Schema Therapy for Personality Disorders

Transcription

1 Schema Therapy for Personality Disorders Arnoud Arntz University of Amsterdam Maastricht University

2 Supported by grants Disclosure ZonMW, the Netherlands Organization for Health Research and Development Netherlands Institute for Advanced Studies in the Humanities and Social Sciences (NIAS) NWO/DFG (ORA international; with Gitta Jacob & Rainer Göbel) Australian Rotary Health (PI Chris Lee) Else Kröner-Fresenius-Stiftung (PI Gitta Jacob) Netherlands Foundation for Mental Health Netherlands Health Care Insurance Board. (Developmental Medicine) Author of books & chapters on ST Wiley, Guilford, Nieuwezijds, Beltz Occasional trainer/workshop leader in ST

3 Overview Theory and therapy model Effectiveness research Techniques or Therapeutic Alliance? Latest developments

4

5

6 1. Theory (Personality) psychopathology develops when basic needs are not met Emotional needs (children, adults) Safety (secure attachment, stability, care, acceptance) Expression of needs, emotions, opinions Social Inclusion Autonomy, competence, identity Spontaneity, play Realistic limits, self-control

7 Schema Mode Concept Schema = trait ; mode = state Mode = emotional-cognitive-behavioral state Combination of activated schema & coping Modes vary from functional (flexible, adaptive) to dysfunctional (inflexible, nonadaptive)

8 Schema Modes 4 clusters Child Modes Vulnerable; Impulsive; Undisciplined; Angry Child Internalized parent modes Punitive; Demanding Parent Coping modes Surrender, Avoidance, Overcompensation Healthy modes Healthy adult, happy child

9 Bamelis et al., Journal of Personality Disorders, 25(1), 41 58, 2010

10 Borderline Personality Disorder: 5 modes (HEALTHY ADULT MODE) Reassure; Empathic Make Educate, it model, reinforce confrontation; less functional behaviors & Develop necessary attitudes functional ways to assert & get needs met DETACHED PROTECTOR MODE Give safety; process traumas PUNITIVE PARENT MODE ANGRY/IMPULSIVE CHILD MODE ABANDONED/ABUSED CHILD MODE Combate; Get it out patient s system

11 Integrative Ingredients of ST cognitive therapy, behavior therapy, experiential therapies (Gestalt), psychodynamic therapy, attachment theory & other developmental insights not eclectic! 3 foci: Therapeutic relationship Childhood (traumas; early relationships) Present

12 Therapeutic Relationship Limited reparenting Therapist meets needs within professional limits Includes limit setting and frustration! Therapist is more caring and more personal (e.g., self-disclosure) than usual Mimics development: Initially directive, offering safety and care Lots of psycho-education Gradual less directive & stimulates autonomy

13 Therapeutic alliance over time by condition Significant effects: ST > TAU COP linear trend Baseline severity

14 Prediction of Recovery by WAI Early working alliance predicts recovery at 3 yrs, but doesn t (fully) explain ST effects Early WAI-P OR=1.02 p =.019 Severity OR=0.32 p <.001 ST vs TAU OR=3.02 p =.017 COP vs TAU OR=0.70 p =.42

15 Childhood memory processing Mainly with experiential technics Imagery Rescripting Multiple chair techniques Psychodrama Memory activated Correction is imagined / played Though fantasy, effects are powerful Full exposure to all trauma details not necessary Working mechanism Change in meaning Interpersonal context!

16 Improvement in Composite Score exploration past present Weertman & Arntz; Behav. Res. Ther. 2007

17 3. Effectiveness of ST for PDs 6 outpatient studies (4 RCTs, 1 case series, 1 pilot) 4 individual ST 3 BPD 1 six other PDs 2 group-st (BPD) 3 inpatient studies (pilots)

18 BPD trial: ST vs. TFP Giesen-Bloo et al (2006). Archives of General Psychiatry, 63: Compared Schema Therapy to Transference-Focused Psychotherapy N = 86 BPD patients 4 sites 3 yrs. treatment (2 sessions a week) Assessment every 3 months Main outcome instrument: BPDSI

19 number of patients Patients in treatment or succesfully completed days of therapy Schema TFP Survival analyses: Cox regression RR 2.39 df (1), p = 0.016

20 BPD severity assessed with BPDSI per year incl. 1 year follow-up TFP ST (p =.004)

21 BPDSI: the 9 DSM-IV criteria emptiness 1. abandonment 2. unstable relations 3. identity disturbance anger impulsivity 5. self injury & suicidal behavior 6. affective instability dissociation & paranoid ideation

22 Secondary Measures Incl. Follow-up (ITT) (medians composite scores) * BL yr 1 yr 2 yr 3 FU * ST TFP * P <.05

23 Outpatient ST: Change in BPD severity: effect sizes Within condition change pooled Cohen s d = 2.38**

24

25 Meta-analyses of effectiveness of various treatments: BPD specific outcomes Kliem et al. (JCCP, 2010) DBT Cohen s d =.56 (0.50 for general effects) Cochrane review (2012) reports controlled effect sizes for BPD-specific outcomes (vs. TAU) DBT: (pooled mean.42) MBT: (pooled mean 1.52) ST: (pooled mean 1.66)

26 Pooled dropout % in 1st year compared (BPD trials) Arntz & Jacob, submitted

27 ST for 6 PDs (Cl-C; Nar, Par, Histr) Multicenter RCT: ST vs TAU vs COP N = session ST protocol 40 sessions in yr 1; 10 booster sessions in yr 2 TAU = psychological treatment (most psychodynamic) Bamelis et al. (2014) Am J Psychiatry

28

29 Primary Diagnosis 5% 4% 1% 11% 28% 51% Avoidant Obs-Comp Dependent Narcissistic Paranoid Histrionic

30 ST for 6 PDs Drop out % from indicated treatment (mixed log regression estimates) ** ST TAU COP * Bamelis et al. (2014) Am J Psychiatry

31 % still in treatment at 3 yr ST TAU COP *p <.05

32 % recovered from PD at 3 yr primary outcome analysis ST TAU COP ST vs TAU: OR=4.1, p <.001 ST vs COP: OR=2.9, p <.05

33 % having depression at 3-yr follow-up controlled for baseline depression 50 Baseline depression 44.7% ST TAU COP ST vs TAU: OR=0.23, p =.033

34 GAF & SOFAS scores over 3 yrs GAF: ST > TAU SOFAS: ST > TAU=CCT Cohen s d ST 1.76 COP 1.11 TAU 1.27 Cohen s d ST 1.65 COP 0.93 TAU 1.05

35 But No differences on self-report questionnaires Extra effects of ST apparent on objectifiable indices (interviews), not on subjective reports Change in self-view might lag behind objective changes (e.g., behavior) Might also point to the important behavioral change part of ST

36 Cost-effectiveness Helps decision making: which health care should be implemented, which not? Two aspects: Effects (e.g., new vs. usual treatment) Societal costs (of new vs. usual treatment) Societal to prevent that other sectors than health care get extra costs or are ignored (e.g., work; disability compensation, etc)

37 Therapist Cohorts Cohort 1 trained mainly by lectures and viewing video examples (4 days) Two centers withdrew before starting; extra therapists had to be trained Cohort 2 trained mainly by practicing (role plays) techniques after short explanation & life demonstration (4 days)

38 Cohort 1 25% Theory Practice

39 Cohort 2 75% Theory Practice

40 Drop out % ** 5 0 ST cohort 1 ST cohort 2 TAU

41 % Recovered by cohort Cohort 1 Cohort TAU ST TAU ST % recovered 0 % recovered

42 Therapeutic Alliance Assessed with the WAI 3 subscales Task, Bond & Goal Total score presented Here only patient report (most important) After 3 sessions 6 months 12 months

43 Therapeutic alliance over time by condition and cohort No initial difference between ST in cohort 1 and 2

44 ST-technique use over time by condition and cohort ST > CCT, TAU & linear + quadratic time effects ST cohort 2 > ST cohort 1

45 Use of ST techniques (independent raters blind for condition) 1.8 P < ST cohort 1 ST cohort 2

46 Therapeutic Relation or ST Techniques? At start Cohort 2 already used more ST techniques But therapeutic alliance was similar at start in both cohorts Later, alliance further improved in Cohort 2 Suggesting as result of ST techniques Conclusion: the better training in ST techniques caused the better effects in cohort 2

47 Prediction of Recovery by WAI Early working alliance predicts recovery at 3 yrs, but doesn t explain ST and ST-cohort effects Severity OR=0.32 p <.001 ST OR=3.02 p =.017 CCT OR=0.70 p =.42 ST x cohort OR=10.28 p =.011 Early WAI-P OR=1.02 p =.019

48 Training Matters!

49 Forensic ST trial: Intermediate findings with 50% of final N (Bernstein et al, in progress) RCT in high security forensic hospitals Compares ST to Treatment as Usual (TAU) All PD patients (highest recidivism) Antisocial; Borderline, Narcissistic, Paranoid PD Approx 50% psychopaths 3 year treatment (study period): ST vs TAU

50 Forensic PD treatment study: Patients Retained in Therapy for the Duration of the 3-Year Study ST > TAU, p =.06, (Chi-Square Test) N = 53

51 Intermediate independent outcome: granted leave requests Patients submit requests for short leaves from the high security hospital, e.g. to visit family Independent committee decides using risk assessment procedures

52 Forensic PD RCT Cumulative % Receiving Supervised Leave months 18 months 24 months 30 months 36 months ST TAU ST > TAU, p<.05, Kaplan-Meier Survival Analysis, N=53

53 Forensic PD RCT Cumulative % Receiving Unsupervised Leave ST TAU months 24 months 30 months 36 months ST > TAU, p<.05, Kaplan-Meier Survival Analysis, N=53

54 New Developments Group Schema Therapy Use of group cohesion, and recognition, validation and confrontation by group members Catalyzes ST techniques (Farrell et al. 2009) First studies show large effects Discussion: only group or with individual ST?

55 BPD severity assessed with 40 BPDSI (per year) RCT vs. group-st studies TFP Indiv. ST Giesen-Bloo et al (2006) Archives of General Psychiatry Dickhaut & Arntz (2014) J Behav Ther & Exp Psychiatry Group + Indiv. ST

56 Group-ST pilot: effects are more than symptom reduction BPDSI Happiness general population BPD 15 cutoff score = Dickhaut & Arntz (2014) J Behav Ther & Exp Psychiatry

57 Current RCT on group-st for International trial BPD 14 sites in 5 countries, N = 448 Group-ST vs TAU (specialist CTBE) Two subforms of group-st also compared: Only group vs. combined group-individual ST Effectiveness, cost-effectiveness, perspectives of patients & therapists

58 Further new developments ST for Dissociative Identity Disorder Central = help patients to view identities as schema modes Next, treat as personality disorder Case series (pilot) study starts this year ST for traumatized psychotic patients Case series study in preparation (Chris Taylor, UK) Computerized ambulatory ST self-help support (GAIA, Gitta Jacob et al.)

59 Conclusions Schema therapy is effective and highly acceptable treatment for personality problems Also cost-effective (cheaper and more effective) Group-ST increasingly popular Catalyzes change processes? Clinical impression: optimal in combination with individual ST More research needed into intimate relations, social, and societal functioning Recovery from PD symptoms does not always imply a full life

60 Conclusions - 2 What makes ST specifically effective? Mode model offers meta-cognitive understanding processing (traumatic) childhood experiences (~ PTSD: trauma-focused CBT > other CBT) Specific therapeutic relationship (limited reparenting = corrective experience) Focus on behavioral change later in treatment Push for societal participation and healthy relationships

61 Conclusions - 3 Specific techniques are essential Training matters influences effectiveness on patient outcome level

62 Thank you!

63 Design Screening (check on in-exclusion criteria) SCID-1 SCID-2 SFT m1 S1-40 m2 m3 S41-50 m4 FU TAU CCT Informed consent Baseline m0 Central independent randomization WAI at session 3, 6 mnths, 12 mnths (therapeutic alliance) Year 1 Year 2 Year 3 Primary outcome (loss of PD diagnosis on SCID- 2) SCID-1 also taken

64 2 1 Treatment Integrity Check (audiotapes raters blind for condition) ST-techniques COP-techniques ** 2 ** 1 0 ST TAU COP 0 ST TAU COP 4 Facilitative conditions ** 4 Explicit directiveness ** ST TAU COP 2 ST TAU COP ** p <.01

65 Proportion in indicated treatment or succesfully completed Wald (2) = 10.48, p =.005 ST vs TAU p =.004 ST vs COP n.s. 2nd cohort ST-therapists had less drop-out, p <.02

66 Drop out % from indicated treatment (mixed log regression estimates) ST cohort 1 ST cohort 2 TAU COP

67 Therapeutic alliance over time by condition Significant effects: ST > TAU CCT linear trend Baseline severity

68 Qualitative Study Ten Napel-Schutz MC, Abma TA, Bamelis L, Arntz A., in prep. Two focus groups of 8 therapists each, who had just begun with applying ST context of the Bamelis et al RCT all trained in first cohort Thematic analysis Main results Large group was unsafe Not enough room to address their resistance Missed training in specific techniques Missed regular supervision / access to expert

69 Group ST vs TAU for BPD Main Outcome Measures (Farrell et al.) Mean ES Cohen s d (BPD measures) ST = 2.62 TAU = 0.04 Recovery ST 94% TAU 25% Drop-Out ST 0 % TAU 25% Farrell et al. (2009), J. Beh. Ther. & Exp. Psychiatry.

70 Group-ST pilots: effects of training in Farrell & Shaw s approach BPDSI SCL90 pre-post d=2.72 pre-post d=1.49 SMI-dysf WhoQol pre-post d=1.33 pre-post d=1.16 Dickhaut & Arntz, in prep.

71 Conclusions ST has higher nonspecific therapeutic relationship ratings Training matters Small groups Address questions and resistance Clear instructions & models Practice in roleplays trainer helps What explains superior effects, nonspecific relationship or specific factors?

72 Conclusions Cohort 2 had higher use of specific ST techniques, but not better nonspecific working alliance Cohort 2 had less drop-out and superior effectiveness Thus, superior effects seem related to better training and thus better application of specific techniques Relational and nonrelational ST techniques This does not imply that the nonspecific working alliance is not important Probably: Necessary but not sufficient

73 Sample 56.6% female Age = 39 (sd 9.5) # previous treatments = 2.4 (sd 2.3) 34.4% disability compensation 10.9% welfare 90 % Cluster-C as primary diagnosis ~ 70% life time depression on axis % current depression/dysthymia 58.4% anxiety disorder(s) 50.9% medication (at baseline)

74 Treatments ST = 50-session protocol 40 weekly sessions in year 1 10 booster sessions in year 2 (to be) published in English & German COP = clarification-oriented psychotherapy (CCT model by Rainer Sachse) weekly sessions, open ended TAU = optimal treatment at local site according to indication staff indicated: 41.8% insight-oriented psychotherapy 32.1% supportive psychotherapy 20.9% CBT/EMDR

75 Early WAI does not predict recovery in ST cohort 2