Objective: To investigate the effectiveness of a patient education (PE) program with or without the added effect of manual therapy (MT) compared to a minimal control intervention (MCI). Methods: In a single-center university hospital setting, a total of 118 patients with clinical and radiographic unilateral hip osteoarthritis (OA) from primary care were randomized into one of three groups: PE, PE plus MT or MCI. The PE was taught by a physiotherapist involving five sessions. The MT was delivered by a chiropractor involving 12 sessions and the MCI included a home stretching program. Primary outcome was self-reported pain severity on an 11-box numeric rating scale (NRS) immediately following a 6-week intervention period. Patients were followed for 1 year. Results: Primary analysis included 111 patients (94%). In the combined group (PE+MT), a clinically relevant reduction in pain severity compared to the MCI of 1.90 points (95% confidence interval (CI) 0.9-2.9) was achieved. Effect size (Cohen's d) for the PE+MT minus the MCI was 0.92 (95% CI 0.41-1.42). Number needed to treat for PE+MT was 3 (95% CI 2-7). No difference was found between the PE and MCI groups, with mean difference 0.0 (95% CI-1.0 to 1.0). At 12 months, not including patients receiving hip surgery the statistically significant difference favoring PE+MT was maintained. Conclusions: For primary care patients with OA of the hip, a combined intervention of MT and PE was more effective than a MCI. PE alone was not superior to the MCI. Trial registration: clinicaltrials.gov NCT01039337.