Older studies have suggested that antibiotic treatment for less than 21 days in children with osteoarticular infections is associated with relapse, and therefore traditionally these infections have been treated for six weeks with an intravenous phase of one week. However, four new prospective studies suggest that if the duration of treatment is individualized and made dependent on the clinical response and the C-reactive protein level, the treatment course may be shortened substantially without increasing the risk of relapse or sequelae. This article reviews older and newer literature on this subject.