Background: It has become an article of faith that appropriate antibiotic therapy is needed for best outcomes during a serious infection. Despite this long-held view, there is some debate about the role of appropriate outcome in serious infections, in particular with nosocomial pneumonia. Therefore, more recent data on adequacy of antibiotic therapy and outcomes were reviewed. Methods: The medical literature from 1997 to 2004 was surveyed for articles that directly dealt with appropriate therapy. Search terms included ‘appropriate and inappropriate antibiotic therapy’, ‘adequate antibiotic therapy’, ‘resistance and antibiotic failures’ and ‘delayed therapy’. The data were abstracted to obtain their essential findings. Results: In bacteremia, data are most persuasive that appropriate and timely therapy significantly influences outcomes. Areas where this may not be the case are studies where coagulase-negative staphylococci are isolated in large numbers or in studies where the incidence of appropriate therapy is high. One area where data are not conclusive concerns the treatment of enteric bacteria carrying extended spectrum betalactamases, where the only cephalosporin of concern is ceftazidime. There is not enough data to compare carbapenems with specific cephalosporins to conclude that these are the most appropriate agents. The studies in regard to nosocomial pneumonias are not as conclusive as those with bacteremias. There appears to be a subset of patients that do not respond to therapy or do not survive, which confounds studies of this population; however, most studies favor a role of appropriate therapy. Conclusions: Appropriate antibiotic therapy has several dimensions. It improves outcomes in most serious diseases. Timing of administration and appropriateness, based on susceptibility, are the most important determinants, but dosing intervals and dose probably play similarly important roles in outcomes that have not been examined exhaustively in humans. Other aspects of appropriate therapy that deserve attention include a shift to more ‘resistance’-proof antibiotics in empiric therapy, which may be accompanied by better outcomes.