Page 235 - PhD GT

significantly different from dysfunctional group scores but the reliability of the mean score for the functioning group is less certain.
Jacobson et al. (1999) argue that the method they propose enables them to distinguish, in clinical populations, “the percentage of clients who improved but did not recover, the percentage of clients who recovered and the percentage of clients who remained unchanged or who deteriorated” (p. 300). Kazdin (1999) suggested that the measure of reliable change alone was useful in that the purpose of judging treatment outcomes might be well served by establishing degrees of improvement that do not necessarily render the individual within the range of well functioning people but can be shown to result in improved coping on the everyday level and therefore constitute a realistic and reasonable goal of treatment.
The analyses in Chapter 9 were based on the decision that all change was of interest in the present study, that confining change to that which was clinically significant would exclude those who had change but who, in the words of Jacobson et al. were improved but not recovered. In the calculation of clinically significant change, the general population score used was derived from a very small sample, though application of the more stringent of the criteria for clinically significant change, that the end point score would fall outside two standard deviations below the mean dependence score at intake (taken as the norm for the dysfunctional group) produced a very small group for analysis. Nonetheless, application of one and of both the criteria did produce some interesting data: 50% of the cohort seen for follow-up at three months had achieved reliable change in the direction of improvement in level of dependence. A third also had clinically significant change, was made up of equal proportions of each of the substance groups and deteriorated over the next nine months. Another group achieved nearly no change in dependence between intake and three months but clinically significant change between three months and twelve months. Further work is required to look at what distinguishes the two groups. In conclusion, all change is worthy of examination, but the fact of change as suggested by statistically significant differences in mean scores does requires further scrutiny. An amount of change will inevitably occur: referred to as regression to the mean, good scores will get worse and bad scores will improve. Criteria are required to establish when change in scores is due to factors other than such natural regression or to unreliability of the measuring instruments and for interpreting the meaning of such change from individually differing starting points. For future calculations of reliable change, larger general population samples at least are required.
230