What is the effectiveness of physiotherapistdirected exercise, advice, or both for subacute low back pain? Randomised controlled trial. 7 university hospitals and primary care clinics in Australia and New Zealand. 259 persons with non-specific, subacute low back pain lasting for at least 6 weeks, but no longer than 12 weeks. Participants were randomised to four groups: exercise and advice, exercise and sham advice, sham exercise and advice, or sham exercise and sham advice. 12 exercise or sham exercise sessions were delivered over 6 weeks. The exercise program included an individualised, progressive, submaximal program of aerobic exercise, stretches, functional activities, activities to build speed, endurance and co-ordination, and trunk and limb-strengthening exercises. Participants also received a home exercise program. The sham exercise intervention consisted of sham shortwave diathermy and sham pulsed ultrasosound. In weeks 1, 2 and 4 participants received advice or sham advice. Advice sessions aimed to encourage a graded return to normal activities. During the sham advice sessions the patients talked about their problems but received no advice. Participants were not informed whether their group allocation was active or sham for either intervention. Primary outcomes were average pain over the last week (0 to 10 scale), global perceived effect (–5 to 5 scale) and function (Patient Specific Functional Scale, 0 to 10) at 6 weeks and 12 months. The effect of exercise (the adjusted difference in outcomes between exercise and sham exercise groups) at 6 weeks was –0.8 points (95% CI –1.3 to –0.3 points) on the pain scale and 0.5 points (95% CI 0.1 to 1.0 points) on the global perceived effect scale. The effect of advice at 6 weeks was –0.7 points (95% CI –1.2 to –0.2 points) on the pain scale and 0.8 points (95% CI 0.3 to 1.2 points) on the global perceived effect scale. The effect of advice on the function scale was significant at 6 weeks and 12 months. For pain, function, and global perceived effect, the effect of combined treatments was larger than the effect of exercise or advice alone. Physiotherapydirected exercise and advice was slightly more effective than placebo at 6 weeks. The effect was greatest when the interventions were combined. At 12 months a small effect on participant-reported function was still reported.