Page 238 - PhD GT

Substance group emerged as a significant predictor (p<.05) at three months and dependence at three months was a significant predictor of clinically significant change at twelve months. Thus in the main, the two types of analysis produced the same predictor variables with the exception of a role identified for substance group in the prediction of change at three months. No significant difference had been found between the mean change scores for the heroin compared to the alcohol group at three months in the entire sample but the mean change score for the heroin group (16.8) was significantly higher (p<.01) than the mean change score for the alcohol group (11.6) in those who had achieved clinically significant change at three months.
The question of whether changes in use of the substance are more directly influenced by the independent variables measured was examined. The model constructed for the purpose of examining the predictors of change in dependence was used to examine the predictors of a dichotomous use outcome (abstinence or use at three months and at twelve months) and a continuous use outcome (change in number of days using between intake and three months and between three months and twelve months). Treatment emerged as a significant predictor of abstinence at three months but not at twelve months. The independent variable which accounted for the largest amount of variance in the abstinence outcome at twelve months was dependence at intake but it only reached a level of significance of < .07. When predictors of the change in the number of days using between intake and three months was explored, three significant independent variables emerged: drug group (p<.03), dependence at intake (p<.01) and number of treatments (p<.001). Dependence at three months alone emerged as a significant predictor (p<.05) of change in number of days using between three months and twelve months.
It may be that the assumptions used in the construction of the model for regression analysis were not well founded or that the measurement points were not appropriate for finding factors predictive of dependence. In constructing the model for examining predictors of change in dependence, the question of how coping itself might have changed and how, in turn, such change might have affected change in dependence emerged. Coping at three months and at twelve months might have thrown light on this question but could not be used in the analyses of predictors of dependence at the same time points due to the possibility of other factors influencing the change in coping at follow-up. The change in coping itself could be subjected to further critical analysis.
233