Direct costs in patients hospitalised with community-acquired pneumonia after non-response to outpatient treatment with macrolide antibacterials in the US.

Introduction: Antibacterial cost-containment programmes emphasise the use of narrow-spectrum generic agents whenever possible. The use of these agents is driven by their lower purchase prices; the consequences of treatment failure are rarely considered. This study was conducted to identify the costs of treating patients hospitalised with community-acquired pneumonia (CAP) associated with Streptococcus pneumoniae following failure to respond to outpatient treatment with macrolide antibacterials.

Methods: A multicentre, retrospective, observational study was performed in patients with CAP due to S. pneumoniae who were admitted to 31 North American hospitals following a lack of response to >or=2 days of outpatient treatment with a macrolide antibacterial. Direct medical costs (year 2004 values) of infection-related hospital resources, including antibacterials (purchase, preparation, dispensing, administration and monitoring), diagnostic tests, therapeutic procedures, treatment of adverse events and therapeutic failures, and hospitalisation per diem (ward, critical care and ventilator days), were analysed. Total hospital costs were then compared with standard diagnosis-related group (DRG) reimbursement.

Results: A total of 122 patients were enrolled. Patients were frequently bacteraemic (52%) and infected with macrolide-resistant strains of S. pneumoniae (71%). Initial inpatient antibacterial treatment was not successful in 17 patients (14%) and seven patients (5.7%) died. The mean length of stay was 8.7 days (SD 7) including 1.3 days (SD 2.9) in a critical care unit and 1.4 days (SD 4.4) of mechanical ventilation. The mean cost of hospitalisation was US dollars 12,678 (SD 13 346) but standard DRG reimbursement averaged only US dollars 8,634.

Conclusions: Patients who do not respond to outpatient treatment with a macrolide antibacterial and who are subsequently hospitalised with CAP caused by S. pneumoniae are likely to be infected with a non-susceptible strain, are frequently bacteraemic, are at an increased risk for mortality compared with previously published estimates in patients with CAP due to S. pneumoniae, and incur hospital costs that far exceed standard DRG reimbursement for CAP.