Table G13Cognitive behavioral therapy compared with supportive counseling

Outcome: Number of Studies; Number of Subjects; DesignRisk of BiasConsistencyDirectnessPrecisionMagnitude of Effect: Summary Effect Size (95% CI)Strength of Evidence
Incidence of PTSD at end of treatment: 3; 105; RCTMediumaConsistentDirectImprecisebCIDI-PTSD, CAPS, RR (95% CI) = 0.27 (0.05 to 1.29) favors CBTLow
Incidence of PTSD at 6 months: 3; 105; RCTMediumaConsistentDirectImprecisecCIDI-PTSD, CAPS, RR (95% CI) = 0.46 (0.21 to 1.01) favors CBTLow
PTSD symptom reduction at end of treatment: 3; 105; RCTMediumaConsistentDirectPrecisedIES-I, WMD −7.85 (−11.18 to −4.53) favors CBTModerate
PTSD symptom reduction at 6 months: 3; 105; RCTMediumaConsistentDirectPreciseeIES-I, WMD −8.19 (−11.79 to −4.58) favors CBTModerate
PTSD symptom reduction at end of treatment: 3; 105; RCTMediumaConsistentDirectPrecisefIES-A, WMD −14.04 (−19.37 to −8.71) favors CBTModerate
PTSD symptom reduction at 6 months: 3; 105; RCTMediumaConsistentDirectPrecisegIES-A, WMD −9.94 (−15.06 to −4.83) favors CBTModerate
Depression symptom reduction at end of treatment: 3; 105; RCTMediumaInconsistentDirectImprecisehBDI-2, SMD −0.15 (−0.53 to 0.24)Low
Depression symptom reduction at 6months: 3; 105; RCTMediumaInconsistentDirectImpreciseiBDI-2, SMD −0.21 (−0.70 to 0.27)Low
Anxiety symptom reduction at end of treatment: 3; 105; RCTMediumaConsistentDirectImprecisejBAI, STAI, SMD −0.25 (−0.64 to 0.13)Moderate
Anxiety symptom reduction at 6 months: 3; 105; RCTMediumaConsistentDirectImprecisekBAI, STAI, SMD −0.28 (−0.67 to 0.11)Moderate
Incidence/severity of comorbid conditions: 0; 0NANANANANAInsufficient
Quality of Life: 0; 0NANANANANAInsufficient
Return to work/return to active duty or ability to work: 0; 0NANANANANAInsufficient
Incidence of self-injurious or suicidal thoughts, attempts, or behaviors (including suicide): 0; 0NANANANANAInsufficient
Incidence of aggressive or homicidal thoughts, attempts, or behaviors (including homicide) 0; 0NANANANANAInsufficient
Perceived utility: 0; 0NANANANANAInsufficient
a

Reasons for downgrading include lack of reporting on number of treatment sessions completed (1 study), high (> 20%) overall attrition rate (1 study), unclear randomization scheme (1 study)

b

Although the direction of effects was consistent, the meta-analysis had considerable statistical heterogeneity (I2=71.8%), reflecting the fact that two of the three medium risk of bias trials found large magnitudes of benefit but one medium risk of bias study found no difference between treatment groups. When we repeated the analysis including an additional high risk of bias study that found a small benefit, the heterogeneity was reduced (I2=58.78%). Even though the direction of effect was consistent across trials, we rated the findings as imprecise and thus graded the SOE as low rather than moderate.

c

Although the direction of effects was consistent, the meta-analysis had considerable statistical heterogeneity (I2=44.9%), reflecting the fact that two of the three medium risk of bias trials found large magnitudes of benefit but one medium risk of bias study found no difference between treatment groups. When we repeated the analysis including an additional high risk of bias study that found a small benefit, the heterogeneity was reduced (I2=32.0%). Even though the direction of effect was consistent across trials, we rated the findings as imprecise and thus graded the SOE as low rather than moderate.

d

The analysis found very low statistical heterogeneity (I2=1.3%) and a subsequent sensitivity analysis (n=136) including one high risk of bias study indicated a slightly larger benefit of CBT (WMD, −8.39; 95% CI, −11.45 to −5.34) with no statistical heterogeneity (I2=0.0%), increasing our confidence in the finding of a moderate effect size and finding a consistent, precise result.

e

The analysis found very low statistical heterogeneity (I2=6.8%). A subsequent sensitivity analysis (n=136) including one high risk of bias study indicated a slightly smaller benefit of CBT (WMD, −7.91; 95% CI, −10.85 to −4.98) with no statistical heterogeneity (I2=0.0%), reinforcing our confidence in the finding of a moderate effect size and a consistent, precise result.

f

Although the direction of the effect was consistent, the analysis found moderate statistical heterogeneity (I2=53.8%). A subsequent sensitivity analysis (n=136) including one high risk of bias study indicated a slightly larger benefit of CBT (WMD, −14.17; 95% CI, −17.82 to −10.51) with reduced statistical heterogeneity (I2=31.9%), reinforcing our confidence in the finding of a large effect size and a consistent, precise result.

g

Although the direction of the effect was consistent, the analysis found moderate statistical heterogeneity (I2=44.0%). A subsequent sensitivity analysis (n=136) including one high risk of bias study indicated a slightly larger benefit of CBT (WMD, −11.49; 95% CI, −16.09 to −6.90) albeit with greater statistical heterogeneity (I2=52.7%), which did not substantively change our confidence in the finding of a large effect size and a consistent, precise result.

h

The analysis found no statistical heterogeneity (I2=0.0%) and the direction of effect was not consistent across trials ranging from a very low effect size in favor of SC to a moderate effect size in favor of CBT. A subsequent sensitivity analysis (n=136) including one high risk of bias study indicated lower statistical heterogeneity (I2=0.0%) and a slightly larger but still insignificant benefit of CBT (SMD, −0.22; 95% CI, −0.56 to 0.12).

i

The analysis found moderate statistical heterogeneity (I2=30.0%) and the direction of effect was not consistent across trials. A subsequent sensitivity analysis (n=136) including one high risk of bias study indicated a slightly larger but still insignificant benefit of CBT (SMD, −0.25; 95% CI, −0.62 to 0.12) with low statistical heterogeneity (I2=10.1%).

j

The analysis found no statistical heterogeneity (I2=0.0%) and a subsequent sensitivity analysis (n=136) including one high risk of bias study indicated a slightly larger but still insignificant benefit of CBT (SMD, −0.39; 95% CI, −0.74 to −0.04) with very low statistical heterogeneity (I2=2.2%).

k

The analysis found no statistical heterogeneity (I2=0.0%) and a subsequent sensitivity analysis (n=136) including one high risk of bias study indicated a larger but still insignificant benefit of CBT (SMD, −0.59; 95% CI, −1.16 to −0.01) with very low statistical heterogeneity (I2=2.2%).

Abbreviations: BAI = Beck Anxiety Inventory; BDI-2 = Beck Depression Inventory-2; CAPS = Clinician Administered PTSD Scale; CBT = Cognitive behavioral therapy; CI = confidence interval; CIDI-PTSD = Composite International Diagnostic Interview PTSD Module; CT = Cognitive therapy; IES-A = Impact of Event-Avoidance subscale; IES-I = Impact of Event-Intrusion subscale; n = number of participants; NA = not applicable; NS = Not significant; OR = Odds ratio; PTSD = posttraumatic stress disorder; RCT = randomized controlled trial; RR = risk ratio; SC = Supportive counseling; SMD = standardized mean difference; SOE = strength of evidence; STAI = State-Trait Anxiety Inventory; WMD = weighted mean difference

Reasons for downgrading include lack of reporting on number of treatment sessions completed (1 study), high (> 20%) overall attrition rate (1 study), unclear randomization scheme (1 study)

Although the direction of effects was consistent, the meta-analysis had considerable statistical heterogeneity (I2=71.8%), reflecting the fact that two of the three medium risk of bias trials found large magnitudes of benefit but one medium risk of bias study found no difference between treatment groups. When we repeated the analysis including an additional high risk of bias study that found a small benefit, the heterogeneity was reduced (I2=58.78%). Even though the direction of effect was consistent across trials, we rated the findings as imprecise and thus graded the SOE as low rather than moderate.

Although the direction of effects was consistent, the meta-analysis had considerable statistical heterogeneity (I2=44.9%), reflecting the fact that two of the three medium risk of bias trials found large magnitudes of benefit but one medium risk of bias study found no difference between treatment groups. When we repeated the analysis including an additional high risk of bias study that found a small benefit, the heterogeneity was reduced (I2=32.0%). Even though the direction of effect was consistent across trials, we rated the findings as imprecise and thus graded the SOE as low rather than moderate.

The analysis found very low statistical heterogeneity (I2=1.3%) and a subsequent sensitivity analysis (n=136) including one high risk of bias study indicated a slightly larger benefit of CBT (WMD, −8.39; 95% CI, −11.45 to −5.34) with no statistical heterogeneity (I2=0.0%), increasing our confidence in the finding of a moderate effect size and finding a consistent, precise result.

The analysis found very low statistical heterogeneity (I2=6.8%). A subsequent sensitivity analysis (n=136) including one high risk of bias study indicated a slightly smaller benefit of CBT (WMD, −7.91; 95% CI, −10.85 to −4.98) with no statistical heterogeneity (I2=0.0%), reinforcing our confidence in the finding of a moderate effect size and a consistent, precise result.

Although the direction of the effect was consistent, the analysis found moderate statistical heterogeneity (I2=53.8%). A subsequent sensitivity analysis (n=136) including one high risk of bias study indicated a slightly larger benefit of CBT (WMD, −14.17; 95% CI, −17.82 to −10.51) with reduced statistical heterogeneity (I2=31.9%), reinforcing our confidence in the finding of a large effect size and a consistent, precise result.

Although the direction of the effect was consistent, the analysis found moderate statistical heterogeneity (I2=44.0%). A subsequent sensitivity analysis (n=136) including one high risk of bias study indicated a slightly larger benefit of CBT (WMD, −11.49; 95% CI, −16.09 to −6.90) albeit with greater statistical heterogeneity (I2=52.7%), which did not substantively change our confidence in the finding of a large effect size and a consistent, precise result.

The analysis found no statistical heterogeneity (I2=0.0%) and the direction of effect was not consistent across trials ranging from a very low effect size in favor of SC to a moderate effect size in favor of CBT. A subsequent sensitivity analysis (n=136) including one high risk of bias study indicated lower statistical heterogeneity (I2=0.0%) and a slightly larger but still insignificant benefit of CBT (SMD, −0.22; 95% CI, −0.56 to 0.12).

The analysis found moderate statistical heterogeneity (I2=30.0%) and the direction of effect was not consistent across trials. A subsequent sensitivity analysis (n=136) including one high risk of bias study indicated a slightly larger but still insignificant benefit of CBT (SMD, −0.25; 95% CI, −0.62 to 0.12) with low statistical heterogeneity (I2=10.1%).

The analysis found no statistical heterogeneity (I2=0.0%) and a subsequent sensitivity analysis (n=136) including one high risk of bias study indicated a slightly larger but still insignificant benefit of CBT (SMD, −0.39; 95% CI, −0.74 to −0.04) with very low statistical heterogeneity (I2=2.2%).

The analysis found no statistical heterogeneity (I2=0.0%) and a subsequent sensitivity analysis (n=136) including one high risk of bias study indicated a larger but still insignificant benefit of CBT (SMD, −0.59; 95% CI, −1.16 to −0.01) with very low statistical heterogeneity (I2=2.2%).

From: Appendix G, Psychological, Pharmacological, and Emerging Interventions: Strength-of-Evidence Grades

Cover of Interventions for the Prevention of Posttraumatic Stress Disorder (PTSD) in Adults After Exposure to Psychological Trauma
Interventions for the Prevention of Posttraumatic Stress Disorder (PTSD) in Adults After Exposure to Psychological Trauma [Internet].
Comparative Effectiveness Reviews, No. 109.
Gartlehner G, Forneris CA, Brownley KA, et al.

NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.