Outcomes following a physician-led CBT skills group for anxiety and depression in primary care: a longitudinal cohort study

Cognitive Behavioural Therapy (CBT) is an evidence-based psychotherapy used to treat a range of mental health disorders. This study aimed to evaluate the long-term outcomes of a novel primary care-based CBT Skills Group program for anxiety and depression, using patient-reported measures at 3, 6, and 12 months post-intervention. This observational cohort study followed adult patients with mild to moderate anxiety and/or depression who completed an 8-week, primary care-based CBT Skills Group program. Participants were assessed using validated self-report measures (PHQ-8 and GAD-7) at baseline, program completion, and 3, 6, and 12-month follow-up. Reporting followed the STROBE guidelines for observational studies. Of 206 invited participants, 67 formed the analytic cohort. Paired complete-case analyses included 57 and 55 participants at 3 months, 43 and 45 at 6 months, and 26 and 25 at 12 months for the PHQ-8 and GAD-7, respectively. PHQ-8 and GAD-7 scores were significantly lower than baseline at all three follow-up points under complete-case analysis, last observation carried forward, and MICE. Under the more conservative extreme-value approach, only the GAD-7 improvement at 3 months remained significant, and at 12 months both measures showed a statistically significant change in the opposite (apparent deterioration) direction, reflecting the sensitivity of longer-term findings to assumptions about missing data amid substantial attrition. Qualitative responses at 12 months ( n = 25) suggested most respondents continued using CBT skills, though this reflects only those who remained engaged with follow-up. Group-based CBT Skills training in a primary care setting was associated with reductions in depression and anxiety symptoms, with the strongest evidence at three months; findings beyond three months varied by missing-data method. Observed-data, LOCF, and MICE analyses suggested continued benefit, but outcome availability declined to approximately 40% by 12 months and longer-term findings were highly sensitive to missing-data assumptions and should be considered exploratory rather than confirmatory. This model may offer a scalable, accessible option for mental health support in publicly funded healthcare systems, pending confirmation of longer-term outcomes with improved retention.