Therapists’ professional and personal characteristics as predictors of outcome in long-term psychodynamic psychotherapy and psychoanalysis

Therapists’ professional and personal characteristics as predictors of outcome in long-term psychodynamic psychotherapy and psychoanalysis

G Model EURPSY-3089; No. of Pages 10 European Psychiatry xxx (2013) xxx–xxx Available online at www.sciencedirect.com Original article Therapists...

448KB Sizes 3 Downloads 12 Views

G Model

EURPSY-3089; No. of Pages 10 European Psychiatry xxx (2013) xxx–xxx

Available online at

www.sciencedirect.com

Original article

Therapists’ professional and personal characteristics as predictors of outcome in long-term psychodynamic psychotherapy and psychoanalysis E. Heinonen a, P. Knekt a,b,*, T. Ja¨a¨skela¨inen a, O. Lindfors a a b

National Institute for Health and Welfare, Finland Social Insurance Institution, Finland

A R T I C L E I N F O

A B S T R A C T

Article history: Received 21 March 2013 Received in revised form 14 July 2013 Accepted 22 July 2013 Available online xxx

Background: Whether long-term psychodynamic therapy (LPP) and psychoanalysis (PA) differ from each other and require different therapist qualities has been debated extensively, but rarely investigated empirically. Methods: In a quasi-experimental design, LPP was provided for 128 and PA for 41 outpatients, aged 20–46 years and suffering from mood or anxiety disorder, with a 5-year follow-up from start of treatment. Therapies were provided by 58 experienced therapists. Therapist characteristics, measured pretreatment, were assessed with the Development of Psychotherapists Common Core Questionnaire (DPCCQ). General psychiatric symptoms were assessed as the main outcome measure at baseline and yearly after start of treatment with the Symptom Check List, Global Severity Index (SCL-90-GSI). Results: Professionally less affirming and personally more forceful and less aloof therapists predicted less symptoms in PA than in LPP at the end of the follow-up. A faster symptom reduction in LPP was predicted by a more moderate relational style and work experiences of both skillfulness and difficulties, indicating differences between PA and LPP in the therapy process. Conclusions: Results challenge the benefit of a classically ‘‘neutral’’ psychoanalyst in PA. They also indicate closer examinations of therapy processes within and between the two treatments, which may benefit training and supervision of therapists. ß 2013 Elsevier Masson SAS. All rights reserved.

Keywords: Unipolar depression Anxiety disorders Psychodynamic therapy Other psychotherapy Education in psychiatry

1. Introduction An ambivalent relationship has existed between mainstream, systematic empirical psychotherapy outcome research, and the theory and practice of psychoanalysis and psychoanalytical longterm therapy [14,15,27,61,68,70]. More specifically, the ‘‘gold standards’’ of large-scale outcome clinical research (e.g., randomized designs, manualized treatments, standardized measures) have been recognized as difficult to apply to long-term treatments which necessitate therapist’s flexibility and patient’s long-term commitment and may involve highly individual outcomes [9,11,26,52,62]. Nevertheless, recently evidence has accumulated from several research projects evaluating the effectiveness of longterm psychodynamic therapy and psychoanalysis showing their similar, if not greater, effectiveness across various problems and outcome domains in comparison to other treatments [10,11,34,37].

* Correspondence at: National Institute for Health and Welfare, Mannerheimintie 166, P.O. Box 30, FIN-00271 Helsinki, Finland. Tel.: +358 29 524 8774; fax: +358 29 524 8760. E-mail address: [email protected]fi (P. Knekt).

Alongside demonstrating effectiveness, however, such investigations may also shed further light on the therapeutic and analytic process and its moderators and mediators [30,49,51]. One such longstanding [2,3,18,19,45,64,67] as well as recently debated question, involving conceptual, clinical, educational and political implications, is whether and how psychoanalysis ‘‘proper’’ differs from psychoanalytically informed long-term therapies [8,31,35,71]. While the two treatments can be categorically distinguished based on extrinsic criteria, such as frequency of sessions, use of couch, and a fully trained psychoanalyst [18,35], differentiation based on the intrinsic treatment process is more difficult [65]. Some have argued psychoanalysis and psychotherapy to differ in their goals or the strategies and techniques that practitioners employ to reach them [8,17–19,31]; others have argued that these differences are insignificant in view of, e.g., the diversity of patient needs, concrete resources and constraints on practice, and cross-cultural clinical variations, and have instead advocated viewing psychoanalytic treatment as one wide-ranging ‘‘family of treatments’’ [35,66]. Empirical investigations of the issue have thus far yielded inconsistent results. A meticulous U.S. study found psychoanalysis to differ from long-term psychoanalytic therapy, showing that in

0924-9338/$ – see front matter ß 2013 Elsevier Masson SAS. All rights reserved. http://dx.doi.org/10.1016/j.eurpsy.2013.07.002

Please cite this article in press as: Heinonen E, et al. Therapists’ professional and personal characteristics as predictors of outcome in long-term psychodynamic psychotherapy and psychoanalysis. European Psychiatry (2013), http://dx.doi.org/10.1016/ j.eurpsy.2013.07.002

G Model

EURPSY-3089; No. of Pages 10 2

E. Heinonen et al. / European Psychiatry xxx (2013) xxx–xxx

psychoanalysis the practitioners adhered more to the ideal prototypic analytic process, distinguished in terms of therapist techniques (e.g., pointing out defensive maneuvers, drawing connections between therapeutic and other relationships), therapist manner (e.g., therapist neutrality), and session focus (e.g., discussing dreams and fantasies) [1]. Yet, a German study aiming at replication and focusing specifically on techniques found little distinction between therapist techniques in the two treatments, with the differing findings from the two studies tentatively ascribed to cultural differences [57]. These studies have nevertheless emphasized scrutinizing the treatment providers; they also suggest further inquiries on whether other practitioner qualities may influence differentially the process and outcome in these treatments [31,35]. This is important not only for theoretical, but also practical reasons: outcome meta-analyses, encompassing diverse psychotherapies, suggest that therapist effects may often exceed treatment effects [69]. Although research is scarce regarding the differential effect of therapist qualities in long-term psychodynamic therapy versus psychoanalysis, investigations in Sweden of over 400 patients and 200 therapists in these treatments have shown the potential and relevance of such studies [20,49]. Assessing by self-report a variety of therapist qualities, such as professional manner and attitudes, the studies showed that therapists with classically psychoanalytic attitudes – notably, being low on kindness, self-disclosure, and supportiveness – were ineffective when conducting long-term psychodynamic therapy. The authors hypothesized that ‘‘the classically psychoanalytic stance, with less emphasis on support, coping strategies, warmth and openness, may be functional with analysands but much less so with patients in psychotherapy’’ [49]. Besides practitioners’ professional manner or attitudes, however, more personal qualities stemming from private life may also matter, as Blatt and Shahar [5] note, referring to findings that therapists’ attachment styles may influence their professional relational manner and consequently the treatment process and effectiveness [13,48,59]. In sum, both the personal and the professional qualities of practitioners should be explored as potential determinants of process and outcome in long-term psychodynamic therapy and psychoanalysis. The Development of Psychotherapists Common Core Questionnaire (DPCCQ) [44] is a recently developed self-report measure that investigates a broad range of such professional qualities–such as therapists’ skills, encountered difficulties, coping methods, and relational manner–as well as personal characteristics– such as how therapists’ experience their manner in close personal relationships–which have also been linked to therapists’ attachment patterns [56]. While a recent study found these characteristics to predict differentially the outcomes in long-term psychodynamic therapy versus short-term therapies of psychodynamic and solution-focused form [24] as well as the outcomes of mainly psychodynamic but ‘‘rather eclectic’’ open-ended therapies [42], the DPCCQ’s prediction on psychoanalysis has not yet been investigated. To complement the earlier STOPPP project findings [20,49], the present study investigates whether therapists’ professional and personal pre-treatment characteristics, assessed by the DPCCQ, predict differently the process and outcomes of psychoanalysis versus long-term psychodynamic therapy in the treatment of depressive and anxiety disorders during a 5-year follow-up.

2. Subjects and methods 2.1. Patients and settings A total of 506 eligible outpatients were recruited to the Helsinki Psychotherapy Study (HPS) from psychiatric services in the Helsinki region from June 1994 to June 2000 [32]. Eligible patients

were 20–45 years of age and had a long-standing disorder causing work dysfunction. They had to meet DSM-IV criteria [4] for anxiety or mood disorders. Patients with psychotic disorder, severe personality disorder (DSM-IV cluster A personality disorder and/ or lower level borderline personality organization), adjustment disorder, substance abuse or organic disorder were excluded, as were individuals who had undergone psychotherapy within the previous 2 years, psychiatric health employees, and persons known to the research team. Of the 506 patients referred to the HPS, 139 refused to participate. Of the remaining 367 patients, 128 were randomized to long-term psychodynamic psychotherapy (LPP) as part of a clinical trial comparing short- and long-term therapies, and 41 were self-selected to psychoanalysis (PA) [34]. After assignment to a treatment group, participation was refused by 26 patients assigned to LPP, and one assigned to PA. Of the 142 patients starting the assigned therapy, a total of 21 patients in LPP and 5 in PA discontinued the treatment prematurely, but were retained for secondary analyses. The mean length of therapy was 31.3 (SD = 11.9) and 56.3 (SD = 21.3) months in the two treatment groups, respectively. The patients were to be monitored for 10 years following the start of the treatment. Written informed consent was obtained from the patients after giving them a complete description of the study. The HPS follows the Helsinki Declaration and was approved by the ethics council of the Helsinki University Central Hospital. 2.2. Treatments Details of the treatments have been published [32,34]. Briefly, LPP is an open-ended, intensive, transference-based therapeutic approach which helps patients by exploring and working through a broad range of intrapsychic and interpersonal conflicts. Therapy includes both expressive and supportive elements, depending on the patient’s needs. The orientation followed the clinical principles of LPP [16]. The frequency of sessions in LPP was 2–3 times a week for approximately 3 years. PA is an open-ended, highly intensive, transference-based psychodynamic therapeutic approach, which helps patients by analyzing and working through a broad area of intrapsychic and interpersonal conflicts. The therapeutic setting and technique are characterized by facilitating maximum development of transference by the use of a couch and free association for exploring unconscious conflicts, developmental deficits, and distortions of intrapsychic structures [21]. The frequency of sessions in PA was four times a week for approximately 5 years. 2.3. Therapists Psychotherapeutic societies representing the treatments of interest were informed of the HPS, leading to a total of 112 eligible therapists volunteering for the study. Eligible therapists were required to have at least 2 years of experience in relevant therapy after completion of their training. A total of 41 therapists who did not have room for new patients or for some other reason could not attend to clients at the beginning of the study were excluded, as were six therapists who provided only solution-focused therapy and six who provided only short-term psychodynamic therapy as part of the original HPS design. If the patient was treated by more than one therapist, only the therapist information from the therapist who treated him or her the longest was used in the analysis. Therefore, one more therapist who provided LPP, and whose only patient changed to another therapist early in the treatment, was excluded from the present study. The final therapist population in this study thus comprised 58 therapists of whom 28 provided LPP, 18 PA, and 12 both LPP and PA. The

Please cite this article in press as: Heinonen E, et al. Therapists’ professional and personal characteristics as predictors of outcome in long-term psychodynamic psychotherapy and psychoanalysis. European Psychiatry (2013), http://dx.doi.org/10.1016/ j.eurpsy.2013.07.002

G Model

EURPSY-3089; No. of Pages 10 E. Heinonen et al. / European Psychiatry xxx (2013) xxx–xxx

3

caseload of the therapists varied between 1 and 7 patients (M = 2.5, SD = 1.5, mode = 2, median = 2). All the therapists providing LPP had received standard training in psychoanalytically oriented psychotherapy, approved by one of the psychoanalytic or psychodynamic training institutes in Finland. Likewise, psychoanalysts providing PA had received standard training at a psychoanalytic training institute. During their training, the psychodynamic therapists received a minimum of 3 years’ training in psychodynamic psychotherapy and analysts a minimum of 4 years’ training in psychoanalytic treatment. The treatments were conducted in accordance with clinical practice, where the therapists might modify their interventions according to the patient’s needs within the respective framework. Accordingly, no manuals were used and no adherence monitoring was organized.

scores, where pre-determined categorical cut-off points existed (Orlinsky and Rønnestad: personal communication). This was done to avoid potential biases resulting from the linearity assumption inherent in the use of continuous variables [7]. Due to the skewness of distributions and the resulting unbalanced tertiles, three of the original DPCCQ lower-level factors and four of the single-item questions that constituted the broad composite scores were excluded from being used as predictors.

2.4. Measures

2.5. Outcome

2.4.1. Therapist questionnaire The Development of Psychotherapists Common Core Questionnaire (DPCCQ), filled prior to initiating the treatments, was used as the predictor [44]. It is a comprehensive self-administered questionnaire soliciting information on therapists’ self-experiences in their professional and personal lives. In the professional domain, it covers aspects such as clinicians’ experiences of their skills, feelings, difficulties and coping strategies in therapy work as well as their interpersonal agency and manner with clients. Regarding private life, it assesses therapists’ self-experiences of their interpersonal and temperamental qualities in close personal relationships. Reliable multi-item scales have been constructed for assessing characteristics in these domains, also used in the present study; before this, the factor structure in the HPS therapist data was compared to and found in accordance with Orlinsky and Rønnestad’s analyses [44], to ensure the validity of the translation and the applicability of the scales (data not shown). Predetermined indicators consisted in the professional domain of composite scores Healing Involvement and Stressful Involvement and, in the domain of personal life, of composite scores Genial, Forceful and Reclusive. Briefly, Healing Involvement is characterized by therapists’ experience of investment in their work (Invested), affirmative manner with patients (Affirming), sense of basic relational and overall skillfulness (Basic Relational Skills, Current Skillfulness) and efficacy (Efficacious), and constructive coping (Constructive Coping Skills). Stressful Involvement, in turn, is characterized by therapists’ experience of frequent difficulties (Frequent Difficulties), feelings of anxiety (Anxiety) or boredom (Boredom), and coping with difficulties by avoidance of issues (Avoidant Coping Skills). Healing Involvement and Stressful Involvement have been respectively posited as determinants of more and less effective therapy processes [44]. In the personal domain, a Genial personal identity consists of experiences of warmth (Warm), openness (Open) and optimism in private relationships, Forcefulness of intense (Intense) and task-oriented (Task-Oriented) temperament and assertive (Assertive) interpersonal manner, and Reclusiveness of aloofness (Aloof), skepticism, privateness, and subtleness. Both the broad composite scores and their lower-level constituents have been found predictive of therapy alliances and outcomes [23,24,41,42], although the broad composite scores are perhaps not as significant as originally hypothesized. Hence, both the composite scores and their aforementioned lower-level constituents were investigated as predictors to gain the most detailed picture of the potentially important therapist characteristics. For the analyses, therapists were classified into ‘low’, ‘middle’, and ‘high’ groups of having more or less of a characteristic, based on therapists’ factor scores on the aforementioned dimensions, divided into tertiles, except for the personal identity composite

The Global Severity Index (GSI) of the Symptom Check List (SCL90), a 90-item self-report questionnaire [12] describing the severity of general psychiatric symptoms during the past month, was assessed as the outcome measure at 12, 24, 36, 48 and 60 months after baseline assessment.

2.4.2. Measures of background variables Further data on patients’ and therapists’ sociodemographic data were assessed using questionnaires. Patients’ previous psychiatric treatment was assessed by linking the study population to nationwide health registers [32].

2.6. Statistical methods The design of this study was longitudinal with repeated measurements of the psychiatric status of the patients at predetermined time points during the follow-up. The statistical analyses were based on linear mixed models (i.e., a model containing both fixed effects and random effects) [29,60]. The primary analyses were based on the assumption of ignorable dropouts [29,33] and in secondary analyses, missing values were replaced by multiple imputation. The imputation was based on the Markov chain Monte Carlo methods [47]. Model adjusted outcome means and mean differences were calculated for different measurement points [36]. The delta method was used for the calculation of confidence intervals [40] and the statistical significance of the model used was tested with the Wald test. The statistical analyses were carried out with SAS software, version 9.2 [53]. The dependent variable (outcome variable) in all analyses was SCL-90-GSI. In the analyses, three models were considered: a basic model included as independent variables the main effects of the therapist measure considered (categorical), therapy groups compared (categorical), and time (categorical), their first- and second-order interactions, a correction term (i.e. the first-order interaction of the difference between theoretical and realized date of measurement, time and the therapist measure), and SCL-90-GSI at baseline (continuous). Of these, the correction term ensured comparability of the time-dependent outcome across patients, as it adjusted for the between-patients differences regarding the specific times of outcome measurement. Also, as the data were not randomized with regard to the therapist factor, the baseline symptom level was a potentially confounding factor that differed between the groups of the predictors, and thus needed to be adjusted for to avoid bias. While therapist background factors (profession, age, level of training, and years of experience) were initially also considered as potential confounders, they did not satisfy the criteria for confounding (range of p-values for the significance of the correlations between these potential confounding factors and outcome were found to be between 0.29 and 0.80), and thus a second model with confounding factors was not run [46]. A third model was created by adding variables describing waiting time from randomization to initiation of treatment and degree of participation (i.e. withdrawal from or discontinuation of treatment) during follow-up as main effects to the complete model, and including any auxiliary treatment (hospitalization,

Please cite this article in press as: Heinonen E, et al. Therapists’ professional and personal characteristics as predictors of outcome in long-term psychodynamic psychotherapy and psychoanalysis. European Psychiatry (2013), http://dx.doi.org/10.1016/ j.eurpsy.2013.07.002

G Model

EURPSY-3089; No. of Pages 10 4

E. Heinonen et al. / European Psychiatry xxx (2013) xxx–xxx

psychotropic medication, other psychotherapy) as a time-dependent covariate. The therapist was included as a random effect in all models. The models were based both on the original and the imputed data. The independent variable of main interest (predictor) was the interaction term between the therapist factor, therapy group, and time. Since no major differences were found in this variable during follow-up between the different models, and since imputation, carried out to study the reliability of the ignorable drop-out assumption, attenuated somewhat but did not noticeably alter the results, the results from the basic model were presented. The significance of the therapist factor in predicting the outcome of long-term psychodynamic therapy versus psychoanalysis during the 5-year follow-up was evaluated by testing the statistical significance of the interaction term between the therapist factor and the therapy group throughout the followup. The Wald test was used. For the interpretation of the results, outcomes in LPP vs. PA were compared as a function of the level of therapist characteristics: LPP more efficient (prognosis group 1), PA more efficient (prognosis group 2), and both of the compared treatments equally efficient (prognosis group 3). This was done by assessing the statistical significance of the change in outcome from baseline to the different measurement points for each therapy group and category (low and high) of therapist factor. Therapy was considered to be sufficient for the patients who experienced and maintained a statistically significant reduction in symptoms in comparison with those at baseline during the 5-year follow-up; patients who did not experience such a symptom reduction were not considered to benefit sufficiently from therapy. To separate prognosis groups 1 and 2 from each other, the statistical significance of the model-adjusted difference in outcome between the therapy groups in the therapist factor categories was measured at the different measurement points. Long-term psychodynamic therapy was considered to work more efficiently in conjunction with particular (‘low’ or ‘high’) therapist qualities when patients benefited significantly more from long-term psychodynamic therapy (prognosis group 1), whereas psychoanalysis was considered more efficient when patients gained significantly more from psychoanalysis (prognosis group 2). In the remaining cases, when no significant differences emerged at any time during the follow-up, prognosis group 3 was applied. Finally, when prognosis groups differed between the ‘low’ and ‘high’ categories or when the global term of interaction between therapy group and therapist factor proved significant, a differential effect of therapist characteristic on LPP vs. PA was indicated. 3. Results 3.1. Description of study population The predominant patient problems were mood disorders, with close to half of the sample also suffering from co-morbid other Axis I or personality disorders, and roughly a fourth to third from comorbid anxiety disorders (Table 1). The patients’ background characteristics in the two treatments were quite similar, the biggest difference being that the patients in PA were more often academically educated. Treatments were provided by relatively experienced therapists, with no notable differences in their background characteristics or their DPCCQ composite scores across the LPP and PA conditions. 3.2. Prediction of therapist characteristics At the 5-year follow up, a statistically significant symptom reduction was found for all patient groups, irrespective of the characteristics of therapists they were seeing, in both LPP and PA,

with one exception (low Affirming manner in LPP) (Table 2). There was one statistically significant (Aloof, p = 02) and two borderline significant (Affirming, p = .06, Forceful, p = .07) interactions between therapist characteristics and therapy group. In addition, numerous consistent statistically significant differences in symptom development between LPP and PA at different points of followup were found in relation to the level of therapist characteristics which are described below in more detail. 3.2.1. Prediction of professional characteristics With therapists who rated high Stressful Involvement, Frequent Difficulties, Boredom, Constructive Coping and Flow, patients who were treated in the LPP rather than the PA condition experienced statistically significantly less symptoms at the 36-month followup point (range of SCL-90-GSI 0.56–0.70 and 0.92–1.25, for the two conditions respectively, among the ‘high’ tertiles of these therapist characteristics), around the end of the LPP treatments (Table 2). Among therapists of higher Current Skillfulness, a similar significant difference for the benefit of patients in LPP as contrasted to PA was observed at the 48-month follow-up (SCL-90-GSI 0.57 and 0.88, respectively). None of these differences remained significant at the 60-month follow-up, however, around the end of the PA treatments. However, at the 5-year follow-up significantly less symptoms were noted for the patients treated in the PA as compared to the LPP condition with therapists who rated being low in Affirming relational manner (SCL-90-GSI 0.27 and 0.90, respectively). 3.2.2. Prediction of personal characteristics With therapists who rated higher Geniality, lower Forcefulness or higher Aloofness, patients experienced significantly less symptoms when treated in the LPP rather than the PA condition at the 36-month follow-up point (range of SCL-90-GSI 0.62–0.69 and 0.95–1.00, for the two conditions, respectively). In contrast, patients with therapists who rated lower Aloofness experienced less symptoms when treated in the PA rather than the LPP condition at the 60-month follow-up (SCL-90-GSI 0.32 and 0.88, respectively). With therapists who rated a less Intense temperament or a more Assertive interpersonal manner, significantly less symptoms were experienced by patients treated in the LPP rather than the PA condition at the 36-month follow-up point (range of SCL-90-GSI 0.61–0.66 and 0.97–1.01, respectively); however, these differences did not remain significant at the 5-year follow-up. 4. Discussion The present study showed both therapists’ professional and personal pre-treatment characteristics to predict the level of psychiatric symptoms differently in PA and LPP during a 5-year follow-up. The results converge with some earlier findings [49], complement extant studies in process and outcome research [1,20,24,49], and add to the theoretical discussion on beneficial therapist qualities in LPP and PA [54]. While the results may raise more questions than provide conclusive answers, we offer some tentative interpretations and outline perspectives for future investigations, along with noting the study’s limitations. First, the choice of general psychiatric symptoms (SCL-90-GSI) as the outcome measure needs to be noted. This choice was based on the desire to compare our findings to the earlier ones by Sandell et al. [49]; the moderate to strong association of the SCL-90-GSI (which covers, e.g., depressive, anxiety and interpersonal symptoms) to other standard outcome measures assessing global psychopathology [55] and general psychiatric symptoms [43]; as well as the measure’s sensitivity to change [28,49]. While psychiatric symptoms may likely be less of a direct focus of intervention in PA and LPP than in some other treatments, they

Please cite this article in press as: Heinonen E, et al. Therapists’ professional and personal characteristics as predictors of outcome in long-term psychodynamic psychotherapy and psychoanalysis. European Psychiatry (2013), http://dx.doi.org/10.1016/ j.eurpsy.2013.07.002

G Model

EURPSY-3089; No. of Pages 10 E. Heinonen et al. / European Psychiatry xxx (2013) xxx–xxx

5

Table 1 Mean (SD) levels of baseline characteristics of the patients and the therapists. Characteristic

All

Long-term psychodynamic psychotherapy

Psychoanalysis

Patient characteristics Socioeconomic variables Age (years) Males (%) Living alone (%) Academic education (%) Psychiatric diagnosis Axis I Mood disorder only (%) Anxiety disorder only (%) Co-morbidity of mood and anxiety disorders (%) Any Axis II disorder (%) Psychiatric co-morbidity on Axis I or with Axis II disorders (%) Previous psychiatric treatment Psychotherapy (%) Psychotropic medication (%) Hospitalization (%) Psychiatric symptoms Symptom Check List, Global Severity Index, (SCL-90-GSI) Therapist characteristics Demographics Age (years) Males (%) Training Psychologist (%) Psychiatrist (%) Other (%) General therapy experience (years) Specific therapy experience (years) Work involvement Healing involvement Stressful involvement Personal identity Genial Forceful Reclusive

N = 169

N = 128

N = 41

31.3 (6.4) 23.7 52.1 32.5

31.6 (6.6) 21.1 49.2 28.1

30.4 (5.6) 31.7 61.0 46.3

62.7 10.7 26.6 14.2 39.6

63.3 11.7 25.0 12.5 36.7

61.0 7.3 31.7 19.5 48.8

21.0 15.2 2.4

19.1 17.6 3.1

26.8 7.7 0.0

1.29 (0.55) N = 58

1.27 (0.55) N = 40

1.34 (0.52) N = 30

49.9 (5.8) 32.2

49.7 (5.6) 24.4

51.1 (5.7) 36.7

74.6 13.6 11.9 18.1 (5.4)

80.5 7.3 12.2 18.0 (5.6) 17.6 (5.6)

73.3 16.7 10.0 19.3 (5.0) 14.8 (5.4)

11.19 (1.12) 3.87 (1.32)

11.16 (0.98) 3.84 (1.32)

11.21 (1.28) 4.02 (1.39)

2.18 (0.32) 1.64 (0.38) 1.16 (0.49)

2.21 (0.32) 1.62 (0.34) 1.16 (0.52)

2.14 (0.30) 1.66 (0.38) 1.25 (0.49)

have been found to be the most sensitive to change even in these treatments and may plausibly be taken as potential indicators of deeper structural change [49]. The findings of our study are reviewed next. It was found that the end results of the two treatments differed noticeably and systematically at the 5-year follow-up, when investigated as a function of therapists’ professional and personal relational qualities. Patients in PA had significantly less symptoms than ones in LPP with therapists who experienced themselves in professional manner as less Affirming. Also, less symptoms in PA than in LPP were observed with practitioners who experienced themselves in personal relationships as less Aloof and more Forceful. One of these corresponding global interaction terms for treatment group and therapist characteristic was significant and two were borderline significant, indicating a robust association. As being less Affirming (i.e., less ‘accepting’, ‘friendly’, ‘tolerant’, and ‘warm’) [44] in professional relational manner predicted a relatively high symptom level in LPP throughout the 5-year follow-up, our results seemed to support Sandell et al.’s [49] earlier findings of therapists low on kindness and supportiveness being especially unbeneficial in LPP. The other two qualities that proved especially beneficial in PA – high Forcefulness (i.e., highly ‘assertive’, ‘authoritative’, ‘demanding’, ‘directive’, ‘challenging’, ‘critical’, ‘pragmatic’, ‘organized’, ‘demonstrative’, ‘determined’, ‘energetic’, ‘intuitive’, ‘intense’) and low Aloofness (i.e., low on ‘cold’, ‘guarded’, and ‘reserved’) – may also be noted to clearly not embody the stereotypical ‘‘detached’’ psychoanalyst which arguably was in fact not advocated by Freud [38,58]. It has been suggested instead that Freud merely advocated that analysts should confine their personal characteristics to the service of building the working alliance or a positive transference in a

realistic sense [38]. Indeed, taken together, our current findings suggest a pattern in which patients do best in PA when their analysts are restrained in showing affirmation professionally but are personally highly ‘present’ (i.e., Forceful and not Aloof in close personal relationships). That such a dynamics between the professional and the personal might not only be beneficial, but also not an atypical pattern in psychoanalysts was suggested by a recent international study of over 4000 therapists [25]. Using the same DPCCQ therapist instrument as the present study, it was shown that clinicians tend to adjust their professional relational style with clients from their manner in private life according to the expectations of their theoretical approaches – that is, either to ‘‘tone up’’ the qualities recommended by their theoretical orientations or ‘‘tone down’’ personal qualities that are viewed as detrimental in those frameworks. Thus, in contrast to the stereotype of the neutral analyst, psychoanalytic and psychodynamic therapists were, for example, found to experience themselves as significantly more warm and friendly, but also more directive and demanding in their private life than in their professional manner, as compared to the magnitude of differences in therapists endorsing some other theoretical orientations, such as cognitive-behavioral or humanistic approaches. Naturally, given the present findings that being less Aloof and more Forceful in close personal relationships predicted notably good outcomes in PA, an interesting question that still remains is: how exactly do these experiences in personal relationships then carry over to therapists’ professional work? While this question merits further investigation, we note that it has been argued that outcomes might be served by replacing the traditional ideal of a neutral analyst with a conception of ‘‘psychoanalysis plus’’ that involves a more supporting and encouraging therapist [54]. With our current data,

Please cite this article in press as: Heinonen E, et al. Therapists’ professional and personal characteristics as predictors of outcome in long-term psychodynamic psychotherapy and psychoanalysis. European Psychiatry (2013), http://dx.doi.org/10.1016/ j.eurpsy.2013.07.002

12

0.84

36

48

0.68

0.61

60

0.69

N

Groupa 12

32 3

0.98

24

0.78

36

48

0.71

0.71

0.95 0.96 0.10 0.28 ( 0.44,0.23) ( 0.65,0.10) 0.80 0.70

0.66 0.54 13 0.04 0.14 ( 0.36,0.28) ( 0.23,0.52) 0.75 0.79 26 3

1.08 0.10 ( 0.39,0.18) 0.82

0.89 0.11 ( 0.42,0.21) 0.76

0.96 0.24 ( 0.58,0.09)

0.97 0.83 0.17 0.13 ( 0.52,0.17) ( 0.50,0.24) 0.86 0.77

0.83 0.72 14 0.08 0.07 ( 0.41,0.26) ( 0.30,0.43) 0.68 0.72 35 3

1.04 0.22 ( 0.48,0.04) 0.89

0.92 0.16 ( 0.47,0.16) 0.85

1.16 0.60 ( 0.95,

0.75 0.83 0.10 0.06 ( 0.26,0.46) ( 0.45,0.33) 0.82 0.75

0.64 0.54 10 0.03 0.18 ( 0.30,0.37) ( 0.21,0.57) 0.69 0.62 39 3

1.17 1.02 0.28 0.17 ( 0.56,0.001) ( 0.50,0.16) 0.94 0.89

0.93 0.30 ( 0.64,0.04)

0.67 0.72 0.16 0.03 ( 0.18,0.50) ( 0.32,0.38) 0.80 0.63

0.60 0.46 12 0.10 0.16 ( 0.21,0.40) ( 0.21,0.52) 0.51 0.54 38 1

1.20 0.25 ( 0.54,0.04) 0.97

1.04 0.15 ( 0.48,0.18) 0.85

0.93 0.26 ( 0.61,0.08)

0.72 0.63 13 0.21 0.09 0.02) ( 0.51,0.09) ( 0.44,0.26) 0.81 0.90 12 2

1.20 0.22 ( 0.51,0.06) 1.01

1.07 0.23 ( 0.54,0.09) 0.83

0.84 1.01 0.04 0.37 ( 0.37,0.30) ( 0.73, 1.00 0.95

60

0.56

0.26)

0.63

0.67

0.78

N

Groupa Inter- Effectc actionb

31 3

0.77 0.06 ( 0.35,0.23) 0.52

0.69 17 0.09 ( 0.24,0.42) 0.39 31 1

0.78 0.25 ( 0.56,0.06) 0.57

0.56 17 0.17 ( 0.52,0.18) 0.61 38 1

0.88 0.32 ( 0.62, 0.63

0.77 15 0.17 ( 0.50,0.17) 0.68 42 3

0.01)

0.72

0.93 0.30 ( 0.62,0.01) 0.70

0.79 14 0.11 ( 0.46,0.24) 0.71 30 1

1.08 0.37 ( 0.70, 0.73

0.98 0.28 ( 0.58,0.03) 0.72

0.72 16 0.01 ( 0.36,0.34) 0.76 45 1

0.03)

1.00 0.64 0.00 0.31 ( 0.54,0.55) ( 0.15,0.77) 0.83 0.74 0.97 0.91 0.14 0.17 ( 0.47,0.18) ( 0.51,0.18) 0.80 0.70

0.51 0.27 8 0.30 0.63 ( 0.11,0.71) (0.13,1.13) 0.64 0.75 32 3 0.62 0.55 15 0.02 0.20 ( 0.29,0.33) ( 0.17,0.56) 0.63 0.71 29 3

1.25 0.24 ( 0.50,0.02) 1.03 1.07 0.04 ( 0.35,0.26) 1.01

1.15 1.25 0.32 0.52 ( 0.62, 0.03) ( 0.83, 0.22) 0.81 0.70 0.91 1.25 0.10 0.55 ( 0.47,0.27) ( 0.91, 0.18) 0.84 0.58

0.93 0.21 ( 0.49,0.07) 0.70 0.92 0.22 ( 0.54,0.10) 0.50

0.77 0.01 ( 0.33,0.31) 0.71 0.81 0.10 ( 0.47,0.27) 0.44

0.87 1.05 0.07 0.35 ( 0.44,0.30) ( 0.75,0.05) 0.77 0.66

0.74 0.63 12 0.11 0.08 ( 0.48,0.25) ( 0.33,0.49) 0.52 0.65 27 3

1.11 0.10 ( 0.48,0.28) 0.94

0.93 0.09 ( 0.49,0.31) 0.85

1.03 0.45 ( 0.89, 0.003) 0.78

0.75 0.25 ( 0.63,0.13) 0.72

0.42 12 0.02 ( 0.44,0.48) 0.64 40 3

0.91 0.64 0.14 0.03 ( 0.58,0.31) ( 0.47,0.52) 0.96 0.82

0.59 0.49 7 0.07 0.16 ( 0.51,0.38) ( 0.35,0.67) 0.79 0.83 31 3

0.98 0.04 ( 0.28,0.20) 0.88

0.77 0.08 ( 0.19,0.35) 0.77

0.95 0.17 ( 0.48,0.14) 0.59

0.62 0.10 ( 0.15,0.35) 0.52

0.49 22 0.16 ( 0.16,0.47) 0.50 45 1

0.87 0.86 0.09 0.04 ( 0.29,0.47) ( 0.46,0.37)

0.62 0.58 10 0.17 0.25 ( 0.20,0.53) ( 0.16,0.66)

0.91 0.03 ( 0.26,0.20)

0.81 0.04 ( 0.31,0.23)

0.92 0.33 ( 0.62, 0.04)

0.64 0.12 ( 0.37,0.14)

0.47 20 0.03 ( 0.26,0.33)

0.89

NO

0.37

YES

0.38

YES

0.10

NO

0.44

NO

0.06

YES

0.44

YES

0.39

YES

0.30

NO

0.76

YES

15

35 1 11

14 1

E. Heinonen et al. / European Psychiatry xxx (2013) xxx–xxx

Professional characteristics LPP Healing 0.95 involvement 1.11 PA LPP-PAd 0.16 ( 0.46,0.12) 1.00 Stressful LPP involvement 1.13 PA LPP-PAd 0.14 ( 0.44,0.17) 1.03 Current LPP skillfulness 0.95 PA LPP-PAd 0.09 ( 0.25,0.42) LPP 0.96 Basic relational skills 0.88 PA LPP-PAd 0.08 ( 0.21,0.37) 0.93 Efficacious LPP agency PA 1.10 LPP-PAd 0.17 ( 0.45,0.12) Affirming LPP 1.08 manner 0.82 PA LPP-PAd 0.26 ( 0.21,0.74) 0.99 Feelings: Flow LPP PA 1.09 LPP-PAd 0.10 ( 0.40,0.20) 1.06 Feelings: LPP Boredom PA 1.05 LPP-PAd 0.01 ( 0.32,0.35) 0.99 Feelings: LPP Anxiety 1.02 PA LPP-PAd 0.03 ( 0.42,0.37) 1.09 Constructive LPP coping skills 1.03 PA LPP-PAd 0.06 ( 0.27,0.40)

High 24

G Model

Therapy Low

EURPSY-3089; No. of Pages 10

Characteristic

6

Please cite this article in press as: Heinonen E, et al. Therapists’ professional and personal characteristics as predictors of outcome in long-term psychodynamic psychotherapy and psychoanalysis. European Psychiatry (2013), http://dx.doi.org/10.1016/ j.eurpsy.2013.07.002

Table 2 Mean values of SCL-90-GSI in long-term psychodynamic therapy and psychoanalysis and mean value differences (95% confidence intervals) at 12, 24, 36, 48 and 60 month follow-up according to the lower and higher values of therapist professional and personal characteristics.

Table 2 (Continued )

Avoidant LPP coping skills PA LPP-PAd Frequent difficulties

LPP PA LPP-PAd

Personal characteristics Geniale LPP PA LPP-PAd e

Forceful

Reclusivef

Interpersonal: Warm

LPP PA LPP-PAd LPP PA LPP-PAd LPP PA LPP-PAd

Interpersonal: Assertive

LPP PA LPP-PAd

Interpersonal: Aloof

LPP PA LPP-PAd

Temperament: LPP Intense PA LPP-PAd

1.03

High 24 0.75

36 0.68

48 0.68

60 0.68

N

Groupa 12

21 3

0.85

24 0.79

36 0.55

48

60

0.46

0.37

N

Groupa Inter- Effectc actionb

36 3

1.23 0.21 ( 0.52,0.11) 1.01

1.10 0.93 0.34 0.25 ( 0.79,0.01) ( 0.63,0.14) 0.82 0.69

0.87 0.81 13 0.18 0.13 ( 0.51,0.15) ( 0.50,0.24) 0.69 0.73 32 3

1.06 0.21 ( 0.51,0.09) 0.84

0.81 0.02 ( 0.36,0.32) 0.78

0.88 0.32 ( 0.68,0.04) 0.57

0.63 0.17 ( 0.48,0.14) 0.56

0.39 11 0.02 ( 0.39,0.35) 0.47 34 1

1.12 0.11 ( 0.39,0.17)

0.91 0.90 0.08 0.21 ( 0.43,0.26) ( 0.55,0.14)

0.75 0.60 14 0.06 0.14 ( 0.36,0.24) ( 0.21,0.48)

0.98 0.14 ( 0.42,0.14)

0.87 0.10 ( 0.43,0.24)

1.14 0.57 ( 0.95, 0.19)

0.79 0.23 ( 0.56,0.11)

0.61 13 0.13 ( 0.50,0.23)

0.97 1.16 0.19 ( 0.50,0.12) 0.97 1.09 0.11 ( 0.31,0.09) 1.03 0.97 0.06 ( 0.23,0.34) 0.97

0.80 0.72 0.93 0.97 0.13 0.24 ( 0.50,0.43) ( 0.62,0.13) 0.86 0.69 0.95 0.95 0.09 0.26 ( 0.31,0.14) ( 0.51, 0.01) 0.89 0.77 0.84 0.95 0.05 0.18 ( 0.30,0.40) ( 0.51,0.15) 0.84 0.74

0.63 0.60 0.91 0.82 0.28 0.22 ( 0.61,0.05) ( 0.62,0.17) 0.64 0.64 0.82 0.71 0.19 0.07 ( 0.40,0.03) ( 0.32,0.18) 0.77 0.79 0.74 0.51 0.03 0.28 ( 0.27,0.32) ( 0.06,0.62) 0.65 0.63

22 3

0.98 1.01 0.04 ( 0.24,0.17) 1.05 1.00 0.05 ( 0.29,0.39) 0.94 1.10 0.16 ( 0.38,0.07) 1.00

0.87 0.89 0.02 ( 0.25,0.21) 0.81 0.78 0.04 ( 0.37,0.44) 0.82 0.94 0.12 ( 0.37,0.13) 0.82

0.69 0.96 0.27 ( 0.52, 0.02) 0.73 1.11 0.38 ( 0.78,0.03) 0.65 0.93 0.28 ( 0.57,0.02) 0.72

0.65 0.65 0.00 ( 0.23,0.23) 0.67 0.56 0.11 ( 0.28,0.50) 0.56 0.76 0.20 ( 0.44,0.05) 0.68

0.67 0.48 0.19 ( 0.08,0.45) 0.68 0.35 0.33 ( 0.09,0.74) 0.56 0.67 0.11 ( 0.39,0.17) 0.67

1.14 0.16 ( 0.51,0.19) 1.05

0.98 0.97 0.15 0.23 ( 0.57,0.28) ( 0.63,0.18) 0.91 0.77

0.97 0.88 10 0.32 0.25 ( 0.68,0.05) ( 0.68,0.17) 0.74 0.79 36 3

1.06 0.06 ( 0.34,0.22) 0.88

0.87 0.05 ( 0.35,0.26) 0.82

0.94 0.22 ( 0.56,0.12) 0.61

0.56 0.12 ( 0.20,0.44) 0.54

0.55 13 0.12 ( 0.24,0.47) 0.53 41 1

0.87 0.18 ( 0.13,0.49) 1.10

0.86 0.80 0.05 0.03 ( 0.30,0.39) ( 0.40,0.34) 0.92 0.81

0.59 0.48 11 0.15 0.32 ( 0.18,0.48) ( 0.06,0.69) 0.86 0.88 30 2

1.09 0.21 ( 0.46,0.05) 0.90

0.89 0.07 ( 0.37,0.23) 0.76

1.01 0.40 ( 0.71, 0.09) 0.62

0.74 0.20 ( 0.48,0.08) 0.50

0.44 16 0.09 ( 0.23,0.40) 0.57 39 1

0.59 0.51 (0.15,0.87) 0.93

0.40 0.51 (0.03,1.00) 0.85

0.46 0.40 (0.03,0.78) 0.60

1.09 0.19 ( 0.42,0.05) 1.01

0.91 0.15 ( 0.42, 0.13) 0.85

1.00 0.38 ( 0.70, 0.06) 0.80

0.76 0.26 ( 0.52, 0.74

0.91 0.10 ( 0.19,0.38)

0.68 0.17 ( 0.24,0.59)

0.98 0.19 ( 0.54,0.17)

0.67 0.08 ( 0.25,0.40)

0.98 0.17 ( 0.72,0.39) 0.66

0.99 0.97 1.16 0.24 0.14 0.31 ( 0.47, 0.01) ( 0.41,0.12) ( 0.61, 0.01)

0.32 0.56 (0.09,1.02) 0.58

26 3 13

74 1 28

40 3 15

8

58 1

0.83 0.72 18 0.23 0.14 ( 0.49,0.02) ( 0.44,0.16)

0.11

NO

0.54

YES

0.21

YES

0.07

YES

0.47

NO

0.25

NO

0.18

YES

0.02

YES

0.26

YES

72 1 25

24 3 10

58 3 23

46 3

0.61 18 0.04 0.003) ( 0.35,0.27) 0.77 27 3 0.42 14 0.36 ( 0.01,0.72)

7

Underlined symptoms have changed statistically significantly since 0 months symptom status. a Prognosis groups. 1: therapist characteristic efficient especially in long-term psychodynamic therapy. 2: therapist characteristic efficient especially in psychoanalysis. 3: therapist characteristic not differentially efficient in longterm psychodynamic psychotherapy or psychoanalysis. b p-value for global test of interaction between therapist factor and therapy length. c YES: significant global test of interaction or prognoses different for therapists with low and high values of a characteristic, i.e. indicating a differential effect of therapist characteristic on outcome in long-term psychodynamic therapy and psychoanalysis. NO: no significant global test of interaction and prognoses same for therapists with low and high values of a characteristic, i.e. no differential effect of therapist characteristic. d Mean value difference of SCL-90-GSI between long-term psychodynamic psychotherapy and psychoanalysis. e In consideration of pre-determined categorical cut-off points and skewness of distributions, ‘high’ therapists consisted of strongly genial and forceful therapists with values over 2 on the factor dimension and ‘low’ therapists consisted of moderately and minimally genial and forceful therapists with values under 2 on the 0–3 scale. f In consideration of pre-determined categorical cut-off points and skewness of distributions, ‘highly’ reclusive therapists consisted of strongly and moderately reclusive therapists with values over 1 on the factor dimension and ‘low’ therapists of minimally reclusive therapists with values under 1.

G Model

12

EURPSY-3089; No. of Pages 10

Therapy Low

E. Heinonen et al. / European Psychiatry xxx (2013) xxx–xxx

Please cite this article in press as: Heinonen E, et al. Therapists’ professional and personal characteristics as predictors of outcome in long-term psychodynamic psychotherapy and psychoanalysis. European Psychiatry (2013), http://dx.doi.org/10.1016/ j.eurpsy.2013.07.002

Characteristic

G Model

EURPSY-3089; No. of Pages 10 8

E. Heinonen et al. / European Psychiatry xxx (2013) xxx–xxx

we may only hypothesize that a therapist more strongly engaged in personal life might have less inhibitions in also engaging the patient in the analytic process beyond a classical neutral manner, as indicated by Schachter and Ka¨chele [54]. Also, while it was a less Affirming professional manner that predicted good outcomes in PA, possibly the fact that patients in the PA condition were seen four times per week might in itself represent a sufficient experience of affirmation to patients so as to not need much other overt expression of acceptance, warmth, friendliness, and tolerance (i.e., components of the Affirming scale) [44] in PA [65]. Although our outcome criteria also indicated that therapists’ personal manner characterized by high Geniality, high Assertiveness, and low Intensity predicted a faster symptom reduction in LPP than in PA, the significant differences did not hold up until the end of the follow-up; also, the actual differences between the low and high ends of these dimensions were comparably small, not indicating a clear superiority of these qualities in either of the two treatments. Several professional qualities reflecting therapists’ experience of therapy work also predicted significant differences between the two treatments during the follow-up. Significantly less symptoms were observed in LPP as compared to PA at the 3-year – and, in one case, at the 4-year – follow-up points, as predicted by somewhat contradictory professional qualities: self-experiences of high Current Skillfulness (4-year follow-up point), high Constructive Coping Skills, and high Flow, but also of high Frequent Difficulties, high Boredom, and high Stressful Involvement. In the cases of high Flow, high Frequent Difficulties, and high Stressful Involvement, even something of a recurring peak in symptoms was observed in PA. However, although contradictory, these characteristics are not mutually exclusive and if considered together, they may be seen as largely approximating a therapist’s strong personal commitment to therapy work or, as described by developers of the DPCCQ, Orlinsky and Rønnestad [44], a work experience of ‘‘Challenging Practice’’: experienced by a practitioner ‘‘facing but apparently overcoming difficulties’’. Furthermore, in the end such therapists were not less effective in PA than in LPP, as no differences were observed at the 5-year follow-up. Indeed, one possible interpretation might be that a considerable symptom level would reflect an intentional ‘‘working through’’ of problems in ongoing psychoanalysis, plausibly associated with such a committed therapist [67]; it may likewise be that the symptom peaks represent a phase in the middle of the analytic process where the patient’s suffering and distress become more alive within the here-and-now therapeutic relationship, as the neurotic conflicts unfold, to be further confronted and understood by a committed analyst [63]. On the other hand, in LPP which ended at the 3-year follow-up point, a more appropriate end state would be the patient’s greater mastery of problems and lesser experience of symptoms [8]. The aforementioned interpretations would need more detailed research into the process of therapy to be confirmed. However, the present results also do mirror some earlier findings on therapists’ global self-experiences of professional skillfulness, confidence, and enjoyment as predictors of patient-rated therapy relationships and outcome in long-term psychodynamically oriented treatments. Studies using the DPCCQ, for example, have shown overall professional experiences of skillfulness and enjoyment versus stress and difficulties – that is, the broad second-order factors of Healing Involvement versus Stressful Involvement – to not predict better versus worse patient-rated alliances [23,41] or outcomes [24] in long-term psychodynamically oriented treatments as straightforwardly and obviously as perhaps might be hypothesized [44]. Other studies using different therapist measures have also failed to find associations between therapists’ professional lack of confidence and patient-rated outcomes in long-term psychodynamic treatments and psychoanalyzes [50,51]. When looked at

solely from the therapist’s viewpoint, however, therapists’ evaluations of better working alliances [23], better session flow and overall better session evaluation [72] are unequivocally predicted by their greater self-experienced skillfulness, confidence, and enjoyment in therapy work, which seem like expectable findings. These findings are challenged and complicated, however, by the paradoxical findings that therapists’ self-rated, more specific advanced relational skills have in fact predicted worse patient-rated alliances and outcomes, whereas professional selfdoubt has proved beneficial for both [41,42]. These complex associations suggest the need for more detailed explorations of how therapists’ perceptions of their own skillfulness are manifested in actual therapy practice, using not only therapist-, but also patient-and observer-rated perspectives on therapists and the impacts of their actions [23]. 4.1. Methodological considerations Several strengths were related to the present study. First, the relatively large therapist and patient sample with frequent assessments during a long follow-up allowed for a detailed investigation of potential differences from the start to the endpoints of both treatments. Second, therapists’ pre-treatment assessment allowed measuring their qualities independently of an ongoing treatment process and thus best indicated the independent contribution of what the therapists bring into treatment. Third, the results complement conceptually and methodologically the outcome literature that has compared these two treatments utilizing other therapist measures focused on therapists’ general treatment attitudes, likewise measurable prior to treatment [49], as well as process measures assessing therapist techniques utilized in actual treatments [1]. Specifically, as the DPCCQ assesses therapist characteristics not particular to any therapy theory or orientation, it may provide one valuable perspective into the generic beneficial, professional and personal therapist qualities [1,54]. However, there are also some methodological limitations. First, although sociodemographic characteristics (profession, age, level of training, and years of experience) were considered as potential confounders, confounding due to other, unmeasured therapist variables cannot fully be excluded, likewise, the investigated DPCCQ therapist qualities cannot be examined truly in isolation from each other, as they are intertwined with each other. However, the research strategy was chosen to gain as comprehensive a picture as possible of the potentially important – both second- and first-order – characteristics in LPP and PA. Second, it may be noted that the study focused on interpreting the fixed effects of therapist characteristics, while the therapist random effects were quite negligible (data not shown); greater between-therapist outcome variability reflected in random effects might present themselves for investigation in a sample and setting with greater heterogeneity and patient-to-therapist ratios [39]. Third, no adherence checks or treatment monitoring was conducted; however, this corresponded with the study’s aim to investigate normal clinical practice. Fourth, as some of the treatment effects in LPP versus PA may only become apparent years after their terminations [6,49], longer follow-ups are needed. Fifth, it is not clear how these DPCCQ characteristics manifest themselves in practice, how therapist qualities would be rated from patients’ and external observers’ perspectives, and what are the DPCCQ’s other correlates in the treatment process. However, initial findings are accumulating in these areas [22,72], which may facilitate interpreting the findings in the future. Finally, as the present study focused on psychiatric symptoms, future outcome studies assessing, e.g., self-concept or interpersonal functioning, could provide additional perspectives.

Please cite this article in press as: Heinonen E, et al. Therapists’ professional and personal characteristics as predictors of outcome in long-term psychodynamic psychotherapy and psychoanalysis. European Psychiatry (2013), http://dx.doi.org/10.1016/ j.eurpsy.2013.07.002

G Model

EURPSY-3089; No. of Pages 10 E. Heinonen et al. / European Psychiatry xxx (2013) xxx–xxx

5. Conclusion The present study showed both therapists’ professional and personal self-experiences to predict the treatment process and outcomes in PA and LPP. The results challenge the benefits of a ‘‘neutral’’ therapist in PA, in line with earlier theoretical writings [54]. Although these results need further exploration, they could, if confirmed, have implications for the training and supervision of therapists – for example, in helping to achieve the right balance between a professionally analytic and properly restrained, yet personally ‘present’ manner [25]. The findings also indicated that patients of therapists experiencing themselves skillful, coping constructively with problems and enjoying therapy practice or, contradictorily, experiencing stress and frequent difficulties could encounter significantly high symptom levels during the middle phases of PA as compared to LPP that would have ended approximately at that same time. As no differences were observed in the end at the 5-year follow-up, the meaning of this finding is not clear but it does suggest some further differences between the two treatments that interact with therapist qualities and supports closer examination of the processes in both treatments, using more nuanced process and outcome measures, as well as follow-ups extending further beyond the endpoints of both treatments. Disclosure of interest The authors declare that they have no conflicts of interest concerning this article. Acknowledgments The Helsinki Psychotherapy Study Group was responsible for collection of the data [32]. The study was financially supported by the Academy of Finland (grant No. 138876). The financial support of the Emil Aaltonen Foundation and the Psychiatric Research Foundation (Finland) to the first author is also gratefully acknowledged. References [1] Ablon JS, Jones EE. On analytic process. J Am Psychoanal Assoc 2005;53:541– 68. [2] Alexander F. Psychoanalysis and psychotherapy. J Am Psychoanal Assoc 1954;2:722–33. [3] Alexander F. Psychoanalysis and psychotherapy: developments in theory, technique, and training. New York, NY: Norton; 1956. [4] A.P.A.. Diagnostic and statistical manual of mental disorders. Washington, DC: American Psychiatric Association; 1994. [5] Blatt SJ, Shahar G. Psychoanalysis–with whom, for what, and how? Comparisons with psychotherapy. J Am Psychoanal Assoc 2004;52:393–447. [6] Blomberg J, Lazar A, Sandell R. Long-term outcome of long-term psychoanalytically oriented therapies: first findings of the Stockholm Outcome of Psychotherapy and Psychoanalysis Study. Psychother Res 2001;11:361–82. [7] Breslow NE, Day NE. Statistical methods in cancer research. The analysis of case-control studies IARC Scientific Publications no 32, Vol. 1. Lyon: International Association for Research on Cancer; 1980. [8] Busch F. Distinguishing psychoanalysis from psychotherapy. Int J Psychoanal 2010;91:23–34. [9] de Maat S, Dekker J, Schoevers R, de Jonghe F. The effectiveness of long-term psychotherapy: methodological research issues. Psychother Res 2007;17:59– 65. [10] de Maat S, de Jonghe F, Schoevers R, Dekker J. The effectiveness of long-term psychoanalytic therapy: a systematic review of empirical studies. Harv Rev Psychiatry 2009;17:1–23. [11] de Maat S, de Jonghe F, de Kraker R, Leichsenring F, Abbass A, Luyten P, et al. The current state of the empirical evidence for psychoanalysis: a metaanalytic approach. Harv Rev Psychiatry 2013;21:107–37. [12] Derogatis LR, Lipman RS, Covi L. SCL-90: an outpatient psychiatric rating scale– preliminary report. Psychopharmacol Bull 1973;9:13–28. [13] Dozier M, Cue KL, Barnett L. Clinicians as caregivers: role of attachment organization in treatment. J Consult Clin Psychol 1994;62:793–800.

9

[14] Eagle MN, Wolitzky DL. Systematic empirical research versus clinical case studies: a valid antagonism? J Am Psychoanal Assoc 2011;59: 791–818. [15] Eagle MN, Wolitzky DL. Response to commentaries. J Am Psychoanal Assoc 2012;60:153–68. [16] Gabbard GO. Long-term psychodynamic psychotherapy: a basic text. Arlington, VA: American Psychiatric Publishing, Inc; 2004. [17] Gill MM. Ego psychology and psychotherapy. Psychoanal Q 1951;20:62–71. [18] Gill MM. Psychoanalysis and exploratory psychotherapy. J Am Psychoanal Assoc 1954;2:771–97. [19] Gill MM. Psychoanalysis and psychotherapy. A revision. Int J Psychoanal 1984;11:161–79. [20] Grant J, Sandell R. Close family or mere neighbours? Some empirical data on the differences between psychoanalysis and psychotherapy. In: Richardson P, Renlund C, Ka¨chele H, editors. Research on psychoanalytic psychotherapy with adults. London: Karnac Books; 2004. p. 81–107. [21] Greenson RR. The technique and practice of psychoanalysis. London: Hogarth Press and the Institute of Psycho-analysis; 1985. [22] Hartmann A, Zeeck A, Orlinsky D.In: Accuracy of therapist’s perception of patient’s alliance: explaining the error. Paper presented at the 42nd International meeting of the Society for Psychotherapy Research; 2011 June 29–July 2. [23] Heinonen E, Lindfors O, Ha¨rka¨nen T, Virtala E, Ja¨a¨skela¨inen T, Knekt P. Therapists’ professional and personal characteristics as predictors of working alliance in short-term and long-term psychotherapies. Clin Psychol Psychother 2013. http://dx.doi.org/10.1002/cpp.1852. [24] Heinonen E, Lindfors O, Laaksonen MA, Knekt P. Therapists’ professional and personal characteristics as predictors of outcome in short- and long-term psychotherapy. J Affect Disord 2012;138:301–12. [25] Heinonen E, Orlinsky DE. Psychotherapists’ personal identities, theoretical orientations, and professional relationships: elective affinity and role adjustment as modes of congruence. Psychother Res 2013. http://dx.doi.org/ 10.1080/10503307.2013.814926. [26] Hoffman IZ. Doublethinking our way to scientific legitimacy: the desiccation of human experience. J Am Psychoanal Assoc 2009;57:1043–69. [27] Hoffman IZ. Response to Eagle and Wolitzky. J Am Psychoanal Assoc 2012;60:105–20. [28] Holi M, Sammallahti P, Aalberg V. A Finnish validation study of the SCL-90. Acta Psychiatr Scand 1998;97:42–6. [29] Ha¨rka¨nen T, Knekt P, Virtala E, Lindfors O. A case study in comparing therapies involving informative drop-out, non-ignorable non-compliance and repeated measurements. Stat Med 2005;24:3773–87. [30] Kazdin AE. Mediators and mechanisms of change in psychotherapy research. Annu Rev Clin Psychol 2007;3:1–27. [31] Kernberg OF. Psychoanalysis, psychoanalytic psychotherapy and supportive psychotherapy: contemporary controversies. Int J Psychoanal 1999;80: 1075–91. [32] Knekt P, Lindfors O. editors. A randomized trial of the effect of four forms of psychotherapy on depressive and anxiety disorders. Design, methods, and results on the effectiveness of short-term psychodynamic psychotherapy and solution-focused therapy during a one-year follow-up. Helsinki: The Social Insurance Institution; 2004 [http://www.thl.fi./hps, Accessed March 12, 2013]. [33] Knekt P, Lindfors O, Ha¨rka¨nen T, Va¨likoski M, Virtala E, Laaksonen MA, et al. Randomized trial on the effectiveness of long- and short-term psychodynamic psychotherapy and solution-focused therapy on psychiatric symptoms during a 3-year follow-up. Psychol Med 2008;38:689–703. [34] Knekt P, Lindfors O, Laaksonen MA, Renlund C, Haaramo P, Ha¨rka¨nen T, et al. Quasi-experimental study on the effectiveness of psychoanalysis, long-term and short-term psychotherapy on psychiatric symptoms, work ability and functional capacity during a 5-year follow-up. J Affect Disord 2011;132: 37–47. [35] Ka¨chele H. Distinguishing psychoanalysis from psychotherapy. Int J Psychoanal 2010;91:35–43. [36] Lee J. Covariance adjustment of rates based on the multiple logistic regression model. J Chronic Dis 1981;34:415–26. [37] Leichsenring F, Rabung S. Long-term psychodynamic psychotherapy in complex mental disorders: update of a meta-analysis. Br J Psychiatry 2011;199:15–22. [38] Lipton SD. Later developments in Freud’s technique (1920–1939). In: Wolman BB, editor. Psychoanalytic technique: a handbook for the practicing psychoanalyst. New York, NY: Basic Books; 1967. p. 51–92. [39] Lutz W, Leon SC, Martinovich Z, Lyons JS, Stiles WB. Therapist effects in outpatient psychotherapy: a three-level growth curve approach. J Couns Psychol 2007;54:32–9. [40] Migon HS, Gamerman D. Statistical inference: an integrated approach. London: Arnold; 1999. [41] Nissen-Lie H, Monsen JT, Rønnestad MH. Therapist predictors of early patientrated working alliance: a multilevel approach. Psychother Res 2010;20: 627–46. [42] Nissen-Lie H, Monsen JT, Ulleberg P, Rønnestad MH. Psychotherapists’ selfreports of their interpersonal functioning and difficulties in practice as predictors of patient outcome. Psychother Res 2013;23:86–104. [43] Øiesvold T, Bakkejord T, Sexton JA. Concurrent validity of the Health of the Nation Outcome Scales compared with a patient-derived measure, the Symptom Checklist-90-Revised in out-patient clinics. Psychiatry Res 2011;187: 297–300.

Please cite this article in press as: Heinonen E, et al. Therapists’ professional and personal characteristics as predictors of outcome in long-term psychodynamic psychotherapy and psychoanalysis. European Psychiatry (2013), http://dx.doi.org/10.1016/ j.eurpsy.2013.07.002

G Model

EURPSY-3089; No. of Pages 10 10

E. Heinonen et al. / European Psychiatry xxx (2013) xxx–xxx

[44] Orlinsky DE, Ronnestad MH. How psychotherapists develop: a study of therapeutic work and professional growth. Washington, DC: American Psychological Association; 2005. [45] Rangell L. Similarities and differences between psychoanalysis and dynamic psychotherapy. J Am Psychoanal Assoc 1954;2:734–44. [46] Rothman K, Greenland S, Lash T. Modern epidemiology, 3rd ed., Philadelphia, PA: Lippincott Williams & Wilkins; 2008. [47] Rubin DB. Multiple imputation for nonresponse in surveys. New York, NY: Wiley; 1987. [48] Rubino G, Barker C, Roth T, Fearon P. Therapist empathy and depth of interpretation in response to potential alliance ruptures: the role of therapist and patient attachment styles. Psychother Res 2000;10:408–20. [49] Sandell R, Blomberg J, Lazar A, Carlsson J, Broberg J, Schubert J. Varieties of long-term outcome among patients in psychoanalysis and long-term psychotherapy: a review of findings in the Stockholm Outcome of Psychoanalysis and Psychotherapy Project (STOPP). Int J Psychoanal 2000;81:921–42. [50] Sandell R, Lazar A, Grant J, Carlsson J, Schubert J, Broberg J. Therapist attitudes and patient outcomes. III. A latent class analysis of therapists. Psychol Psychother 2006;79:629–47. [51] Sandell R, Lazar A, Grant J, Carlsson J, Schubert J, Broberg J. Therapist attitudes and patient outcomes: II. Therapist attitudes influence change during treatment. Psychother Res 2007;17:196–204. [52] Sandell R. Research on outcomes of psychoanalysis and psychoanalysisderived psychotherapies. In: Gabbard GO, Litowitz BE, Williams P, editors. Textbook of psychoanalysis. Arlington, VA: American Psychiatric Press; 2012. p. 385–403. [53] SAS Institute Inc. SAS Online Doc. SAS/STAT User’s Guide, Version 9.2. Cary, NC: SAS Institute Inc.; 2008. [54] Schachter J, Ka¨chele H. The analyst’s role in healing: psychoanalysis-plus. Psychoanal Psychol 2007;24:429–44. [55] Schmitz N, Kruse J, Heckrath C, Alberti L, Tress W. Diagnosing mental disorders in primary care: the General Health Questionnaire (GHQ) and the Symptom Check List (SCL-90-R) as screening instruments. Soc Psychiatry Psychiatr Epidemiol 1999;34:360–6. [56] Schofield M, Orlinsky D, Wiseman H, Grant J.In: Therapists attachment styles, experience of self in close personal relationships and relational manner in therapy. Paper presented at the 42nd International meeting of the Society for Psychotherapy Research; 2011 June 29–July 2.

[57] Seybert C, Desmet M, Erhardt I, Ho¨rz S, Mertens W, Ablon JS, et al. Therapists’ techniques in psychoanalysis and short-term and long-term psychoanalytic psychotherapy: are they different? J Am Psychoanal Assoc 2011;59:592–602. [58] Stolorow RD. Converting psychotherapy to psychoanalysis: a critique of the underlying assumptions. Psychoanal Inq 1990;10:119–30. [59] Tyrrell CL, Dozier M, Teague GB, Fallot RD. Effective treatment relationships for persons with serious psychiatric disorders: the importance of attachment states of mind. J Consult Clin Psychol 1999;67:725–33. [60] Verbeke G, Molenberghs G. Linear mixed models in practice: a SAS-oriented approach. New York, NY: Springer; 1997. [61] Vivona JM. Between a rock and hard science: how should psychoanalysis respond to pressures for quantitative evidence of effectiveness? J Am Psychoanal Assoc 2012;60:121–9. [62] Wachtel PL. Beyond ESTs: problematic assumptions in the pursuit of evidencebased practice. Psychoanal Psychol 2010;27:251–72. [63] Waldron S, Scharf RD, Hurst D, Firestein SK, Burton A. What happens in a psychoanalysis? A view through the lens of the analytic process scales (APS) Int J Psychoanal 2004;85:443–66. [64] Wallerstein RS. Introduction to panel on psychoanalysis and psychotherapy: the relationship of psychoanalysis to psychotherapy: current issues. Int J Psychoanal 1969;50:117–26. [65] Wallerstein RS. 42 lives in treatment. A study of psychoanalysis and psychotherapy. New York, NY: Guilford; 1986. [66] Wallerstein RS. The psychotherapy research project of the Menninger Foundation: an overview. J Consult Clin Psychol 1989;57:195–205. [67] Wallerstein RS. The talking cures: the psychoanalyses and the psychotherapies. New Haven, CT: Yale University Press; 1995. [68] Walls G. Is systematic quantitative research scientifically rigorous? Methodological and statistical considerations. J Am Psychoanal Assoc 2012;60:145–52. [69] Wampold BE. The great psychotherapy debate: models, methods, and findings. Mahwah, NJ: Lawrence Erlbaum Associates Publishers; 2001. [70] Warren CS. Faustian science and the future of psychoanalysis. J Am Psychoanal Assoc 2012;60:131–44. [71] Widlo¨cher D. Distinguishing psychoanalysis from psychotherapy. Int J Psychoanal 2010;91:45–50. [72] Zeeck A, Orlinsky DE, Hermann S, Joos A, Wirsching M, Weidmann W, et al. Stressful involvement in psychotherapeutic work: therapist, client and process correlates. Psychother Res 2012;22:543–55.

Please cite this article in press as: Heinonen E, et al. Therapists’ professional and personal characteristics as predictors of outcome in long-term psychodynamic psychotherapy and psychoanalysis. European Psychiatry (2013), http://dx.doi.org/10.1016/ j.eurpsy.2013.07.002