Frequently Asked Questions

Yes. In this randomized controlled trial of adults with pharmacotherapy-resistant major depressive disorder receiving psychiatric specialty care, adding CBT to usual medication management produced greater improvement in clinician-rated depressive symptoms than treatment as usual alone at 16 weeks. The least squares mean change in GRID-HDRS17 score was −12.7 with CBT plus treatment as usual versus −7.4 with treatment as usual alone, for a between-group difference of −5.4 points (95% CI, −8.1 to −2.6; P < .001).

The benefit was not statistically significant at 8 weeks, but it was significant by 16 weeks. The study reported no difference in treatment effect at 8 weeks (P = .11), whereas by 16 weeks the CBT group had significantly greater improvement in GRID-HDRS17 scores than the treatment-as-usual group (difference = −5.4; 95% CI, −8.1 to −2.6; P < .001).

Yes. The study found that the effect of supplementary CBT on clinician-rated depressive symptoms was maintained for at least 12 months after randomization. The between-group differences in GRID-HDRS17 change remained significant at 3 months after treatment (−3.7; 95% CI, −6.4 to −0.9; P = .01), 6 months (−3.4; 95% CI, −6.2 to −0.6; P = .02), and 12 months (−4.4; 95% CI, −7.2 to −1.6; P = .002).

Yes. At 16 weeks, patients assigned to CBT plus treatment as usual were 2.38 times more likely to achieve treatment response than those receiving treatment as usual alone (RR = 2.38; 95% CI, 1.48 to 3.84), with a number needed to treat of 5 (95% CI, 3 to 83). Remission was also more likely with supplementary CBT at 16 weeks (RR = 2.13; 95% CI, 1.04 to 4.35; NNT = 5; 95% CI, 3 to 36), and remission benefit was maintained over 12 months except at 3 months, when the confidence interval included 1 (RR = 1.67; 95% CI, 0.95 to 2.93).

No clear between-group advantage was shown on the self-rated depression scales used in this trial. Although participants in the CBT group had milder depressive symptoms on the Beck Depression Inventory-II and the Quick Inventory of Depressive Symptomatology Self-Report over the study period, the differences were not statistically significant at each time point.

Not on the quality-of-life measure reported here. The study found no statistically significant between-group differences over time in SF-36 mental or physical component summary scores.

Participants assigned to CBT were offered 16 individual 50-minute sessions scheduled weekly, with up to 4 additional sessions if the therapist judged them clinically appropriate. Treatment as usual in both groups consisted of medication management, education about medication and dosing, review of adverse effects, supportive guidance, and QIDS-SR monitoring at visits that occurred roughly every 2 weeks and lasted 10 to 15 minutes.

The trial enrolled outpatients aged 20 to 65 years with DSM-IV major depressive disorder, a GRID-HDRS17 score of 16 or higher, a Maudsley Staging Method score of 3 or higher, and at least 8 weeks of antidepressant treatment at adequate therapeutic levels without sufficient improvement. A total of 80 patients were randomized, 78 (97.5%) were assessed for the primary outcome, and 73 (91.3%) completed 12-month follow-up.

The study reported no serious adverse events during the intervention period in either group, and nearly all patients in the CBT arm completed the full course of therapy, with a completion rate of 97.5%. During postintervention follow-up, 2 participants in the treatment-as-usual group were hospitalized for worsening depression, and 1 of those patients died by suicide shortly after discharge, 10 months after the intervention period ended.

The main limitations were a relatively small sample size, recruitment from only 2 sites, and very high adherence and low dropout, which may limit generalizability because the sample may have been especially motivated and treatment-seeking. The authors also noted that usual care did not allow full control of antidepressant medication, that there was no psychotherapy control condition, so nonspecific treatment effects such as expectations could have contributed, and that the study duration of 6 years means potential effects of time cannot be fully excluded.