Consensus definitions have emerged for the discrimination between infected and uninfected prosthetic joints but diagnostic uncertainty often occurs. We examined the accuracy of orthopaedic surgeons' assessments to diagnose the infected prosthetic hip or knee and elucidated the added value of laboratory parameters. A prospective cohort study of patients undergoing revision arthroplasty of hip or knee was conducted over a one-year period. Orthopaedic surgeons' determinations prior to arthroplasty were recorded. A reference diagnostic standard was determined retrospectively by independent review from 3 infectious diseases physicians. Patients were followed up to 12 months. For 198 patients enrolled, 228 surgical encounters (110 knee, 118 hip) were classified by independent reviewers as 176 uninfected and 52 infected. Orthopaedic surgeons' preoperative diagnoses of infection had high diagnostic accuracy (sensitivity 89%, specificity 99%, PPV 98%, NPV 97%). Addition of intraoperative findings and histopathology improved their diagnostic accuracy. Addition of culture and PCR results improved sensitivity of diagnostic determinations but not specificity. We provide evidence that clinical acumen has high diagnostic accuracy using routine preoperative parameters. Histopathology from intraoperative specimens would improve surgeons' diagnostic accuracy but culture and PCR from intraoperative specimens could create greater diagnostic uncertainty. This study is critical to further our understanding of the added value, if any, of laboratory testing to support clinical decision making for the suspected infected joint and allow us to identify diagnostic gaps for emerging technologies to fill that will improve our ability to diagnose the infected prosthetic joint.