Objectives: To evaluate the effectiveness of a two session guided self-help (GSH) intervention provided by primary care graduate mental health workers (PCGMHWs) in a primary care mental health service. Design: Pragmatic randomized trial, with a wait list control design. Method: Patients presenting with significant anxiety and depression problems were given one or more self-help booklets at screening and randomly allocated to an immediate (ITG) or delayed treatment group (DTG). Following this, a two-session GSH intervention was provided by one of two PCGMHWs, with a review session to decide on the need for further intervention. The DTG began the intervention 8 weeks after the screening and the primary outcome was Clinical Outcomes in Routine Evaluation – Outcome Measure (CORE-OM) scores after 8 weeks. Results: A total of 63 patients were allocated to the ITG, 59 to the DTG. Analysis of covariance, carried out on an intention to treat basis, showed a significant treatment effect, F(1,98) = 15, p < .001, and a comparison of means at 8 weeks showed a significant difference, t(116) = 2.1 (95% CI [1.1, 5.9]), p= .042 with an effect size, d= 0.375. Taking the two groups together, CORE-OM scores for patients who completed the intervention reduced between screening and the review session by an average of 7.9 (95% CI [6.3, 9.5]), effect size of 1.2. Between screening and the review session, 47% showed a reliable and clinically significant improvement. Conclusions: The study provides some support for the effectiveness of a two-session GSH intervention and a stepped-care service model.