Four decades of outcome research on psychotherapies for

  • Slides: 56
Download presentation
Four decades of outcome research on psychotherapies for adult depression: What's next? Norwegian Psychological

Four decades of outcome research on psychotherapies for adult depression: What's next? Norwegian Psychological Association, Oslo, September 2 nd 2016 Pim Cuijpers

OVERVIEW • What have we learned about psychotherapies for adult depression? • • The

OVERVIEW • What have we learned about psychotherapies for adult depression? • • The effects Comparisons between psychotherapies Comparisons with ADM Characteristics of patients Characteristics of therapies Other outcomes Causes of overestimation of the effects • The need for new treatments • New directions • Conclusions 2

OVERVIEW • What have we learned about psychotherapies for adult depression? • • The

OVERVIEW • What have we learned about psychotherapies for adult depression? • • The effects Comparisons with other psychotherapies Comparisons with ADM Characteristics of patients Characteristics of therapies Other outcomes Causes of overestimation of the effects • The need for new treatments • New directions • Conclusions 3

METHODS • • Database RCTs on therapies for depression >60 published meta-analyses Methods: Cuijpers

METHODS • • Database RCTs on therapies for depression >60 published meta-analyses Methods: Cuijpers et al. , BMC Psychiatry 2008; 8: 36. Data can be downloaded by other researchers: www. evidencebasedpsychotherapies. org • Overview: Cuijpers et al. , Nord J Psychiatry 2011 • Not only significance, but also size of effect: – Small: – Moderate: – Large: d=0. 20 d=0. 50 d=0. 80 NNT=8. 93 NNT=3. 62 NNT=2. 34 • Threshold for clinical relevance: d=0. 24 (Cuijpers et al. , Depr Anx, 2014) 4

Freely available at: http: //bit. do/meta-analysis

Freely available at: http: //bit. do/meta-analysis

RANDOMIZED TRIALS ON PSYCHOTHERAPIES FOR ADULT DEPRESSION (N=400) Cuijpers et al. , Curr Opin

RANDOMIZED TRIALS ON PSYCHOTHERAPIES FOR ADULT DEPRESSION (N=400) Cuijpers et al. , Curr Opin Psychiatry 2015 6

440 RANDOMIZED TRIALS • Effects of different psychotherapies versus control groups • Direct comparisons

440 RANDOMIZED TRIALS • Effects of different psychotherapies versus control groups • Direct comparisons between major types of psychotherapy with other psychotherapies • Direct comparisons of psychotherapy with pharmacotherapy • Comparisons of psychotherapy with combined treatment • Comparisons of pharmacotherapy with combined treatment • Randomized trials on psychotherapy for inpatients • Direct comparisons of individual and group therapy • Direct comparisons of face-to-face therapy with guided self-help • Randomized trials on self-guided therapy for depression 7

OVERVIEW • What have we learned about psychotherapies for adult depression? • • The

OVERVIEW • What have we learned about psychotherapies for adult depression? • • The effects Comparisons between psychotherapies Comparisons with ADM Characteristics of patients Characteristics of therapies Other outcomes Causes of overestimation of the effects • The need for new treatments • New directions • Conclusions 8

440 RANDOMIZED TRIALS • Effects of different psychotherapies versus control groups • Direct comparisons

440 RANDOMIZED TRIALS • Effects of different psychotherapies versus control groups • Direct comparisons between major types of psychotherapy with other psychotherapies • Direct comparisons of psychotherapy with pharmacotherapy • Comparisons of psychotherapy with combined treatment • Comparisons of pharmacotherapy with combined treatment • Randomized trials on psychotherapy for inpatients • Direct comparisons of individual and group therapy • Direct comparisons of face-to-face therapy with guided self-help • Randomized trials on self-guided therapy for depression 9

EFFECTS OF PSYCHOTHERAPIES COMPARED TO CONTROL GROUPS CBT N 159 d 0. 60 NNT

EFFECTS OF PSYCHOTHERAPIES COMPARED TO CONTROL GROUPS CBT N 159 d 0. 60 NNT 3 Behavioral activation Interpersonal Psychother. Problem-solving therapy Supportive therapy Psychodynamic therapy MBCT Other 11 22 21 17 8 6 51 0. 82 0. 67 0. 48 0. 52 0. 44 0. 73 0. 70 2 3 4 4 4 3 3 p for difference is 0. 71 10

DIFFERENT TYPES OF CONTROL GROUPS (ONLY CBT) Waiting list Care as usual Pill placebo

DIFFERENT TYPES OF CONTROL GROUPS (ONLY CBT) Waiting list Care as usual Pill placebo and others p for difference is 0. 003 11 N 55 26 13 d 0. 83 0. 59 0. 51 NNT 2 3 4

OVERVIEW • What have we learned about psychotherapies for adult depression? • • The

OVERVIEW • What have we learned about psychotherapies for adult depression? • • The effects Comparisons between psychotherapies Comparisons with ADM Characteristics of patients Characteristics of therapies Other outcomes Causes of overestimation of the effects • The need for new treatments • New directions • Conclusions 12

MAIN PSYCHOTHERAPIES FOR ADULT DEPRESSION • • Cognitive behavior therapy Behavioral activation therapy Interpersonal

MAIN PSYCHOTHERAPIES FOR ADULT DEPRESSION • • Cognitive behavior therapy Behavioral activation therapy Interpersonal psychotherapy Problem solving therapy Psychodynamic therapy Non-directive supportive counseling Social skills training • Families or types of therapy? • Difficult to assess from papers which type of therapy is used Cuijpers et al. , J Consult Clin Psychol 2008 13 Faculty of Psychology and Education

DIFFERENCES BETWEEN PSYCHOTHERAPIES? N d NNT CBT vs all other 56 0. 03 167

DIFFERENCES BETWEEN PSYCHOTHERAPIES? N d NNT CBT vs all other 56 0. 03 167 Supportive vs all other 30 -0. 17 * 10 Behavioral Activ. vs all other 21 0. 14 13 Psychodynamic vs all other 16 -0. 07 25 PST vs all other 7 0. 40 5 IPT vs all other 8 0. 21 * 8 Social Skills vs all other 7 0. 05 36 Cuijpers et al. , J Consult Clin Psychol 2008; Barth et al. , Plos Med 2013 14

COMPARATIVE EFFECTS OF PSYCHOTHERAPIES Barth et al. , Plos Med 2013 16

COMPARATIVE EFFECTS OF PSYCHOTHERAPIES Barth et al. , Plos Med 2013 16

EFFECTIVE THERAPIES OR EFFECTIVE MECHANISMSIN TREATMENT GUIDELINES FOR DEPRESSION? • If all therapies are

EFFECTIVE THERAPIES OR EFFECTIVE MECHANISMSIN TREATMENT GUIDELINES FOR DEPRESSION? • If all therapies are (about) equally effective, why not focus on effective mechanisms? • Universal / non-specific mechanisms may explain that • So many possible mediators/moderators that there are therapy specific mechanisms but they are not seen • If effects are comparable that is not automatically evidence for universal mechanisms • The evidence for causal mechanisms is very weak • Safe choice is evidence based therapies: CBT, IPT, but also BA Cuijpers et al. , Depress Anx 2013 17 Faculty of Psychology and Education

OVERVIEW • What have we learned about psychotherapies for adult depression? • • The

OVERVIEW • What have we learned about psychotherapies for adult depression? • • The effects Comparisons between psychotherapies Comparisons with ADM Characteristics of patients Characteristics of therapies Other outcomes Causes of overestimation of the effects • The need for new treatments • New directions • Conclusions 18

COMPARISONS WITH PHARMACOTHERAPY N d Psychotherapy vs pharmacotherapy 37 -0. 07 25 Psychotherapy vs

COMPARISONS WITH PHARMACOTHERAPY N d Psychotherapy vs pharmacotherapy 37 -0. 07 25 Psychotherapy vs combined treatment 19 0. 35 5 22 0. 30 6 16 0. 25 7 Pharmacotherapy vs combined treatment Psychotherapy + pharmacotherapy vs PSY+ placebo Cuijpers et al. , J Clin Psychiatry 2008; 2009; 2010; Depress Anx 2008; Acta Psychiatrica Scand 2009; World Psychiatry 2013; 2014 19 NNT

PSYCHOTHERAPY VS ADM IN MOOD ANXIETY • 67 trials (40 depression; 27 anxiety disorders)

PSYCHOTHERAPY VS ADM IN MOOD ANXIETY • 67 trials (40 depression; 27 anxiety disorders) • In mood anxiety disorders • Psychotherapy less effective in dysthymia (g=-0. 30) • Psychotherapy more effective in OCD (g=0. 64) • Counseling less effective than pharmacotherapy (g=-0. 33) • TCAs less effective than psychotherapies (g=0. 21) • Remains significant in multivariate metaregression analyses, except for dysthymia Cuijpers et al. , World Psychiatry, 2013 20

ARE TREATMENTS OF DEPRESSION EFFECTIVE? A) Pharmacotherapy b) Psychotherapy c) Combined therapyd) Cohen’s d

ARE TREATMENTS OF DEPRESSION EFFECTIVE? A) Pharmacotherapy b) Psychotherapy c) Combined therapyd) Cohen’s d 0. 31 0. 25 0. 52 NNT 5. 75 7. 14 3. 50 Only comparisons with pill placebo b) Based on Turner et al. , Nw Engl J Med 2008; adjusted for publication bias c) Cuijpers et al. , Psychol Med 2013 (N=12) d) Data from our database, not (yet) published; only combined versus placebo (N=6) a) 21

11 STUDIES WITH PLA - PSY – PHA COMB vs placebo PHA vs COMB

11 STUDIES WITH PLA - PSY – PHA COMB vs placebo PHA vs COMB PHA vs placebo PSY vs COMB PSY vs placebo N 11 11 11 g 0. 74 0. 37 0. 35 0. 38 0. 37 95% CI 0. 48~1. 01 0. 12~0. 63 0. 21~0. 49 0. 16~0. 59 0. 11~0. 64 Cuijpers et al. , World Psychiatry 2013 22 NNT 3 5 5

LONG TERM OUTCOMES (POSTRANDOMIZATION) N Therapy vs control (acute) • response at >6 mn

LONG TERM OUTCOMES (POSTRANDOMIZATION) N Therapy vs control (acute) • response at >6 mn • response at >12 mn Therapy vs contr. (respnd) • Sustained resp. >6 mn • Sustained resp. >2 yr • No relapse >6 mn • No relapse >12 mn 22 1. 96 1. 50~2. 55 11 1. 59 1. 14~2. 21 16 6 11 5 Karyotaki et al. , 2014 23 OR 2. 37 2. 19 3. 34 2. 46 1. 78~3. 14 1. 17~4. 09 1. 60~3. 41 1. 26~4. 82

LONG-TERM EFFECTS • Acute CBT with (almost) no maintenance vs discontinued pharmacotherapy, at 12

LONG-TERM EFFECTS • Acute CBT with (almost) no maintenance vs discontinued pharmacotherapy, at 12 months follow-up (N=8): – OR = 2. 61 (95% CI: 1. 58~4. 31), p<0. 001 • Acute CBT with (almost) no maintenance vs maintenance pharmacotherapy, at 12 months follow-up (N=5): – OR = 1. 62 (95% CI: 0. 97~2. 72), p=0. 07 – NNT = 9. 5 Cuijpers et al. , BMJ open 2013 24

LONG TERM OUTCOMES (POSTRANDOMIZATION) N OR Combined vs ADM • response at >6 months

LONG TERM OUTCOMES (POSTRANDOMIZATION) N OR Combined vs ADM • response at >6 months 12 2. 72 1. 83~4. 04 • response at >12 months 8 2. 72 1. 50~4. 96 Combined vs PSY • response at >6 months 7 Karyotaki et al. , 2014 25 1. 30 0. 76~2. 22

OVERVIEW • What have we learned about psychotherapies for adult depression? • • The

OVERVIEW • What have we learned about psychotherapies for adult depression? • • The effects Comparisons between psychotherapies Comparisons with ADM Characteristics of patients Characteristics of therapies Other outcomes Causes of overestimation of the effects • The need for new treatments • New directions • Conclusions 26

EFFECT SIZES PER TARGET GROUP (N=295) Adults in general Older adults Students N 144

EFFECT SIZES PER TARGET GROUP (N=295) Adults in general Older adults Students N 144 40 17 g 0. 76 0. 70 0. 95 95% CI 0. 67~0. 84 0. 54~0. 85 0. 67~1. 23 Women with PPD General medical Other 23 37 34 0. 64 0. 60 0. 43~0. 85 3 0. 49~0. 80 3 0. 43~0. 77 3 p for difference: 0. 26 Cuijpers et al. , 2006; 2008; 2009; 2014; Van Straten et al. , 2011 27 NNT 2 3 2

SETTINGS Primary care • referral • systematic screening Inpatients Cuijpers et al. , Br

SETTINGS Primary care • referral • systematic screening Inpatients Cuijpers et al. , Br J Gen Pract 2009; Cuijpers et al. , Clin Psychol Rev 2010 28 N g NNT 20 11 7 15 0. 31 0. 43 0. 13 0. 29 6 4 14 6

CLINICAL CHARACTERISTICS N g NNT 15 0. 27 7 15 0. 17 10 8

CLINICAL CHARACTERISTICS N g NNT 15 0. 27 7 15 0. 17 10 8 0. 26 7 9 0. 31 6 • versus control 8 0. 23 8 • versus ADM 10 -0. 31 6 • COMB vs ADM 9 0. 23 8 • COMB vs psychotherapy 4 0. 45 4 18 0. 35 5 Comorbid alcohol • depression • alcohol • 6 -12 months depression • 6 -12 months alcohol Chronic depression/dysthymia Subclinical depression Riper et al. , Addiction 2014; Cuijpers et al. , Clin Psychol Rev 2010; Br J Psychiatry 2014 29

OTHER CHARACTERISTICS • Proportion ethnic minorities: no association • No association between outcome and

OTHER CHARACTERISTICS • Proportion ethnic minorities: no association • No association between outcome and baseline severity • In low- and middle income countries (depression & anxiety) • 17 studies • Brazil, Uganda (4), Turkey, Pakistan, China (2), Cambodia, India, Iran (1) • g: 1. 02 (95% CI: 0. 76~1. 28); NNT=1. 89 • Depression (8 studies): ES=1. 07 (95% CI: 0. 71~1. 43): NNT=1. 82 Ünlü et al. , Psychiatr Serv 2014; Driessen et al. , J Consult Clin Psychol 2010; Van ‘t Hof et al. , 2011 30

INDIVIDUAL PATIENT DATA META-ANALYSES • 16 trials comparing CBT with ADM and pill placebo

INDIVIDUAL PATIENT DATA META-ANALYSES • 16 trials comparing CBT with ADM and pill placebo • N>1800 patients • Moderators: • No difference between men and women • No association between baseline severity and outcome • No association between other sociodemographics and outcome • No difference between ADM and CBT in melancholia or atypical depression • Sleep is not a predictor or moderator of outcome Cuijpers et al. , Depress Anx 2014; Weitz et al. , JAMA Psychiatry 2015 31

OVERVIEW • What have we learned about psychotherapies for adult depression? • • The

OVERVIEW • What have we learned about psychotherapies for adult depression? • • The effects Comparisons between psychotherapies Comparisons with ADM Characteristics of patients Characteristics of therapies Other outcomes Causes of overestimation of the effects • The need for new treatments • New directions • Conclusions 32

TREATMENT FORMAT (N=295) Individual Group Guided self-help Mixed/other p for difference: 0. 92 33

TREATMENT FORMAT (N=295) Individual Group Guided self-help Mixed/other p for difference: 0. 92 33 N g NNT 133 108 41 13 0. 71 0. 73 0. 71 0. 62 3 3

DIFFERENCES BETWEEN FORMATS? • Direct comparisons Individual vs group : • N= 19; d

DIFFERENCES BETWEEN FORMATS? • Direct comparisons Individual vs group : • N= 19; d = 0. 20; p<0. 01; NNT = 9 (individual is superior) • Drop-out is higher in group: OR=0. 56; p<0. 01 • Internet-based therapies is effective compared to control groups (12 studies) • With support: large effects (d=0. 61) • Without support: small effects (d=0. 25) • • Effects of self-guided therapy are smaller: 7 large trials d=0. 28, NNT=6 Direct comparisons between GSH and FTF therapies for depression/anxiety (21 studies): no difference Cuijpers et al. , Eur J Psychiatry 2008; Psychol Med, 2010; Plos One 2011; Spek et al. , Psychol Med 2007; Andersson & Cuijpers, 2009; Andrews et al. , Plos One 2010; 34

TREATMENT INTENSITY • 70 studies on individual therapies • Only small association between number

TREATMENT INTENSITY • 70 studies on individual therapies • Only small association between number of sessions and treatment effect • Negative association between length of treatment and outcome • Strong positive association between number of sessions per week and outcome • Two in stead of 1 session per week would increase effect size with 0. 45 • Sessions per week remained significant in multivariate metaregression analysis Cuijpers et al. , JAD 2013 35

NUMBER OF SESSIONS 36

NUMBER OF SESSIONS 36

NUMBER OF SESSIONS PER WEEK 37

NUMBER OF SESSIONS PER WEEK 37

OVERVIEW • What have we learned about psychotherapies for adult depression? • • The

OVERVIEW • What have we learned about psychotherapies for adult depression? • • The effects Comparisons between psychotherapies Comparisons with ADM Characteristics of patients Characteristics of therapies Other outcomes Causes of overestimation of the effects • The need for new treatments • New directions • Conclusions 38

OTHER OUTCOMES N d NNT Quality of life 31 0. 33 6 Suicidality 4

OTHER OUTCOMES N d NNT Quality of life 31 0. 33 6 Suicidality 4 0. 12 n. s. 15 Hopelessness 18 1. 10 2 Social functioning 31 0. 46 4 Social support 15 0. 38 5 Mental health children 7 0. 40 5 Mother-child interaction 8 0. 35 5 Parental functioning 5 0. 67 3 Cuijpers et al. , Ment H Phys Act 2014; EACP 2015; Renner et al. , Psychol Med 2013; Park et al. , COTR 2015; Kolovos, Br J Psychiatry, in press 39

OVERVIEW • What have we learned about psychotherapies for adult depression? • • The

OVERVIEW • What have we learned about psychotherapies for adult depression? • • The effects Comparisons with other psychotherapies Comparisons with ADM Characteristics of patients Characteristics of therapies Other outcomes Causes of overestimation of the effects • The need for new treatments • New directions • Conclusions 40

RISK OF BIAS Risk of bias N 0 (high) 1 2 3 4 (low)

RISK OF BIAS Risk of bias N 0 (high) 1 2 3 4 (low) g 15 1. 06 118 0. 88 45 0. 81 33 0. 69 84 0. 47 p for difference: <0. 001 See also: Cuijpers et al. , Psychol Med 2010 41 NNT 2 2 2 3 4

PUBLICATION BIAS? • Duvall & Tweedie’s trim & fill procedure: – Unadjusted effect size:

PUBLICATION BIAS? • Duvall & Tweedie’s trim & fill procedure: – Unadjusted effect size: d=0. 67 – Adjusted effect size: d=0. 42 (p<0. 001) – Imputed studies: 51 • Other tests: Begg & Mazumdar; Egger’s test • Very significant results (p<0. 001) • No indication for publication bias in IPT Cuijpers et al. , Br J Psychiatry 2010 42

PUBLICATION BIAS? UNADJUSTED 43

PUBLICATION BIAS? UNADJUSTED 43

PUBLICATION BIAS? ADJUSTED 44

PUBLICATION BIAS? ADJUSTED 44

A BETTER ESTIMATE OF THE EFFECTS (ALL THERAPIES) N 45 g NNT All studies

A BETTER ESTIMATE OF THE EFFECTS (ALL THERAPIES) N 45 g NNT All studies 295 0. 71 3 No waiting list 161 0. 58 3 Low risk of bias 62 0. 39 5 Adjusted for publ. bias 74 0. 31 6

A BETTER ESTIMATE OF THE EFFECTS (CBT) N 46 g NNT All studies 159

A BETTER ESTIMATE OF THE EFFECTS (CBT) N 46 g NNT All studies 159 0. 71 3 No waiting list 77 0. 54 3 Low risk of bias 34 0. 39 5 Adjusted for publ. bias 38 0. 34 5

OVERVIEW • What have we learned about psychotherapies for adult depression? • • The

OVERVIEW • What have we learned about psychotherapies for adult depression? • • The effects Comparisons between psychotherapies Comparisons with ADM Characteristics of patients Characteristics of therapies Other outcomes Causes of overestimation of the effects • The need for new treatments • New directions • Conclusions 47

EFFECTIVE TREATMENTS ARE AVAILABLE • Evidence-based treatments are available: CBT and other psychotherapies, pharmacotherapy,

EFFECTIVE TREATMENTS ARE AVAILABLE • Evidence-based treatments are available: CBT and other psychotherapies, pharmacotherapy, others • Considerable improvements in treatment • Comparable effects in mental health as in general medical field (Leucht, Br J Psychiatry 2012) 48 Faculty of Psychology and Education

BUT • Only 1/3 of disease burden reduced by treatments (Andrews et al. ,

BUT • Only 1/3 of disease burden reduced by treatments (Andrews et al. , 2004) • High relapse (~50% in 2 years, ~80% in 5 years) • Treatments not more effective then 40 years ago • Most progress: how to apply the treatments • Prevalence does not go down with more treatment • Treatment effects are overestimated • Publication bias • Low quality of studies • Researcher allegiance • Underserved populations and access 49 Faculty of Psychology and Education

OVERVIEW • What have we learned about psychotherapies for adult depression? • • The

OVERVIEW • What have we learned about psychotherapies for adult depression? • • The effects Comparisons between psychotherapies Comparisons with ADM Characteristics of patients Characteristics of therapies Other outcomes Causes of overestimation of the effects • The need for new treatments • New directions • Conclusions 50

RESEARCH PRIORITIES • Better understanding of depression (heterogeneity, comorbidity, aetiology, staging) • Better diagnostic

RESEARCH PRIORITIES • Better understanding of depression (heterogeneity, comorbidity, aetiology, staging) • Better diagnostic tools • Better treatments: but how? 51 Faculty of Psychology and Education

WE DO NOT NEED • No new therapies for acute treatment Ø To examine

WE DO NOT NEED • No new therapies for acute treatment Ø To examine superiority: trials of ~1000 patients 52 • No trials examining different treatment formats • No trials comparing different therapies • No trials examing therapies in specific target groups • Unless there is a convincing reason Faculty of Psychology and Education

WE DO NEED Prevention of depression (reduction of incidence 20 -25%) Improvement of treatments:

WE DO NEED Prevention of depression (reduction of incidence 20 -25%) Improvement of treatments: • • o o • Focus on chronic, treatment resistant depression (e. g. , CBASP) Focus on relapse Who benefits from which treatment? How do treatments work? But also how does natural recovery work and how can that be stimulated? Scaling up and simplifying treatments • • • Lay health counselors Group therapies Guided self-help/Internet-based/telephone therapies are equally effective Cuijpers et al. , Am J Psychiatry, Van Zoonen et al. , J Clin Epidem 2014; Cuijpers, Curr Opin Psychiatry 2015; World Psychiatry in press 53 Faculty of Psychology and Education

OVERVIEW • What have we learned about psychotherapies for adult depression? • • The

OVERVIEW • What have we learned about psychotherapies for adult depression? • • The effects Comparisons between psychotherapies Comparisons with ADM Characteristics of patients Characteristics of therapies Other outcomes Causes of overestimation of the effects • The need for new treatments • New directions • Conclusions 54

OVERALL CONCLUSIONS • 40 years of research on psychotherapy for depression has resulted in

OVERALL CONCLUSIONS • 40 years of research on psychotherapy for depression has resulted in a huge body of knowledge • Psychotherapies are effective, but less effective than we thought for a long time • Better at the longer term • It is effective in many target groups and settings • But we need to improve treatments, not by developing new ones, but by • focusing on prevention, chronic depression and relapse • scaling up and simplifying treatments 55

Ø Thank you for your attention! Contact: p. cuijpers@vu. nl 56

Ø Thank you for your attention! Contact: p. cuijpers@vu. nl 56