Empiric therapy of methicillin-susceptible Staphylococcus aureus (MSSA) infections with vancomycin is associated with poorer outcome than targeted therapy with β-lactams. Our objective was to evaluate whether rapid determination of methicillin resistance shortens the time from Gram stain to targeted antimicrobial therapy in staphylococcal bacteraemia, thereby reducing vancomycin overuse. This was a single-centre open parallel RCT. Gram-positive cocci in clusters in positive blood culture underwent real-time PCR for rapid species and methicillin resistance determination parallel to conventional microbiology. Patients were randomized 1:1 so that clinicians would be informed of PCR results (intervention group) or not (control group). Eighty-nine patients (intervention 48, control 41) were analysed. MRSA was identified in seven patients, MSSA in 46, and CoNS in 36. PCR results were highly concordant (87/89) with standard microbiology. Median time (hours) from Gram stain to transmission of methicillin-susceptibility was 3.9 (2.8-4.3) vs. 25.4 (24.4-26-7) in intervention vs. control groups (p <0.001). Median time (hours) from Gram stain to targeted treatment was similar for 'all staphylococci' [6 (3.8-10) vs. 8 (1-36) p 0.13] but shorter in the intervention group when considering S. aureus only [5 (3-7) vs. 25.5 (3.8-54) p <0.001]. When standard susceptibility testing was complete, 41/48 (85.4%) patients in the intervention group were already receiving targeted therapy compared with 23/41 (56.1%) in the control group (p 0.004). There was no significant effect on clinical outcomes. Rapid determination of methicillin resistance in staphylococcal bacteraemia is accurate and reduces significantly the time to targeted antibiotic therapy in the subgroup of S. aureus, thereby avoiding unnecessary exposure to vancomycin.