Acta Psychiatr Scand zyxwvutsrqponm 1995’ zyxwvutsrqponmlk 92: 132-137 Printed in Belgium zyxwvutsrqponml - all rights reserved Copyight zyxw 0 Munksgaard 1995 ACTA PSYCHIATRICA SCANDINAVICA ISSN 0001 zyx -690X Improving the usefulness Psychodynamic Profile in from a reliability study of the Karolinska research: proposa Haver zyxwvutsrqp B, Svanborg P, Lindberg S. Improving the usefulness of the Karolinska Psychodynamic Profile in research: proposals from a reliability study. Acta Psychiatr Scand 1995: 92: 132-137. 0 Munksgaard 1995. The interrater reliability of data obtained by use of the Karolinska Psychodynamic Profile (KAPP) was tested among 60 women seeking treatment for drinking problems. The first rater had a psychodynamic background but was minimally trained rating the KAPP subscales and performing KAPP interviews. Independent, blind ratings of audiotaped interviews by an experienced KAPP rater revealed that 8 of the subscales obtained satisfactory reliability, whereas 6 subscales showed unsatisfactory reliability. Furthermore, data for one subscale (normopathy) showed a zero correlation between raters, probably due to the novelty of the construct. Additionally 3 subscales related to bodily aspects were of little clinical significance among the present study group. Our data were compared with data from previous KAPP reliability studies, and the reasons for I similarities and discrepancies of results are discussed. The construction of reliable, valid and applicable scales for rating of psychoanalytical constructs and psychodynamic formulations still represents a chal- lenge within the field of psychotherapy research. Available methods often show variable reliability (l), much the same as for rating symptoms or diagnoses (2, 3). Also, these methods are time-consuming and require extensive training by raters and interviewers to obtain adequate reliability. This applies especially to Bellak’s Ego Function Assessment Scale (EFA) (4-7). Other instruments focus on fairly limited as- pects of the subjects’ mental functioning, i.e., the central psychodynamic conflict or ego resources of importance for short-term anxiety provoking psy- chotherapy (8-lo), or they limit their scope of evalu- ation to defense mechanisms or transference reac- tions (11, 12). Lately, a new set of scales based on object relations theory aiming at a comprehensive evaluation of personality showed moderate reliabil- ity between independent raters (13). Finally, psycho- dynamic formulations made by teams of judges with contrasting theoretical positions may show adequate reliability but nevertheless be significantly different concerning dynamic content (14). With this background in mind, it is interesting when a new set of scales - Karolinska Psycho- dynamic Profile (KAPP) - challenges some of the limitations of previous methods for psychodynamic 1s z B. Haver ’, P. Svanborg’, S. Lindberg’ ’ Karolinska Hospital, Magnus Huss Clinic, Stockholm, * Department of Clinical Neuroscience, Psychiatry Section, Karolinska Institute, Stockholm, Sweden Key words: rating scale; women; alcoholism Brit Haver, PB 3323 Varden, N-5022 Bergen, Norway Accepted for publication March 18, 1995 1 ratings (15-18). The aim has been to develop a set of scales constituting a comprehensive profile of the subjects’s stable mental functioning based on con- structs close to clinically observable phenomena. In the first studies, KAPP was reported to have high interdisciplinary and cross-cultural reliability and stability over time, and minimal training of raters was needed until adequate reliability was obtained (16, 17). Knowledge of psychoanalytic theory and clinical experience of psychodynamic processes are thought to be essential to the satisfactory execution of the KAPP assessment. The interview technique is based upon Kernberg’s structural interview and is estimated to last about 2 h (15, 19). The reliability of the data obtained using the KAPP scales was tested the first time in a study among 44 (57 7; men) patients suffering from ulcer- ative colitis (16-18). The interviews were conducted by one of the psychoanalysts who developed the scales (RMW). The interrater reliability was tested both jointly and independently based on a selection of 12 among 32 available taped interviews (16). Dis- crepancies between the joint raters were discussed at the end of every rating session. Based on this pro- cedure, the interrater reliability varied between 0.2 1 and 0.98 across all subscales, and the mean corre- lation coefficient through all subscales varied be- tween 0.73-0.88 for the individual raters. The inde- 132