OBJECTIVE To evaluate the outcomes of children with acute osteomyelitis and septic arthritis at a hospital where short-duration antibiotic treatment (< or = 3.5 weeks) was considered routine. METHODOLOGY We carried out a retrospective chart review, with telephone interviews to follow up and determine long-term outcomes. Patients were selected to be at low risk for complications (illness < or = 14 days, no underlying disease, uncomplicated presentation). RESULTS Thirty-two children with osteomyelitis (OM), 34 with septic arthritis (SA) and five with OM and SA (OMSA) were included. Blood cultures were positive (mainly Staphylococcus aureus) in 15% of patients who had not had prior antibiotic treatment, and microbiological confirmation (positive blood culture, Gram stain or culture of surgical specimen) was obtained in 36%. The median duration of antibiotic treatment was 5.4, 4.4 and 5.0 weeks for OM, SA and OMSA, respectively. Only 22% of patients received antibiotics for 3.5 weeks or less. Overall, the recurrence rate was 1.4%. At follow-up, only two patients had mild occasional pain at the site of the original infection; all patients had normal function. CONCLUSIONS Contrary to expectations and local protocols, most patients were treated with conventional long-duration therapy. Patients treated for short courses had good outcomes. The low rate of complications may make randomized controlled equivalence trials unfeasible. Increasing evidence of the efficacy and safety of short-duration treatment (3-3.5 weeks) for acute, uncomplicated OM or SA in children suggests that this could be accepted as the standard treatment. However, this should be evaluated prospectively using a register, with at least 12 months' of follow-up.