Skip to main content

Advertisement

Table 2 Parameter estimates (PE) and associated t-values for predictors and moderators in relation to number of unprotected sexual occasions

From: Predictors and moderators of outcomes of HIV/STD sex risk reduction interventions in substance abuse treatment programs: a pooled analysis of two randomized controlled trials

Baseline variable Predictor (Main effect) Moderators (Intervention group by predictor)
  Zero Count Zero Count
Main Hypotheses t PE t PE t PE t PE
Severity of Current Primary Drug Use
ASI Drug Use Composite -1.21 -1.24 2.91 1.17** -.41 -.65 -.60 -.38
Recent Sex Under the Influence .31 .08 -.01 -.00 -1.14 -.54 2.94 .57**
Secondary Hypotheses         
Duration of Drug Problem -.81 -.01 -1.44 -.01 .05 .00 1.98 .02*
In monogamous relationship? 4.29 1.06*** -1.22 -.11 1.69 .83 .83 .15
Gender (men = 1; women = 0) .06 .01 2.52 .26* -.52 -.27 .97 .18
Age -.44 -.01 -4.13 -.02*** .54 .83 -1.19 -.01
Race/Ethnicity         
European American, non-Hispanic vs. -3.43 -.85*** 1.25 .12 -.44 -.20 -.66 -.12
All others         
African American, non-Hispanic vs. 2.87 .93*** -2.83 -.32** -.59 -.35 .05 .01
All others         
Hispanic vs. all others 1.27 .45 .03 .05 2.59 1.60** 2.18 .53*
Main Hypotheses         
Severity of Current Primary Drug Use         
ASI Drug Use Composite .14 .17 1.68 .73 -1.17 -2.83 -.92 -.64
Recent Sex Under the Influence .52 .15 -.52 -.05 -1.14 .82 2.67 .56**
Secondary Hypotheses         
Duration of Drug Problem -.71 -.01 -.25 -.00 1.15 .04 2.17 .02*
In monogamous relationship? 4.06 1.14*** -1.19 -.12 1.00 .65 .93 .18
Gender (men = 1; women = 0) -.57 -.19 2.92 .34** 1.62 -1.06 .73 .15
Age -1.06 -.02 -2.07 -.01* -1.87 -.08 -.81 -.01
Race/Ethnicity         
African American, non- 2.38 .93* -1.35 -.17     
Hispanic vs. all others         
Hispanic vs. all others      3.00 2.45** 2.05 .53*
  1. Note. Parameter estimates and associated t-values from ZINB analyses are given. For univariable tests, the DF for the t-values is 1454, except for Recent Sex Under the Influence, for which the DF was 1365. For multivariate tests, the DF for each t-value is 1353. The parameter estimates for the zero column corresponds to the log-odds coefficients for a unit increase for a given predictor for any unprotected occurrence. Positive coefficients correspond to higher prevalence of occurrences per unit increase in the given predictor; a negative coefficient corresponds to lower prevalence of occurrences per unit in the given predictor. The parameter estimates for the count column corresponds to the log of the multiplicative change in the average number of unprotected occurrence per unit increase in the given predictor. Positive coefficients correspond to an increase in the number of unprotected occurrences per unit increase in the given predictor; a negative coefficient corresponds to a decrease in the number of occurrences per unit increase for the given predictor. Main effects (predictors) are evaluated without interaction terms in the models. Moderators (intervention by predictor) are evaluated with relevant main effects in the model. Outcome is number of unprotected sexual occasions at 3 and 6 months. Analysis is a longitudinal model focusing on estimating the population average over the 3 and 6 month outcome assessments. Analyses are controlling for site, study (gender), and baseline number of unprotected sexual occasions. Sample size is 824. *p < .05; **p < .01 ***p < .001.