Martijn van Teffelen

Imagery-enhanced cognitive restructuring: Efficacy 81 4 cognitive behavioral therapeutic techniques in clinical practice. After training, the students’ performance with a pilot-participant was assessed by the study’s authors (M.T. and J.L.) and the clinical psychologist (E.M.). Also, the study’s authors (M.T. and J.L.) listened to and provided feedback on the first two tapes for all therapists. During the study, the therapists had two supervisory meetings with the clinical psychologist (E.M.). In these meetings, difficult therapeutic situations were discussed to optimize protocol adherence and therapist competency. All sessions were audio taped. Two independent raters evaluated a random selection of 55 (out of 87) tapes on treatment adherence and competence. Together the raters covered all but five audiotapes that were used for training purposes. This selection procedure is comparable to that of other studies (Feeley, DeRubeis, & Gelfand, 1999; Shaw et al., 1999). Twenty-eight tapes were double coded to examine inter- rater reliability. The raters were one master-level and one third year bachelor-level clinical psychology students with previous clinical experience. Like the therapists, these students were extensively trained in conducting the treatment protocols. Both raters were blind for condition. To measure treatment integrity, a short version of the Cognitive Therapy Scale (CTS) (Dobson, Shaw, & Vallis, 1985) and the Collaborative Study Psychotherapy Rating Scale – version 6 (CSPRS-6) (Hollon et al., 1988) was used, supplemented with three items examining the competency of transforming cognitions into mental images (e.g., “To what degree did the therapist direct the client to transform an alternative thought into a helpful mental image?”). Items were scored on a 7-point Likert scale. Higher scores indicate higher treatment integrity. Overall, tapes were rated as good to excellent in 71% of I-CR cases, 97% of CR cases and 75% of AC cases. As expected, compared to the CR and AC conditions, therapists significantly displayed more I-CR specific behaviors in the I-CR condition, F (2, 79) = 1214.77, p < .001. Compared to the I-CR and AC conditions, therapists displayed significantly more CR specific behaviors in the CR condition, F (2, 79) = 200.25, p < .001. Interrater reliability of treatment integrity subscales (i.e., ICC ’s between .71 and .98) was good to excellent according to Cicchetti (1994) and fair to excellent according to Portney and Watkins (2000). There was no significant difference between therapists in assigned number of treatment conditions ( p = .574). Compared to the other therapists, one therapist displayed a significantly reduced overall integrity rating, F (2, 79) = 16.82, p < .001. This difference remained significant after controlling for patient difficulty level (i.e., one CTS item), F (2, 78) = 10.82, p < .001. For this reason, the variable therapist was treated as a random effect within our analyses.