Osborn, Demoncada, Feuerstein, 2006

Meta-analysis investigating effects of cognitive behavioral therapy (CBT) and patient education (PE) on anxiety in adult cancer survivors

The search included databases from 1993-2004: Medline, PsychINFO, and the Cochrane Database. Search words used: cancer, anxiety, depression, quality of life (QOL), fatigue, stress, pain, physical function, social, self-management, evidence-based, interventions, and random/randomized.

Inclusion Criteria: adult patient with cancer (all types of cancer and all stages), control group, randomization, measurable outcomes of interest (anxiety, depression, fatigue, QOL, physical function, and pain) and at least one follow-up assessment beyond post-treatment, which allowed for examination of duration of effects.

Quality assessed by modified version of Jadad six-item checklist (randomization, double blinding, descriptions of withdrawals and dropouts, statistical analyses, inclusion and exclusion criteria, and adverse effects). Studies were excluded if not randomized or controlled, or had score less than four on checklist, or did not report follow-up data, or did not report data on targeted outcomes. Dissertations excluded

CBT interventions on anxiety (four studies) Individual and group CBT: Large effect noted (g=1.99, p 0.01; 95% CI 0.69-3.31) Of these four studies, a sensitivity analysis revealed a large effect size for individual treatment (g=2.41, p0.01; 95% CI 1.2-3.55) and no effect for group interventions (d=0.03, p+0.82; 95% CI -0.20-0.25) Forest plots representing the effect sizes of CBT on anxiety favor the intervention.

Significance to practice: CBT is effective for short-term management (8 months) of anxiety. Individually based interventions were more effective than those delivered in a group format. Various CBT approaches provided in an individual format can assist cancer survivors in reducing the emotional distress of anxiety.

The single trial using PE to decrease anxiety resulted in no short-term effect on anxiety and did not include long-term follow up on anxiety (d= -.02, p= 0.89; CI-0.36-0.31). Analysis did not consider patient adherence to pharmacologic interventions, which is known to be modest in medical patients.

No cost-benefit implications.

694