Skip to main content
. Author manuscript; available in PMC: 2010 Nov 1.
Published in final edited form as: Drug Alcohol Depend. 2009 Aug 21;105(1-2):132–138. doi: 10.1016/j.drugalcdep.2009.06.022

Table 2.

Outcome summary by treatment group

Buspirone Placebo Measure of Treatment Effect
Variable Estimate Estimate Description Estimate Standard Error p-value
Urine Drug Screens for Cannabis
    Mean (SD) % Negative, ITT (all missed assumed positive) 20.3% (32.5%) 6.5% (16.2%) Wilcoxon Rank Sum Test N/A N/A 0.13
    Mean (SD) % Negative, only observed specimens 27.6% (40.9%) 11.7% (27.8%) Wilcoxon Rank Sum Test N/A N/A 0.16
    Model-based estiamte % negative{a} 28.8% 11.0% Risk Difference 17.8% 9.9% 0.07
Timeline Follow-Back
    % Days reporting no use{a} 45.2% 51.4% Risk Difference −6.2% 10.4% 0.55
    % Reduced days using relative to baseline{b} 69.6% 85.2% Fisher's exact N/A N/A 0.30
    % Reduced amount using/using day relative to baseline{b} 91.3% 92.6% Fisher's exact N/A N/A 1.00
Hamilton Anxiety Rating Scale, mean (SD){d}
    Total score −6.10 (7.11) −4.00 (5.68) Wilcoxon Rank Sum Test N/A N/A 0.40
Marijuana Craving Questionnaire, mean (SD){d}
    Compulsion subscale −Q4.50 (4.70) −4.00 (4.72) Wilcoxon Rank Sum Test N/A N/A 0.97
    Emotion subscale −5.25 (5.20) −3.88 (4.74) Wilcoxon Rank Sum Test N/A N/A 0.43
    Expectancy subscale −5.40 (4.47) −5.84 (5.15) Wilcoxon Rank Sum Test N/A N/A 0.88
    Purposefulness subscale −7.50 (5.61) −5.08 (5.05) Wilcoxon Rank Sum Test N/A N/A 0.17
    Total score −22.65 (16.29) −19.40 (15.75) Wilcoxon Rank Sum Test N/A N/A 0.71
Withdrawal checklist, mean (SD){d}
    Total score −10.87 (8.99) −10.40 (7.73) Wilcoxon Rank Sum Test N/A N/A 0.99
{a}

Summaries represent model-based estimates from a generalized linear model using all available data.

{b}

For the Intention-to-Treat (ITT) analysis, the denominators were n=23 for Buspirone and n=27 for Placebo. Only observed results that were negative were considered

{d}

Values are change from baseline using last observation carried forward method of imputation. One placebo subject did not have any post baseline scale-based