Hotline system provides year-round phone consultation service for inquiries from each hospital department. The most common incoming call is, as expected,"We have a patient with a fever!"However, at present there are many physicians who say"We have not done a blood culture,"and"As a stopgap , we started the patient on an antibiotic as of yesterday."In fact, reaching a diagnosis of a nosocomial fever without having performed the first work-up for a fever (i.e., two sets of blood culture, chest X-rays taken in two directions, and general urinalysis and smear culture) is like reaching a diagnosis of myocardial infarction without performing an ECG. The root of these problems is a lack of awareness of the importance of the initial response to febrile patients. Ill-considered administration of an antibiotic can result in delayed (perhaps indefinitely) diagnosis of a patient with infectious endocarditis. In fact, since the causative organism of infectious endocarditis is unknown, in most cases there has been earlier administration of an antibiotic. We carried out a collaborative retrospective clinical study at 17 hospitals affiliated with the Japanese Society of Hospital General Medicine to investigate the diagnostic methods used for causative diseases of fevers of unknown origin (FUO). We found that, even for patients with classical FUO, blood cultures were not performed in 13.2% of the patients with FUO. To exclude bacteremia, it is necessary to perform two sets of blood culture using blood drawn when the patient has not been administered an antibiotic for at least 48 h. When a catheter-related blood infection is suspected, simultaneous culture of the catheter tip and blood should be submitted. In our hospital, the rate of collection of two sets of blood culture and the rate of submission of only catheter culture are reported for each department. The problems with blood inflammatory markers (e.g., WBCs, CRP, ESR, etc.) are that they do not identify the site of inflammation and the assay data do not always match the severity. For inflammation at sites remote from blood, such as meningitis, abscess, etc., the CRP does not always increase. The blood procalcitonin value has been reported to have high specificity for bacterial infections, but evaluations of its clinical usefulness have not shown consistent results. Especially in patients with classical FUO, procalcitonin values were not found to correlate with the causative diseases. Inflammatory markers cannot be relied upon to exclude bacterial infections.