Skip to main content

Table 3 HGLM regression of illegal drug and alcohol use to intoxication onto treatment processes, capacity-building training, client and organizational predictors (n = 497)

From: Therapeutic community drug treatment success in Peru: a follow-up outcome study

  Decrease in Illegal Drug Use Decrease in Alcohol to Intoxication
Predictors Odds Ratio Effect Size (r) T ratio Odds Ratio Effect Size (r) T ratio
Treatment       
TC length of stay 1.19 0.09 2.04* 0.95 -0.02 -0.36
TC Model Implementation fidelity 1.08 0.14 3.03** 0.97 -0.03 -0.74
TC treatment satisfaction 1.01 0.00 0.07 0.62 -0.12 -2.66**
Capacity-building Training       
TC training intensity 1.15 0.03 0.7 1.2 0.06 1.2
TC staff participated in drug treatment training 1.44 0.08 1.67 1.14 0.03 0.68
TC director Daytop training in 1999 1.46 0.04 0.89 0.86 -0.03 -0.62
Client characteristics       
Client age 1.14 0.12 2.62** 1.13 0.11 2.43*
Client education 0.83 -0.07 -1.59 0.91 -0.04 -0.8
Client employment status 0.58 -0.04 -0.8 0.94 -0.02 -0.46
Perceived stigma 0.54 -0.19 -4.27** 0.12 -0.12 -2.67**
Client ethnicity 0.89 -0.04 -0.92 1.06 0.02 0.49
Client attended other treatment 1.14 0.03 0.59 1.56 0.09 1.91
Organizational characteristics       
Percentage of clients that dropped out before 30 days 1.01 0.06 1.22 1 -0.02 -0.51
Institution certified or not 0.87 -0.04 -0.82 1.17 0.04 0.77
Director collaborated with staff 1.42 0.08 1.64 0.9 -0.03 -0.58
Director's time spent with clients 0.41 -0.13 -2.75* 0.82 -0.04 -0.84
Director's ethnicity 0.96 -0.01 -0.26 0.85 -0.06 -1.35
Director's age 1 0.01 0.12 0.96 -0.07 -1.45
Director a recovering addict 1.15 0.03 0.65 1.16 0.02 0.45
Intercept .00 -.08 -1.63 8.21 0.03 .553
Interactions       
Stigma × implementation fidelity --- --- --- 1.03 0.11 2.39*
Client age × Implementation fidelity .99 -.12 -2.53* .99 -.08 -1.75
  1. Note: D.F. = 473; An attrition bias correction score [not in table] was used as a control variable, which increases the generalizability of these results; *p ≤ 0.05, **p ≤ 0.01. Effect size = point – biserial correlation with a small effect size equals .10, a medium effect size equals .24, and a large effect size equals .37[73]. Length of stay coefficient is scaled to show increase in logistic coefficient for every 100 day increase in length of stay. TC = Treatment Center.
\