The impact of attendance by infectious disease specialists (IDS) on hospitalised adults with community-acquired infection was assessed by studying 402 consecutive febrile adults who were admitted randomly to either of two internal medicine wards over a 4-month period and given intravenous antibiotics. In ward 1, patients were attended by IDS, whereas those in ward 2 were attended by physicians from other specialties. In total, 160 patients were treated in ward 1 and 242 in ward 2 (median age 66 years; 49% male). The case-mix was comparable. Only 39% of ward 2 patients underwent minimal fever diagnostic tests compared to 82% in ward 1 (p < 0.001). Ward 1 and 2 patients received 188 and 315 antibiotic courses, respectively, of which 32% and 20% required approval from IDS (p 0.003). Patients in ward 1 were more likely to receive ceftriaxone (7.5% vs. 2%; p 0.002), erythromycin (7% vs. 1.5%; p 0.002) and cefuroxime (48% vs. 26%; p < 0.0001), but were less likely to receive amoxycillin-clavulanate (8% vs. 28%; p < 0.0001). The mean durations of therapy were 3.6 and 3.2 days (not significant), and therapy was deemed to be completely appropriate in 55.5% and 43% of cases, respectively (p 0.012). The crude mortality rates were 6.3% and 7.9%, respectively (not significant), while the medication costs were US dollars 27.4 and US dollars 26.4/patient/antibiotic day, respectively. Regular attendance by IDS resulted in significantly higher rates of accurate diagnosis and appropriate therapy. IDS prescribed more restricted (and expensive) agents, but preferred less expensive agents among unrestricted drugs, thereby offsetting the overall medication costs.