Attachment_4_PTSD_Practice_Guidelines_DJF-18-4210-PR-0008262.pdf

Appendices

1 CLINICAL PRACTICE

GUIDELINEPTSD

CLINICAL PRACTICE

GUIDELINE

PTSD

Table of Contents: Appendices

Appendix A: Descriptions of Treatments and Strength of Recommendations……………125

Appendix B: Treatment Dose, Timing, and Duration as Described in

Reviewed Studies………………………………………………………………………….128

Appendix C: Evidence Profiles Prepared for APA by RTI-UNC Scientists…………………129

Appendix D: Decision Tables……………………………………………………………………….224

Appendix E: Definition of Key Terms……………………………………………………………...608

Appendix F: Additional Tables and Figure for Updated Evidence (2012-2016)…………….612

Appendix G: Trials identified through search update (published 2012 – 2016) matching one or more recommendations

Appendix H: Trials identified through search update (published 2012 – 2016) matching one or more Evidence Profiles but not matching any recommendations………..633

Appendix I: Trials identified through search update (published 2012 – 2016) that would not have met criteria for inclusion in systematic review………………635

Appendix J: APA Declarations/COI form…………………………………………………………..655

Appendix K: Voting Procedures Established by Advisory Steering Committee (ASC)…...672

Appendix L: Strength of evidence ratings for psychotherapies and medications ..………674

Appendix A

Descriptions of Treatments and Strength of Recommendations

Intervention Description of Treatment

Psychological

(Recommendation:

Strong for)

Cognitive behavioral therapy Utilizes behavioral and cognitive strategies, particularly exposure, cognitive restructuring, changes in behavior, and development of coping skills, to address learned and conditioned behaviors, thoughts and emotional and psychophysiological reactions.

Cognitive processing therapy Cognitive Processing Therapy is a specific type of cognitive behavioral therapy that focuses on the cognitions developed as a result of the trauma and the role that inaccurate or distorted cognitions have on emotional responses and on behavior. The primary goals of CPT are to encourage the expression of natural emotions and reduce manufactured emotions related to the trauma; to identify and challenge dysfunctional cognitions (“stuck points”) about the traumatic event(s) as well as current thoughts about self, others, and the world; and to promote a more balanced set of beliefs about oneself, others, and the world. It has four main parts: education about PTSD and CPT, processing the trauma, learning to challenge thoughts about the trauma, and trauma themes.

Cognitive therapy Brief, goal-oriented therapy with the aim of helping a patient analyze distorted thinking in order to reshape existing beliefs and cognitions and to modify associated problematic behaviors.

Prolonged exposure Prolonged exposure therapy (PE) is a specific cognitive behavioral therapy designed to help PTSD sufferers emotionally process their traumatic experiences through repeated revisiting and recounting of their trauma memories (imaginal exposure) and repeated, gradual confrontation of feared situations, places, and things that are objectively safe but feel more dangerous following the traumatic event (in vivo exposure). Patients are asked to expose themselves to trauma cues both during the therapy sessions and as homework between sessions in order to consolidate learning that trauma reminders are not harmful and that they are capable of tolerating the temporary distress associated with confronting these experiences. Education regarding trauma and PTSD as well as breathing retraining are also a part of the therapy.

Conditional for)

Brief eclectic therapy Manualized treatment that combines cognitive behavior therapy techniques and psychodynamic strategies including psychoeducation, relaxation, imaginal exposure, writing, learning from the trauma, meaning and integration and an ending ritual.

Conditional for)

Eye Movement Desensitization and Reprocessing Therapy

EMDR is an eight-phase psychotherapy that facilitates the accessing and processing of traumatic memories to bring these to an adaptive resolution. During EMDR therapy, the patient imaginally attends to emotionally disturbing material in brief sequential doses (sets) while simultaneously focusing (dual attentional focus) on an external bilateral stimulus such as eye movements, sounds, or tapping. Between sets, patients are asked to report what comes to mind and often report associations that they had not previously contemplated. They are also invited to verbally discuss their traumatic experiences as well as their shifting perspectives. Body scans are also utilized to determine the patient’s degree of resolution.

None)

Imagery rehearsal therapy A manualized therapy that aims to reduce intensity and frequency of nightmares by increasing knowledge about sleep and selecting a repetitive target nightmare, scripting a change to their nightmare that increases their sense of mastery or control, and then rehearsing the change in the mind’s eye, each night before sleep.

Relaxation Meditation, muscle relaxation, and deep breathing strategies to lower anxiety and stress levels.

Seeking Safety Brief manualized cognitive behavior therapy for patients with co-morbid PTSD and substance abuse disorder focused on creating safety in personal life and addressing triggers while focusing on personal ideals.

Stress inoculation training SIT exposes patient to stress in a controlled fashion to help with consequences of previous stress and learn adaptive strategies for dealing with the situation in the future.

Trauma affect regulation A manualized therapy that teaches skills for processing and managing trauma-related reactions to stressful situations. The goal is to help patients regulate intense emotions.

Intervention Description of Treatment

Pharmacological

Weak for fluoxetine, paroxetine, sertraline)

Selective serotonin reuptake inhibitors (SSRIs: citalopram, escitalopram, fluoxetine, fluvoxamine, paroxetine, sertraline)

Weak for venlafaxine)

Serotonin and norepinephrine reuptake inhibitors (SNRIs:

desvenlafaxine, venlafaxine, duloxetine)

Anticonvulsants/mood stabilizers (topiramate, tiagabine, lamotrigine, carbamazepine, divalproex)

Atypical antipsychotics (olanzapine, risperidone)

N/A)

Alpha blockers (prazosin)

Benzodiazepines (alprazolam, diazepam, lorazepam, clonazepam)

Other second-generation antidepressants (bupropion, mirtazapine, nefazodone, trazodone)

Tricyclic antidepressants (imipramine, amitriptyline, desipramine)

Note: N/A = Not applicable. The panel did not make recommendations for any interventions for which the strength of evidence from the RTI-UNC systematic review was rated as insufficient or very low.

Appendix B

Treatment Dose, Timing, and Duration as Described in Reviewed Studies

Treatment Dose Range (No. of sessions or mg)

Timing Range Duration Range Session Duration Range, minutes

Exposure 1-30 Weekly 1-30 weeks 60-120 minutes CBT 4-16 Weekly to Bi-weekly 4-16 weeks 60-120 minutes

Cognitive processing 12-17 Weekly to Bi-weekly 6-17 weeks 60-90 minutes

Cognitive 9-16 Weekly 9-16 weeks 60-90 minutes Brief eclectic 16 Weekly 16 weeks 45-60 minutes

EMDR 3-18 Weekly to Bi-weekly 3-8 weeks 60-90 minutes Narrative exposure 4-17 Weekly to Bi-weekly 3-17 weeks 60-120 minutes

Relaxation 8-14 Weekly to Bi-weekly 6-14 weeks 40-90 minutes

Seeking Safety 12-24 Weekly, Bi-weekly, or Tri-weekly

6-12 weeks 75-90 minutes

Fluoxetine 10-80mg Daily 5-12 weeks N/a

Paroxetine 12.5-62.5mg Daily 10-12 weeks N/a

Sertraline 20-200mg Daily 10-12 weeks N/a

Topiramate 12.5-500mg Daily or Twice Daily 12 weeks N/a

Venlafaxine 37.5-300mg Daily 12-24 weeks N/a

Appendix C

Evidence Profiles Prepared for APA by RTI-UNC Scientists

Psychological and Pharmacological Treatments for Adults With Posttraumatic Stress Disorder (PTSD) Prepared for:

American Psychological Association

Prepared by:

RTI International–University of North Carolina Evidence-based Practice Center Investigators:

Daniel E. Jonas, M.D., M.P.H.

Jennifer Cook Middleton, Ph.D.

Roberta Wines, M.P.H.

Cynthia Feltner, M.D., M.P.H.

Gerald Gartlehner, M.D., M.P.H.

Submitted: October 14, 2014

The Evidence Profiles included in this document were derived from the following report: Jonas DE, Cusack K, Forneris CA, Wilkins TM, Sonis J, Middleton JC, Feltner C, Meredith D, Cavanaugh J, Brownley KA, Olmsted KR, Greenblatt A, Weil A, Gaynes BN. Psychological and Pharmacological Treatments for Adults With Posttraumatic Stress Disorder (PTSD). Comparative Effectiveness Review No. 92. (Prepared by the RTI International–University of North Carolina Evidence-based Practice Center under Contract No.

290-2007-10056-I.) AHRQ Publication No. 13-EHC011-EF. Rockville, MD: Agency for Healthcare Research and Quality; April 2013.

www.effectivehealthcare.ahrq.gov/reports/final.cfm.

http://www.effectivehealthcare.ahrq.gov/reports/final.cfm

Table of Contents

Critical and Important Outcomes

Explanation of Effect Sizes Presented in the Evidence Profiles

Psychological Treatments for Adults with PTSD

Question: Is cognitive processing therapy efficacious for adults with PTSD? [Evidence from studies comparing cognitive processing therapy with inactive controls]

Question: Is cognitive therapy (not including cognitive processing therapy) efficacious for adults with PTSD? [Evidence from studies comparing cognitive therapy with inactive controls] Question: Is stress inoculation training (SIT) efficacious for adults with PTSD? [Evidence from studies comparing SIT with waitlist]

Question: Is relaxation efficacious for adults with PTSD? [Evidence from studies comparing relaxation with treatment as usual]

Question: Are exposure-based therapies efficacious for adults with PTSD? [Evidence from studies comparing exposure-based therapies with inactive controls (waitlist or usual care)]

Question: Are CBT interventions efficacious for adults with PTSD? [Evidence from studies comparing CBT therapies with inactive controls (waitlist, usual care)]

Question: Is EMDR efficacious for adults with PTSD? [Evidence from studies comparing EMDR with inactive controls (waitlist, usual care)]

Question: Is imagery rehearsal therapy (IRT) efficacious for adults with PTSD? [Evidence from studies comparing IRT with waitlist]

Question: Is narrative exposure therapy (NET) efficacious for adults with PTSD? [Evidence from studies comparing NET with waitlist or minimal attention]

Question: Is brief eclectic psychotherapy (BEP) efficacious for adults with PTSD? [Evidence from studies comparing BEP with waitlist]

Question: Is trauma affect regulation efficacious for adults with PTSD? [Evidence from studies comparing trauma affect regulation with waitlist] Question: What is the comparative effectiveness of relaxation and cognitive restructuring for adults with PTSD? [Evidence from studies comparing relaxation with cognitive restructuring]

Question: What is the comparative effectiveness of exposure-based therapies and cognitive restructuring for adults with PTSD? [Evidence from studies comparing exposure-based therapies with cognitive restructuring]

Question: What is the comparative effectiveness of exposure-based therapies and cognitive processing therapy for adults with PTSD? [Evidence from studies comparing exposure-based therapies with cognitive processing therapy]

Question: What is the comparative effectiveness of exposure-based therapies and stress inoculation training (SIT) for adults with PTSD? [Evidence from studies comparing exposure-based therapy with SIT] Question: What is the comparative effectiveness of exposure-based therapies and EMDR for adults with PTSD? [Evidence from studies comparing exposure-based therapy with EMDR]

Question: What is the comparative effectiveness of exposure-based therapies and relaxation for adults with PTSD? [Evidence from studies comparing exposure-based therapy with relaxation]

Question: What is the comparative effectiveness of exposure-based therapies and exposure plus cognitive restructuring for adults with PTSD? [Evidence from studies comparing exposure therapies with exposure plus cognitive restructuring] Question: Are CBT interventions better than relaxation for adults with PTSD? [Evidence from studies comparing CBT therapies with relaxation]

Question: Is EMDR better than relaxation for adults with PTSD? [Evidence from studies comparing EMDR with relaxation]

Question: Is seeking safety better than standard community treatment for adults with PTSD? [Evidence from studies comparing seeking safety with standard community treatment]

Question: Is seeking safety better than active controls for adults with PTSD? [Evidence from studies comparing seeking safety with active controls: relapse prevention, psychoeducation, or treatment as usual in a Veteran Affairs substance use disorders clinic] Question: How does brief eclectic psychotherapy (BEP) compare with EMDR for adults with PTSD? [Evidence from studies comparing BEP with EMDR]

Pharmacological Treatments for Adults With PTSD

Question: Are alpha blockers efficacious for adults with PTSD? [Evidence from studies comparing alpha blockers with placebo]

Question: Is divalproex efficacious for adults with PTSD? [Evidence from studies comparing divalproex with placebo]

Question: Is lamotrigine efficacious for adults with PTSD? [Evidence from studies comparing lamotrigine with placebo]

Question: Is tiagabine efficacious for adults with PTSD? [Evidence from studies comparing tiagabine with placebo] Question: Is topiramate efficacious for adults with PTSD? [Evidence from studies comparing topiramate with placebo]

Question: Is olanzapine efficacious for adults with PTSD? [Evidence from studies comparing olanzapine with placebo]

Question: Is risperidone efficacious for adults with PTSD? [Evidence from studies comparing risperidone with placebo]

Question: Are benzodiazepines efficacious for adults with PTSD? [Evidence from studies comparing any benzodiazepine with placebo]

Question: Is citalopram efficacious for adults with PTSD? [Evidence from studies comparing citalopram with placebo] Question: Is fluoxetine efficacious for adults with PTSD? [Evidence from studies comparing fluoxetine with placebo]

Question: Is paroxetine efficacious for adults with PTSD? [Evidence from studies comparing paroxetine with placebo]

Question: Is sertraline efficacious for adults with PTSD? [Evidence from studies comparing sertraline with placebo]

Question: Is venlafaxine efficacious for adults with PTSD? [Evidence from studies comparing venlafaxine with placebo]

Question: Are tricyclic antidepressants efficacious for adults with PTSD? [Evidence from studies comparing tricyclic antidepressants with placebo] Question: Is bupropion efficacious for adults with PTSD? [Evidence from studies comparing bupropion with placebo]

Question: Is mirtazapine efficacious for adults with PTSD? [Evidence from studies comparing mirtazapine with placebo]

Question: Is paroxetine more efficacious than desipramine for adults with PTSD? [Evidence from studies comparing paroxetine with desipramine]

Question: Is venlafaxine ER more efficacious than sertraline for adults with PTSD? [Evidence from studies comparing venlafaxine ER with sertraline]

Question: Is sertraline more efficacious than citalopram for adults with PTSD? [Evidence from studies comparing sertraline with citalopram]

Head-to-Head Trials of Psychological and Pharmacological Treatments Question: What is the comparative effectiveness of fluoxetine and EMDR for adults with PTSD? [Evidence from studies comparing fluoxetine with EMDR]

Question: What is the comparative effectiveness of prolonged exposure plus paroxetine vs. prolonged exposure plus placebo for adults with PTSD? [Evidence from studies comparing prolonged exposure plus paroxetine with prolonged exposure plus placebo]

Question: What is the comparative effectiveness of sertraline plus prolonged exposure vs. sertraline for adults with PTSD? [Evidence from studies comparing sertraline plus prolonged exposure with sertraline]

References

APA GDP Critical and Important Outcomes of Treatment

1Outcome ratings were determined by the APA PTSD guideline development panel in May 2014.

2Harms are defined as adverse events leading to withdrawals and/or other serious adverse events.

Note. The critical and important outcomes were selected by the APA PTSD GDP. Staff members from the RTI-UNC Evidence-based Practice Center prepared the Evidence Profiles for the outcomes selected by the GDP based on the systematic review used as the evidence base for this guideline.

Outcome Critical or Important1 PTSD symptom reduction Critical Harms2 Critical Remission (no longer having symptoms) Important Loss of diagnosis Important Prevention or reduction of comorbid depression Important Prevention or reduction of comorbid anxiety Important

Prevention or reduction of comorbid substance use Important

Prevention or reduction of comorbid pain Important

Quality of life Important

Disability or functional impairment Important

Explanation of Effect Sizes Presented in the Evidence Profiles

When we pooled data for our RTI-UNC Systematic Review (Jonas et al., 2013), we presented data as standardized mean differences (SMDs), mean differences (MDs, also known as WMDs), or risk differences (RDs).

Any data in this document that is presented as odds ratios (ORs) or as other measures were the data presented by individual studies, rather than from results of our meta-analyses.

Explanation of effect sizes reported in this document Effect Size Description/Definition Absolute Effects Describe absolute (rather than relative) differences between groups, e.g., how many more people achieved remission when treated with a drug than when receiving placebo.

Mean Difference (MD) i.e., Weighted mean difference (WMD)

Measures the absolute difference between the mean values of two groups. Used for continuous outcome measures. For example, how much greater reduction in PTSD symptoms were found in one group than in the comparison group.

WMD is used to combine measures on continuous scales (such as PTSD symptoms measured using the CAPS), where the mean, standard deviation and sample size in each group are known. The weight given to the difference in means from each study (e.g. how much influence each study has on the overall results of the meta-analysis) is determined by the precision of its estimate of effect and is equal to the inverse of the variance. This method is used when all trials have measured the outcome on the same scale.

Mean differences reported in the Evidence Profiles are calculated as the active intervention minus the control intervention, for efficacy comparisons involving an active intervention and a control intervention. When lower scores on the outcome measure at the post-treatment assessment are indicative of clinical improvement, a negative mean difference indicates that there was greater clinical improvement in the active than the control group and a positive mean difference indicates that there was greater clinical improvement in the control group than the active intervention group.

Conversely, when higher scores on an outcome measure (e.g., Global Assessment of Function) are indicative of clinical improvement, a negative mean difference indicates that there was greater improvement in the control group than the active treatment group and a positive mean difference means that there was greater improvement in the active treatment group than in the control group. When two active treatments are compared (e.g., CBT vs relaxation), the mean difference is calculated as the difference between treatment 1 and treatment 2, where treatment 1 is the treatment shown farthest to the left in the column headed “No. of patients” and treatment 2 is the treatment shown to the right, in the column headed “No. of patients”.

Relative Effect Sizes Relative effect sizes were not calculated for our RTI-UNC Systematic Review.

Therefore all relative effect size cells within the evidence profiles were left blank

Effect Size Description/Definition unless an individual study reported a relative measure, and just that 1 study was included in that row of the evidence profile.

Risk Difference (RD) Measures the absolute (not relative) difference in risk between two groups for a dichotomous outcome (e.g., achieving remission or loss of diagnosis).

The difference in risk between two groups. For example, if one group has a 25% risk of an outcome, and the other has a 10% risk of the outcome, the risk difference is 15 percentage points.

Standardized Mean Difference (SMD) Measures the difference between two estimated means divided by an estimated standard deviation. Used for continuous outcome measures. SMDs were used when effect estimates from different trials that were combined in a meta-analysis used different instruments to measure the same outcome (e.g., one study used the CAPS and another used the PCL). They were also used when studies used the same instruments, to allow the panel to be able to compare the magnitude of effect across different interventions using a common standard. It is used to combine results from studies using different ways of measuring the same concept. By expressing the effects as a standardised value, the results can be co mbined since they have no units.

SMD data presented in this document are Cohen’s d. For Cohen’s d, a small effect size is 0.20, medium is 0.50, and large is 0.90.

Standardized mean differences (SMD) reported in the Evidence Profiles are calculated as the active intervention minus the control intervention, for efficacy comparisons involving an active intervention and a control intervention. When lower scores on the outcome measure at the post-treatment assessment are indicative of clinical improvement, a negative SMD indicates that there was greater clinical improvement in the active than the control group and a positive SMD indicates that there was greater clinical improvement in the control group than the active intervention group. Conversely, when higher scores on an outcome measure (e.g., Global Assessment of Function) are indicative of clinical improvement, a negative SMD indicates that there was greater improvement in the control group than the active treatment group and a positive SMD means that there was greater improvement in the active treatment group than in the control group. When two active treatments are compared (e.g., CBT vs relaxation), the mean difference is calculated as the difference between treatment 1 and treatment 2, where treatment 1 is the treatment shown farthest to the left in the column headed “No. of patients” and treatment 2 is the treatment shown to the right, in the column headed “No. of patients”.

Psychological Treatments for Adults with PTSD Question: Is cognitive processing therapy efficacious for adults with PTSD? [Evidence from studies comparing cognitive processing therapy with inactive controls] Bibliography: Chard et al., 2005 (71); Forbes et al., 2012 (74); Monson et al., 2006 (70); Resick et al., 2002 (72); Resick, et al., 2003 (77); Resick, et al., 2012 (78).

Quality assessment No of patients Effect Quality

No of studies Design Risk of bias Inconsistency Indirectness Imprecision Other considerations CPT Control Relative

(95%

CI)

Absolute

PTSD symptom reduction: mean change from baseline to end of treatment (follow-up 6-17 weeks; measured with: CAPS; range of scores: 0-136; Better indicated by lower values) 4 randomised trials serious1 no serious inconsistency no serious indirectness no serious imprecision2 none 158 141 - MD 32.2 lower (46.3 to 18.1 lower)3

SMD 1.40 lower (1.95 lower to 0.85 lower)

⊕⊕⊕Ο

MODERATE

Loss of diagnosis (follow-up 6-17 weeks) 4 randomised trials serious1 no serious inconsistency no serious indirectness no serious imprecision4 none 87/158 (55.1%)

14/141 (9.9%)

- 44 more per 100 (from 26 more to 62 more)5

⊕⊕⊕Ο

MODERATE

Prevention or reduction of comorbid depression (follow-up 6-17 weeks; measured with: BDI, mean change from baseline to end of treatment; range of scores: 0-63; Better indicated by lower values) 4 randomised trials serious1 no serious inconsistency no serious indirectness no serious imprecision none 158 141 - MD 10.7 lower (16.5 to 4.9 lower)6

SMD: 1.04 lower (1.60 lower to 0.48 lower)

⊕⊕⊕Ο

MODERATE

Prevention or reduction of comorbid anxiety7 (follow-up 10-12 weeks; measured with: STAI; Better indicated by lower values) 2 randomised trials serious1 serious8 no serious indirectness serious8 none 60 59 - not pooled7 ⊕ΟΟΟ

VERY LOW

Quality of life7,9 (follow-up 12 weeks; measured with: ADAS and the short form of the WHOQOL) 1 randomised trials serious1 very serious9 no serious indirectness very serious9 none 30 29 - not pooled7,9 ⊕ΟΟΟ

VERY LOW

Note: The table does not include rows for outcomes for which we found zero studies (remission, disability of functional impairment, adverse events leading to withdrawals, and other serious adverse events).

1 Included trials rated as medium risk of bias.

2 Although the meta-analysis had considerable statistical heterogeneity (I2=86.5%), the direction of effects were consistent, the differences were only in the magnitude of benefit; all trials found moderate or large magnitudes of benefit. There was no serious imprecision, but the confidence interval was not precise enough to determine whether the magnitude of benefit is moderate or large.

3 When only including the 3 studies that compared CPT with waitlist: MD -35.9, 95% CI: -52.8 to -18.9, N=240 (128 in treatment groups; 112 in control groups).

4 The meta-analyses for loss of diagnosis and for BDI had considerable statistical heterogeneity, but the direction of effects were consistent, the differences were only in the magnitude of benefit; all trials found moderate or large magnitudes of benefit.

5 When only including the 3 studies that compared CPT with waitlist: RD 0.52, 95% CI: 0.37 to 0.67, N=240 (128 in treatment groups; 112 in control groups).

6 When only including the 3 studies that compared CPT with waitlist: MD -11.9, 95% CI: -18.9 to -4.9, N=240 (128 in treatment groups; 112 in control groups).

7 Actual data not reported.

8 Conflicting results: One found cognitive processing therapy to be no more effective in reducing symptoms of anxiety than waitlist (Monson et al., 2006); the other found greater improvement in anxiety for subjects treated with cognitive processing therapy than those receiving usual treatment from intake to posttreatment (p=0.018). (Forbes et al., 2012).

9 Single study.

Abbreviations: ADAS = Abbreviated Dyadic Adjustment Scale; BDI = Beck Depression Inventory; CAPS = Clinician-Administered PTSD Scale; CI = confidence interval; CPT = cognitive processing therapy; MD = mean difference; No= number; PTSD = post-traumatic stress disorder; RCT = randomized controlled trial; RD = risk difference; STAI = State-Trait Anxiety Inventory; WHOQOL = World Health Organization Quality of Life Scale.

Question: Is cognitive therapy (not including cognitive processing therapy) efficacious for adults with PTSD?

[Evidence from studies comparing cognitive therapy with inactive controls] Bibliography: Ehlers et al., 2003 (73); Ehlers et al., 2005 (52); Mueser et al., 2008 (75).

bias Inconsistency Indirectness Imprecision Other considerations Cognitive

Therapy (not including CPT)

Control Relative

(95%

CI)

Absolute

PTSD symptom reduction: mean change from baseline to end of treatment (follow-up 4-12 weeks1; measured with: CAPS; range of scores: 0-136; Better indicated by lower values) 3 randomised trials serious2 no serious inconsistency3 no serious indirectness no serious imprecision3 none 96 125 - SMD 1.22 lower (1.91 lower to 0.53 lower)4

⊕⊕⊕Ο

MODERATE

Loss of diagnosis (follow-up 4-12 weeks1) 3 randomised trials serious2 no serious inconsistency3 no serious indirectness no serious imprecision3 none 67/96 (69.8%)

37/125 (29.6%)

- 51 more per 100 (from 24 more to 78 more)5,6

⊕⊕⊕Ο

MODERATE

Prevention or reduction of comorbid depression (follow-up 4-12 weeks1; measured with: BDI, mean change from baseline to end of treatment; range of scores: 0-63;

Better indicated by lower values) 3 randomised trials serious2 no serious inconsistency no serious indirectness no serious imprecision3 none 96 125 - SMD 0.91 lower (1.20 to 0.62 lower)7

⊕⊕⊕Ο

MODERATE

Prevention or reduction of comorbid anxiety (follow-up 4-12 weeks1; measured with: State-Trait Anxiety Inventory, mean change from baseline to end of treatment; range of scores: 20-80; Better indicated by lower values) 3 randomised trials serious2 no serious inconsistency no serious indirectness no serious imprecision3 none 96 125 - SMD 0.93 lower (1.36 to 0.50 lower)8

⊕⊕⊕Ο

MODERATE

Quality of life (follow-up 12-16 sessions; measured with: SF-12; range of scores: 0-100; Better indicated by higher values) 1 randomised trials serious2 serious9 no serious indirectness serious10 none 54 54 - Better QoL outcomes for CT group than usual care group for the Physical Component (p=0.002), but not for the Mental Component (p=0.13)

⊕ΟΟΟ

VERY LOW

Disability or functional impairment (follow-up 4-12 weeks11; measured with: SDS, change from baseline; range of scores: 0-30; Better indicated by lower values) 2 randomised trials serious2 no serious inconsistency no serious indirectness no serious imprecision3 none 42 71 - SMD 1.13 lower (1.76 to 0.51 lower)12

⊕⊕⊕Ο

MODERATE

Note: The table does not include rows for outcomes for which we found zero studies (remission, adverse events leading to withdrawals, and other serious adverse events).

1 Ehlers et al., 2003, Mean of 9 weeks, 0 to 3 booster sessions (3, 6, and 9 months); Ehlers et al., 2005, 4 to 12 weeks plus up to 3 monthly boosters (3 and 6 months); Mueser et al., 2008 12 to 16 sessions, duration not specified 2 Included trials rated as medium risk of bias.

3 Direction of effects were consistent; magnitude of effects ranged from very large to small. I2 79.6%.

4 When only including the 2 studies that compared CT with waitlist: SMD -1.54, 95% CI: -2.17 to -0.92, N = 85 (42 in treatment groups; 43 in control groups).

5 The estimate of effect is reported as risk difference (RD).

6 When only including the 2 studies that compared CT with waitlist RD 0.66, 95% CI 0.5 to 0.82, N = 42 in treatment groups; 43 in control groups) 7 For all three studies, WMD -8.34, 95% CI: -10.8 to -5.85. When only including the 2 studies that compared CT with waitlist SMD -1.06, 95% CI: -1.52 to -0.60, N=113 (42 in treatment groups; 43 in control groups).

8 For all three studies, WMD -9.22, 95% CI: -11.9 to -6.5. When only including the 2 studies that compared CT with waitlist SMD -1.20, 95% CI: -1.67 to -0.73) N=113 (42 in treatment groups; 43 in control groups).

9 Consistency is unknown, single study 10 Wide CI including clinically distinct conclusions.

11 Ehlers et al., 2003, Mean of 9 weeks, 0 to 3 booster sessions (3, 6, and 9 months); Ehlers et al., 2005, 4 to 12 weeks plus up to 3 monthly boosters (3 and 6 months) 12 For all comparisons, WMD -2.66, 95% CI: -4.0 to -1.33. When only including comparison with waitlist, SMD -1.41, 95% CI: -2.41 to -0.41, N = 85 (42 in treatment groups; 43 in control groups).

Abbreviations: BDI = Beck Depression Inventory; CAPS = Clinician-Administered PTSD Scale; CI = confidence interval; CPT = cognitive processing therapy; CT = cognitive therapy;

MD = mean difference; No = number; PTSD = post-traumatic stress disorder; RCT = randomized controlled trial; RD = risk difference; SDS = Sheehan Disability Scale; SF-12 = Medical Outcome Study Self-Report; SMD = standardized mean difference; WMD = weighted mean difference.

Question: Is stress inoculation training (SIT) efficacious for adults with PTSD? [Evidence from studies comparing SIT with waitlist] Bibliography: Foa et al., 1999 (49); Zoellner et al., 1999 (81).

bias Inconsistency Indirectness Imprecision Other considerations Stress

Inoculation Training

Control Relative

(95%

CI)

Absolute

PTSD symptom reduction (follow-up 9 weeks; measured with: PSS-I; range of scores: 0-51; Better indicated by lower values) 1 randomised trials serious1 serious2 no serious indirectness serious3 none 26 15 - Baseline PSS-I: 29.4 vs. 32.9 for

WL; Endpoint 12.9 vs. 26.9;

p<0.05

⊕ΟΟΟ

VERY

LOW

Loss of diagnosis (follow-up 9 weeks) 1 randomised trials serious1 serious2 no serious indirectness serious3 none 26

15 - 42% vs. 0%, p<0.001 ⊕ΟΟΟ

VERY

LOW

Prevention or reduction of comorbid depression (follow-up 9 weeks; measured with: BDI; range of scores: 0-63; Better indicated by lower values) 1 randomised trials serious1 serious2 no serious indirectness serious3 none 26 15 - Baseline: SIT 21.7 vs. WL 25.2;

Endpoint 10.1 vs. 22.1; p<0.05

⊕ΟΟΟ

VERY

LOW

Prevention or reduction of comorbid anxiety (follow-up 9 weeks; measured with: STAI; range of scores: 20-80; Better indicated by lower values) 1 randomised trials serious1 serious2 no serious indirectness serious3 none 26 15 - Baseline: SIT 51.5 vs. WL 51.4;

Endpoint: 39.1 vs. 50.4; p=0.14

⊕ΟΟΟ

VERY

LOW

Note: The table does not include rows for outcomes for which we found zero studies (remission, quality of life, disability or functional impairment, adverse events leading to withdrawals, and other serious adverse events).

1 Included trial rated as medium risk of bias 2 Consistency is unknown, single study 3 Only 1 small study available. Does not meet optimal information size.

Abbreviations: BDI= Beck Depression Inventory; CI = confidence interval; PSS-I = Posttraumatic Stress Disorder Symptom Scale-Interview; PTSD = post-traumatic stress disorder;

SIT = stress inoculation training; STAI = State-Trait Anxiety Inventory; WL = waitlist.

Question: Is relaxation efficacious for adults with PTSD? [Evidence from studies comparing relaxation with treatment as usual] Bibliography: Carlson et al., 1998 (44).

bias Inconsistency Indirectness Imprecision Other considerations Relaxation Control Relative

(95%

CI)

Absolute

PTSD symptom reduction (follow-up 9 weeks; measured with: Global self rating scale; range of scores: 0-10; Better indicated by lower values) 1 randomised trials serious1 serious2 no serious indirectness serious3 none 13 12 - No benefit found for 3 different measures

⊕ΟΟΟ

VERY

LOW

Loss of diagnosis (follow-up 9 weeks) 1 randomised trials serious1 serious2 no serious serious3 none 13

12 - Trial did not report data for the inactive comparator group

⊕ΟΟΟ

VERY

LOW

Prevention or reduction of comorbid depression (follow-up 9 weeks; measured with: BDI; range of scores: 0-63; Better indicated by lower values) 1 randomised trials serious1 serious2 no serious indirectness serious3 none 13 12 - Results trend in favor of relaxation ⊕ΟΟΟ

VERY

LOW

Prevention or reduction of comorbid anxiety (follow-up 9 weeks; measured with: STAI; range of scores: 20-80; Better indicated by lower values) 1 randomised trials serious1 serious2 no serious indirectness serious3 none 13 12 - Results trend in favor of relaxation ⊕ΟΟΟ

VERY

LOW

Note: The table does not include rows for outcomes for which we found zero studies (remission, quality of life, disability or functional impairment, adverse events leading to withdrawals, and other serious adverse events).

1 Included trial rated as medium risk of bias.

2 Consistency is unknown, single study.

3 Only 1 small study available. Does not meet optimal information size.

Abbreviations: BDI = Beck Depression Inventory; CI= confidence interval; No = number; PTSD = post-traumatic stress disorder; STAI = State-Trait Anxiety Inventory.

Question: Are exposure-based therapies efficacious for adults with PTSD? [Evidence from studies comparing exposure-based therapies with inactive controls (waitlist or usual care)] Bibliography: Asukai et al., 2010 (89); Basoglu et al., 2007 (88); Foa et al, 1999 (49); Zoellner et al., 1999 (81); Foa et al., 2005 (25); Nacasch et al, 2011 (90); Resick et al., 2002 (72); Resick et al., 2003 (77); Resick et al., 2012 (78);

Rothbaum et al., 2005 (87).

bias Inconsistency Indirectness Imprecision Other considerations Exposure-

Based Therapies

Control Relative (95% CI) Absolute

PTSD symptom reduction: (follow-up 4.5 - 15 weeks1; measured with: CAPS or PSS-I; Better indicated by lower values) 7 randomised trials serious2 no serious inconsistency no serious indirectness no serious imprecision

Magnitude of effect

233 154 - SMD 1.27 lower (1.54 to 1 lower) ⊕⊕⊕⊕

HIGH

Remission (no longer having symptoms) (follow-up 10 weeks) 1 randomised trials serious2 serious3 no serious indirectness serious4 none 21/141

(14.9%)

10/143 (7%)

OR 2.43

(1.1 to 5.37)

85 more per 1000 (from 6 more to 218 more)

⊕ΟΟΟ

VERY LOW

Loss of diagnosis (follow-up 4.5 - 9 weeks) 3 randomised trials serious2 no serious inconsistency5 no serious indirectness serious6 Magnitude of effect

71/111 (64%)

3/86 (3.5%)

- 66 more per 100 (from 42 more to 91 more)7

⊕⊕⊕Ο

MODERATE

Prevention or reduction of comorbid depression (follow-up 4.5 - 15 weeks8; measured with: BDI; range of scores: 0-63; Better indicated by lower values) 6 randomised trials serious2 no serious inconsistency no serious indirectness no serious imprecision none 221 142 - MD 8.2 lower (10.3 to 6.1 lower) SMD 0.83 lower (1.05 lower to 0.60 lower)

HIGH

Disability or functional impairment (follow-up 8 -12 weeks; Better indicated by lower values) 2 randomised trials serious2 serious9 no serious indirectness serious6 none 180 41 - Cohen’s d 0.8 at 4 weeks, 0.6 at 8 weeks from one trial (N=31); numerically greater improvements on the Social Adjustment

Scale for another trial (N=190) in the exposure and exposure plus cognitive restructuring groups than for waitlist, but differences were not statistically significant.

⊕ΟΟΟ

VERY LOW

Note: The table does not include rows for outcomes for which we found zero studies (prevention or reduction of comorbid anxiety, quality of life, adverse events leading to withdrawals, and other serious adverse events).

1 Timing of outcome assessment: 8 weeks (Basoglu, 2007), 12 weeks (Foa, 2005), 9 weeks (Foa, 1999), 6 weeks (Resick, 2002), 4.5 weeks (Rothbaum, 2005), “post-treatment” or 8 to 15 weeks (Asukai, 2010), 9 to 15 weeks (Nacasch, 2011).

2 Included trials rated as medium risk of bias.

3 Consistency is unknown, single study.

4 Only one study with a wide confidence interval 5 Although there was some inconsistency in the magnitude of benefit, we did not downgrade the SOE because the direction of effect was the same and the magnitude of benefit was large or very large for the three studies.

6 Very wide confidence interval.

7 The estimate of effect is reported as risk difference (RD).

8 Timing of outcome assessment: 8 weeks (Basoglu, 2007), 12 weeks (Foa, 2005), 9 weeks (Foa, 1999), 6 weeks (Resick, 2002), 4.5 weeks (Rothbaum, 2005), 9 to 15 weeks (Nacasch, 2011) 9 One of the two studies did not find statistically significant differences between the intervention and control groups.

10 One study comparing prolonged exposure, prolonged exposure plus cognitive restructuring, and waitlist (Foa et al., 2005) reported that 4 participants had severe depression and suicidal ideation, but did not report which groups the individuals were in.

Abbreviations: BDI = Beck Depression Scale; CAPS = Clinician-Administered PTSD Scale; CI = confidence interval; MD = mean difference; No = number; OR = odds ratio; PSS-I = PTSD Symptom Scale Interview; PTSD = post-traumatic stress disorder; RD = risk difference; SMD = standardized mean difference.

Question: Are CBT interventions efficacious for adults with PTSD? [Evidence from studies comparing CBT therapies with inactive controls (waitlist, usual care)] Bibliography: Blanchard et al., 2003 (50); Bryant et al., 2003 (63); Cloitre et al., 2002 (51); Cottraux et al., 2008 (64);

Fecteau et al., 1999 (53); Foa et al., 2005 (25); Foa et al, 1999 (49); Zoellner et al., 1999 (81); Hinton et al., 2005 (54);

Hinton et al., 2009 (68); Hollifield et al., 2007 (55); Johnson et al., 2011 (61); Kubany et al., 2003 (56); Kubany et al., 2004 (69); Kruse et al., 2009 (62); Liedl et al., 2011 (57); Litz et al., 2007 (65); McDonagh et al., 2005 (58); Spence et al., 2011 (59); van Emmerik et al., 2008 (60).

bias Inconsistency Indirectness Imprecision Other considerations

CBT

Interventions Control Relative (95% CI) Absolute

PTSD symptom reduction (follow-up 4 - 14 weeks1; measured with: CAPS; range of scores: 0-136; Better indicated by lower values) 8 randomised trials serious2 some inconsistency

(I2=87%) no serious indirectness no serious imprecision

Magnitude of effect

246 230 - MD 31.1 lower (42.6 to

19.6 lower)3

SMD 1.36 lower (1.87 lower to 0.85 lower)

⊕⊕⊕Ο

MODERATE

PTSD symptom reduction (follow-up 4 - 21 weeks4; measured with: various; Better indicated by lower values) 14 randomised trials serious2 some inconsistency

(I2=75%) no serious indirectness no serious imprecision

Magnitude of effect

454 371 - SMD 1.09 lower (1.4 to

0.78 lower)5

⊕⊕⊕Ο

MODERATE

Remission (no longer having symptoms)6 2 randomised trials serious2 no serious inconsistency no serious indirectness no serious imprecision none - - not pooled6 not pooled6 ⊕⊕⊕Ο

MODERATE

Loss of diagnosis (follow-up 8 - 16 weeks) 6 randomised trials serious2 no serious inconsistency no serious indirectness no serious imprecision none 58/152 (38.2%)

19/138 (13.8%)

- 26 more per 100 (from 11 more to 41 more)8

⊕⊕⊕Ο

MODERATE

Prevention or reduction of comorbid depression (follow-up 4 - 20 weeks9; measured with: BDI; range of scores: 0-63; Better indicated by lower values) 10 randomised trials serious2 no serious inconsistency no serious indirectness no serious imprecision none 371 291 - MD 10.4 lower (14.4 to

6.4 lower)10

SMD 0.93 lower (1.27 lower to 0.59 lower

⊕⊕⊕Ο

MODERATE

Prevention or reduction of comorbid anxiety (follow-up 8 - 14 weeks11; measured with: STAI; range of scores: 20-80; Better indicated by lower values) 4 randomised trials serious2 no serious inconsistency no serious indirectness serious12 none 110 83 - MD 11.21 lower (20.04 to 2.37 lower)13

SMD 0.88 lower (1.54 lower to 0.22 lower)

⊕⊕ΟΟ

LOW

Quality of life 3 randomised trials serious2 serious14 no serious indirectness serious14 none 0 - - not pooled ⊕ΟΟΟ

VERY LOW

Question: Are CBT interventions efficacious for adults with PTSD? [Evidence from studies comparing CBT therapies with inactive controls (waitlist, usual care)] (continued) Disability or functional impairment (follow-up 8 - 52 weeks; Better indicated by lower values) 5 randomised trials serious2 no serious inconsistency no serious indirectness serious15 none 140 182 - not pooled ⊕⊕ΟΟ

LOW

Adverse events leading to withdrawals 316 No evidence available serious2 serious no serious serious none -

- - not pooled16

⊕ΟΟΟ

VERY LOW

Other serious adverse events (follow-up 12 weeks) 117 randomised trials serious2 serious no serious serious none -

- - not pooled17

⊕ΟΟΟ

VERY LOW

1 Timing of outcome assessment: 7 weeks (Johnson et al., 2011), 8 to 12 weeks (Blanchard, 2003), 12 weeks (Cloitre et al., 2002), 4 weeks (Fecteau et al., 1999), 12 weeks (Hinton et al., 2005), 4 to 5.5 weeks (Kubany et al., 2004), 4.5 months (Kubany et al., 2003), 14 weeks (McDonagh et al., 2005).

2 Included trials rated as medium risk of bias.

3 When only including the 7 studies that compared CBT with waitlist, WMD -34.4, 95% CI: -45.5 to -23.2, N=406 (211 in treatments groups and 195 in control groups).

4 Timing of outcome assessment: 7 weeks (Johnson et al., 2011), 8 to 12 weeks (Blanchard et al., 2003), 12 weeks (Cloitre et al., 2002), 4 weeks (Fecteau et al., 1999), 12 weeks (Foa et al., 2005), 9 weeks (Foa et al., 1999), 12 weeks (Hinton et al., 2005), 12 weeks (Hollifield et al., 2007) 4 to 5.5 weeks (Kubany et al., 2004), 4.5 months (Kubany et al., 2003), 4.8 months (Liedl et al., 2011), 14 weeks (McDonagh et al., 2005), 8 weeks (Spence et al., 2011), 5 sessions (van Emmerik et al., 2008).

5 When only including the 13 studies that compared CBT with waitlist, SMD -1.16, 95% CI: -1.47 to -0.84, N=755 (419 in the treatment groups, 336 in the control groups).

6 Two trials used different measures for remission, both found a greater percentage of subjects achieving remission: 61% vs. 21%, p=NR using the PCL; 82.4% vs. 0%, P<0.001 using the HTQ. (Spence et al., 2011; Kruse et al., 2009) 7 When only including 3 studies that compared CBT with waitlist RD 0.29, 95% CI: -0.01 to 0.60, N =147 (77 in treatment groups, 70 in control groups).

8 Estimate of effect is reported as risk difference (RD).

9 Timing of outcome assessment: 8 to 12 weeks (Blanchard, 2003), 12 weeks (Cloitre et al., 2002), 4 weeks (Fecteau et al., 1999), 12 weeks (Foa et al., 2005), 9 weeks (Foa et al., 1999), 4.5 months (Kubany et al., 2003), 14 weeks (McDonagh et al., 2005), 5 sessions (van Emmerik et al., 2008), 4 to 5.5 weeks (Kubany et al., 2004), 7 weeks (Johnson et al., 2011).

10 When only including the 9 studies that compared CBT with waitlist WMD -10.4, 95% CI -14.91 to -5.89, N=592 (336 in the treatment groups and 256 in the control groups).

11 Timing of outcome assessment: 12 weeks (Cloitre et al., 2002), 9 weeks (Foa et al., 1999), 14 weeks (McDonagh et al., 2005), 8 weeks (Bryant et al., 2003).

12 Wide confidence interval containing different clinical conclusions.

13 When only including the 3 studies that compared CBT with waitlist, WMD -12.1, 95% CI: -24.11 to -0.12, N = 155 (90 in the treatment groups, 65 in the control groups).

14 One trial (N=60) found no significant difference in change from baseline on the Marks’ Quality of Life Scale (-6.7 vs. -9.6, p=0.26); 64 another found no difference in change from baseline on the Quality of Life Inventory (QOLI: 3.4 vs. 0.4, p=0.63). (McDonagh et al., 2005). One trial (N=70) enrolling Bosnian refugees reported positive effect sizes for both the mental and the physical component summary scales of the SF-36 for CBT subjects compared with usual care subjects (Cohen’s d = 2.1 vs. -0.1, p<0.001, and 1.4 vs. 0.2, p<0.001, respectively). (Kruse et al., 2009; Meuser et al., 2008).

15 Four of the five trials compared CBT interventions with waitlist controls; one compared a CBT intervention with standard care. All four trials with waitlist controls found greater improvements in disability or functional outcomes for subjects who received CBT interventions—all but one reached statistical significance, (Spence et al., 2011) (p=0.07). The trial that compared CBT with standard care found similar changes in both groups. (Cottraux et al., 2008) Taken together, results suggest CBT interventions are efficacious for reducing disability and functional impairment; SOE was low because of some inconsistency and imprecision (low SOE).

16 Just 3 of the 23 included trials of CBT interventions (Blanchard et al., 2003; Cottraux et al., 2008; Hollifield et al., 2007) reported any information about withdrawals due to adverse events. With such a small proportion of trials reporting data, evidence was insufficient to draw conclusions. Any conclusions would be highly subject to reporting bias.

17 The trial reported one postrandomization removal from the trial due to death caused by an unrelated medical condition, but the trial did not report which group the death was in. (Foa et al., 2005)

Abbreviations: BDI = Beck Depression Scale; CAPS = Clinician-Administered PTSD Scale; CBT = Cognitive behavioral therapy; CI = confidence interval; MD = mean difference; No = number; PTSD = post-traumatic stress disorder; RD = risk difference; SMD = standardized mean difference; STAI = State-Trait Anxiety Inventory.

Question: Is EMDR efficacious for adults with PTSD? [Evidence from studies comparing EMDR with inactive controls (waitlist, usual care)] Bibliography: Hogberg et al., 2007 (111); Rothbaum et al., 1997 (112); Rothbaum et al., 2005 (87); Carlson et al., 1998 (44).

bias Inconsistency Indirectness Imprecision Other considerations EMDR Inactive controls (waitlist, usual care)

Relative (95%

CI)

Absolute

PTSD Symptom Reduction (follow-up 4-8 weeks; Better indicated by lower values) 4 randomised trials serious1 serious2 no serious indirectness serious Magnitude of effect size 60 57 - SMD 1.08 lower (1.83 lower to 0.33 lower)3

⊕⊕ΟΟ

LOW

Loss of Diagnosis (follow-up 4-8 weeks) 3 randomised trials serious1 no serious inconsistency no serious indirectness no serious imprecision none 39/50 (78%)

8/45 (17.8%)

- 64 more per 100 (from 46 more to 81 more)4

⊕⊕⊕Ο

MODERATE

Prevention/reduction of comorbid depression (follow-up 4-8 weeks; Better indicated by lower values) 4 randomised trials serious1 no serious inconsistency no serious indirectness no serious imprecision none 60 57 - SMD 1.13 lower (1.52 lower to 0.74 lower)

⊕⊕⊕Ο

MODERATE

Prevention/reduction of comorbid anxiety (follow-up 4-8 weeks; measured with: STAI; Better indicated by lower values) 3 randomised trials serious1 serious2 no serious indirectness very serious5 none 47 46 - MD 11.08 lower (23.06 lower to 0.90 higher)6

SMD 0.82 lower (1.77 lower to 0.13 higher)

⊕ΟΟΟ

VERY LOW

Adverse events leading to withdrawals (follow-up 2 months) 1 randomised trials serious1 serious7 no serious indirectness very serious8 none 1/13

(7.7%)

0/11 (0%)

- - ⊕ΟΟΟ

VERY LOW

Note: The table does not include rows for outcomes for which we found zero studies (remission, quality of life, disability/functional impairment, and other serious adverse events).

1 Included trials rated as medium risk of bias.

2 Conflicting findings across studies.

3 SMD -1.37 (-2.05, -0.69) when only compared with WL (3 trials, N = 95).

4 NNT 2

5 Wide CI that contains clinically distinct conclusions.

6 No statistically significant difference.

7 Single study.

8 The number of events is low. Does not meet optimal information size.

Abbreviations: CI = confidence interval; EMDR = eye movement desensitization and processing; MD = mean difference; No = number; PTSD = post-traumatic stress disorder; SMD = standardized mean difference; SOE = strength of evidence; STAI = State-Trait Anxiety; WL = waitlist.

Question: Is imagery rehearsal therapy (IRT) efficacious for adults with PTSD? [Evidence from studies comparing IRT with waitlist] Bibliography: Krakow et al., 2001 (121).

bias Inconsistency Indirectness Imprecision Other considerations Imagery rehearsal therapy (IRT) Waitlist Relative

(95%

CI)

Absolute

PTSD Symptom Reduction (follow-up 3 sessions1; measured with: CAPS; Better indicated by lower values) 1 randomised trials serious2 serious3 no serious indirectness very serious4 none 88 80 - -32.3 vs. -11.3, p=0.001 ⊕ΟΟΟ

VERY

LOW

Prevention/reduction of…