Seven studies (eight publications) were included: 6 RCTs and 1 controlled before-and-after study. The sample sizes ranged from 67 to 618 patients and from 20 to 286 care providers.
The methodological quality of the studies varied greatly. Methodological problems included lack of clarity about follow-up, unclear concealment of allocation, lack of a power calculation and lack of comparison of treatment groups at baseline. In all but one of the included studies, the outcome assessor was blind to the treatment allocation.
Organisational interventions (3 studies).
Anxiety measures (3 studies): improvements in the main problem and the fear questionnaire for global phobia, phobia and anxiety-depression were found with a nurse substitution intervention (1 study); more patients recovered (on the Panic Disorder Severity Scale and the Anxiety Severity Index) and there were more anxiety-free-days per patient with a collaborative care intervention (1 study); and there was no difference between collaborative care and control in recovery rates (on the Shedler Quick Diagnostics Panel) with the other collaborative care intervention (1 study).
Process-of-care measures (1 study): this study found no significant difference in the use of appropriate medication, adequate dosage duration, or the proportion of patients complying with medication with a collaborative care intervention.
Professional interventions (4 studies).
Anxiety measures (1 study): this study of audit and feedback reported no difference with the intervention when using the Global Severity Index, Highest Anxiety Subscale or SF-36, but reported greater improvements for Global Anxiety Score (SMD -0.18, 95% CI: -0.35, -0.01) and self-reported anxiety symptoms (RR 1.25, 95% CI: 1.01, 1.53).
Process-of-care measures for diagnosis (2 studies): the studies reported significantly increased recognition of symptoms and chart notation with an audit and feedback intervention (1 study), and significantly increased diagnosis of agoraphobia with panic attacks, panic disorder, generalised anxiety disorder and adjustment disorder with anxious mood with a brief education intervention (1 study).
Process-of-care measures for management (4 studies): 1 study found no increase in prescription rates with audit and feedback, while 2 studies reported no consistent effects on prescription rates with educational outreach interventions. The only study assessing referral reported that an audit and feedback intervention significantly increased referrals compared with a control intervention (RR 2.94, 95% CI: 1.33, 6.51).
In 3 of the 4 studies of professional interventions and in all 3 studies of organisational interventions, an external expert provided education or actively participated in the care of patients with anxiety disorders.