IntroductionManagement of thumb base osteoarthritis (OA) using a combination of therapies is common in clinical practice; however, evidence for the efficacy of this approach is lacking. The aim of this study is to determine the effect of a combination of conservative therapies for the treatment of thumb base OA compared with an education control group.Methods and analysisThis is a randomised, controlled, single-centre, two-arm superiority trial with 1:1 allocation ratio; with assessor and statistician blinded. Participants are blinded to the trial's hypothesis and to the interventions received by the opposite group. A total of 204 participants will be recruited from the community and randomised using a computer-generated schedule. The intervention group will receive education for joint protection and OA, a splint for the base of the thumb, hand exercises and topical diclofenac sodium 1% gel over 6 weeks. The control group will receive education for joint protection and OA alone. Main inclusion criteria are pain ≥40 mm (Visual Analogue Scale, 0–100) at the base of the thumb, impairment in hand function ≥6 (Functional Index for Hand Osteoarthritis, 0–30) and radiographic thumb base OA (Kellgren Lawrence grade ≥2). Participants currently receiving any of the intervention components will be excluded. Outcomes will be measured at 2, 6 and 12 weeks. The primary outcome is change in pain and hand function from baseline to 6 weeks. Other outcomes include changes in grip and pinch strength, quality of life, presence of joint swelling and tenderness, duration of joint stiffness, patient's global assessment and use of rescue medication. Analysis will be performed according to the intention-to-treat principle. Adverse events will be monitored throughout the study.Ethics and disseminationThis protocol is approved by the local ethics committee (HREC/15/HAWKE/479). Dissemination will occur through presentations at international conferences and publication in peer-reviewed journals.Trial registration numberACTRN12616000353493; Pre-results.