English was first language	1.07 (0.57 to 1.99)	0.83	-	-
MINI diagnoses
Psychosis	0.34 (0.17 to 0.65)	0.001	0.37 (0.18 to 0.75)	0.01
Major depressive episode	1.55 (0.81 to 2.96)	0.18	1.03 (0.50 to 2.10)	0.95
Alcohol or substance abuse or dependence	1.10 (0.59 to 2.05)	0.77	-	-
Post-traumatic stress disorder	1.75 (0.80 to 3.84)	0.16	1.32 (0.58 to 3.05)	0.51
Number of chronic medical conditions	1.06 (0.95 to 1.18)	0.28	-	-

^aAll univariate analyses were adjusted for treatment groups (adapted HF vs. usual care). Reference categories for categorical variables were as follows (indicated by 0): treatment group (usual care = 0, adapted HF = 1); gender (female/other = 0, male = 1); education (less than high school = 0; completed high school/some college/university = 1); immigrant status (Canadian born = 0; foreign born = 1); English as first language (other languages = 0; English = 1); MINI diagnoses (absence of diagnosis = 0; presence of diagnosis = 1)