Page 85 - PhD GT

21.19 (SD = 6.8, range 0-30, n = 528). Different distributions of LDQ scores are shown for the two substances, with the alcohol group showing a flatter distribution of scores and the opiate group scores being more clearly skewed towards the higher scores. Heather et al. suggested that, due to the fact that the score most often achieved, i.e. the mode, was 30, there may have been a ceiling effect in the measurement of dependence by this instrument.
Heather et al. (submitted) found a significant negative relationship between LDQ score and age in the total sample and in the alcohol sub-sample. No relationship with age was found in the opiate or the ‘other drugs’ group. A significant relationship was found between gender and LDQ score in the total sample and in the alcohol sub-sample. This was not found in either of the other drug groups. These findings are at odds with those of the earlier study of Raistrick et al. (1994) who found no significant relationships between age or gender in either of their substance sub-samples. Although age and gender were proposed in that study to be used as criterion variables for discriminant validity analysis, there may be a case for hypothesising that age would be associated with alcohol dependence which is thought to have a much longer lead-in time than does opiate dependence. This may account for the finding of a significant correlation with age in the larger alcohol sub-sample. In the present main study a significant difference is found in the mean age of the two substance groups and in the duration of use and of problem use prior to attending for treatment (see below, Chapter 8).
Heather et al. (submitted) found a significant difference in mean GHQ scores between their substance sub-samples. They also found a positive, significant correlation between GHQ scores and LDQ scores for the whole sample and for each of the drug group sub-samples. In a multiple regression analysis used to predict LDQ score from other variables they found that age, gender, GHQ score and substance were independently predictive of LDQ score in the whole sample. Younger, male clients with more psychological disturbance (i.e. higher scores on the GHQ) and whose main substance problem was with drugs other than alcohol gained higher LDQ scores. However, while in the alcohol sub-sample all three background variables, namely age, gender and GHQ score were independent predictors of LDQ score, in the opiate group only GHQ and gender were significant predictors and age was not. In the “other drugs” sub-sample GHQ score alone remained as a significant predictor. Where predictors were found in the sub-samples they worked in the same direction as in the total sample.
In the Heather et al. study further support was given to the earlier conclusion that the LDQ was an instrument with good psychometric properties capable of measuring a unitary construct of dependence in clinic attenders using different psychoactive substances. In order to examine its
73