Page 178 - PhD GT

results were described above. However, further examination of the sample seen at all three data collection points, Sample 6c, provided the basis for tracing the changes that occurred over time. Data from this sample provided the basis of examining the question of the rate and sequence of change as well as the nature and predictors of such change.
A total of 101 individuals were seen at all three data collection points. Fifty-nine per cent of them were primary heroin users and 41% were primary alcohol users (see Table 8.11). Due to the greater likelihood of tracing alcohol users compared to heroin users for follow-up, the difference in proportion between the two substance groups is the smallest in this sample. The proportion of men is smaller in the heroin group compared with that in the original heroin group (75% compared with 80.9%) and the proportion of men in the alcohol group is significantly smaller than that in the original sample (66% compared with 75.6%). These differences are due to the finding that there is a greater likelihood of follow-up for women than there is for men at both three months and at twelve months and that this likelihood appears to grow over time.
Heroin users are significantly younger than alcohol users, have a shorter duration of use and a shorter duration of problem use (an average of 2.5 years) than alcohol users (who have an average of 9.6 years of problem use). This suggests that problem heroin users are likely to seek help much earlier in their problem using career than alcohol users.
As in Sample 6, other aspects of the pattern of use of the two substances varied: while similar proportions of individuals were abstinent at t1, marked differences in the proportions using on a daily basis and using on less than a daily basis were observed. A significantly greater proportion of heroin users were daily users of their primary drug (78.3% compared with 48.8% of alcohol users), though no significant differences were found in severity of dependence between the two substance groups at t1. Levene’s test for homogeneity of variances for dependence confirmed no significant differences in the variances of the two drug groups. In one-way analysis of variance no significant difference emerged between the means for the two drug groups (F=1.8; significance = .18) and it was therefore thought satisfactory to explore the nature of change in dependence for the drug groups combined.
No significant differences were found between the substance groups on scores of psychological ill-health or social dissatisfaction in the sample seen at the three data collection points. Relationships between these and dependence scores are further explored in Chapter 9.
173