Table 2.
Reasons for Choosing a Sober Living House (SLH) and Recovery Outcomes – Separate Regression Models.
| Reason for choosing | Abstinent for prior 6 months (dichotomized) |
Percent days abstinent |
ASI Employment Severity |
PDSQ psychiatric symptoms total |
Length of stay |
|---|---|---|---|---|---|
| OR (95 % CI LL, UL) | β (95 % CI LL, UL) | β (95 % CI LL, UL) | IRR (95 % CI LL, UL) | IRR (95 % CI LL, UL) | |
| Affordability | 1.12 (0.79, 1.57) | 2.54 (−4.08, 9.16) | 0.00 (−0.04, 0.04) | 0.84 (0.73, 0.98) | 0.96 (0.00, 290.34) |
| Wanting to live with others in recovery | 0.85 (0.62, 1.18) | −1.44 (−9.06, 6.18) | −0.02 (−0.05, 0.02) | 0.94 (0.80, 1.10) | 1.04 (0.79, 1.28) |
| Location | 1.08 (0.77, 1.51) | 4.58 (−2.13, 11.30) | −0.03 (−0.06, 0.00)* | 0.90 (0.76, 1.06) | 1.03 (0.02, 70.2) |
| No place to live/this is my only option | 1.03 (0.77, 1.38) | −3.31 (−8.88, 2.25) | 0.04 (0.01, 0.06)* | 1.21 (1.02, 1.42)* | 0.94 (0.01, 150.57) |
| Access to transportation | 1.18 (0.85, 1.62) | 4.68 (−1.91, 11.28) | 0.08 (0.05, 0.12)*** | 0.91 (0.75, 1.20) | 0.94 (0.02, 58.60) |
| Appearance of house | 1.16 (0.84, 1.61) | −1.90 (−7.80, 3.99) | 0.01 (−0.02, 0.04) | 0.96 (0.81, 1.12) | 0.92 (0.00, 185.76) |
| Amenities offered | 0.95 (0.72, 1.24) | −2.02 (−7.04, 2.98) | 0.04 (0.00, 0.07)* | 0.98 (0.82, 1.18) | 0.93 (0.00, 412.14) |
| Criminal justice, family, or someone other than myself paying for this | 1.09 (0.69, 1.74) | 4.76 (−3.75, 13.28) | 0.06 (0.03, 0.09)*** | 0.94 (0.78, 1.14) | 0.98 (0.00, 304.25) |
Note. PDSQ = Psychiatric Diagnostic Screening Questionnaire; ASI = Addiction Severity Index.
Outcomes were analyzed using multilevel mixed models (Level-1 N = 462; Level-2 sample size n = 48). Confidence intervals (CI) with their lower limit (LL) and upper limit (UL) for robust standard errors are in parentheses. Logistic regression was used for the dichotomized outcome of abstinence. Due to positive skewness, tobit regression models were used for ASI Employment and the percent days abstinent for the six months prior to the interview. Incident Rate Ratios (IRRs) for Poisson regression models are reported for PDSQ and length of stay. Time, age, sex, and race/ethnicity were also included in the models (except for time in the length of stay models), adjusting for random effects of neighborhoods and within-subjects.
Bold signifies P ≤ 0.05. ***p < 0.001, **p < 0.01, *p < 0.05.