Objectives To evaluate the effectiveness and cost-utility of the addition of different doses of Pilates to an advice for non-specific chronic low back pain (NSCLBP) from a societal perspective.

Design Randomised controlled trial with economic evaluation.

Setting Physiotherapy clinic in São Paulo, Brazil.

Participants 296 patients with NSCLBP.

Interventions All patients received advice and were randomly allocated to four groups (n=74 per group): booklet group (BG), Pilates once a week (Pilates group 1, PG1), Pilates twice a week (Pilates group 2, PG2) and Pilates three times a week (Pilates group 3, PG3).

Main outcome measures Primary outcomes were pain and disability at 6-week follow-up.

Results Compared with the BG, all Pilates groups showed significant improvements in pain (PG1, mean difference (MD)=−1.2, 95% CI −2.2 to −0.3; PG2, MD=−2.3, 95% CI −3.2 to −1.4; PG3, MD=−2.1, 95% CI −3.0 to −1.1) and disability (PG1, MD=−1.9, 95% CI −3.6 to −0.1; PG2, MD=−4.7, 95% CI −6.4 to −3.0; PG3, MD=−3.3, 95% CI −5.0 to −1.6). Among the different doses, PG2 showed significant improvements in comparison with PG1 for pain (MD=−1.1, 95% CI −2.0 to −0.1) and disability (MD=−2.8, 95% CI −4.5 to −1.1). The cost-utility analysis showed that PG3 had a 0.78 probability of being cost-effective at a willingness-to-pay of £20 000 per quality-adjusted life-year gained.

Conclusions Adding two sessions of Pilates exercises to advice provided better outcomes in pain and disability than advice alone for patients with NSCLBP; non-specific elements such as greater attention or expectation might be part of this effect. The cost-utility analysis showed that Pilates three times a week was the preferred option.

Trial registration number NCT02241538, Completed.