Abstract
BACKGROUND: Utilization of diagnostics and biomarkers are the second largest cost drivers in the management of patients hospitalized with community-acquired pneumonia (CAP). The present study aimed to systematically assess the inter-hospital variation in these cost drivers in relation to antibiotic use in CAP. METHODS: Detailed resource utilization data from 300 patients who participated in a multicenter placebo-controlled trial investigating dexamethasone as adjunctive treatment for community-acquired pneumonia was grouped into 3 categories: clinical chemistry testing, radiological exams, and microbiological testing. Based on the identified top 5 items per category, average costs were calculated per category and per hospital. Antibiotic de-escalation at day 3 and secondary ICU admission were assessed as outcomes for proportionality of diagnostics use. RESULTS: The mean costs for diagnostics varied between hospitals from 350 (SD 31) to 841 (SD 37) euro per patient (p < 0.001). This difference was primarily explained by variation in costs for microbiological testing (mean 195 vs. 726 euro per patient, p < 0.001). There was no difference in number of secondary ICU admissions but there was an inverse association between the costs of microbiological testing and level of antibiotic de-escalation. De-escalation occurred most frequently in the hospital with the lowest cost for microbiological testing (48% vs. 30%; p = 0.018). The latter hospital had an automated physician alert system in place to consider a timely iv-to-oral switch of antibiotics. CONCLUSIONS: Large inter-hospital variation exists in resource utilization, mainly in microbiological diagnostics in the management of adult patients with community-acquired pneumonia. A counterintuitive inverse association between the magnitude of these costs and the amount of antibiotic de-escalation was found. Future studies about the optimal cost-effective set of microbiological testing for antimicrobial stewardship in pneumonia patients should acknowledge the interaction between testing, way of communication of results and triggered physician alert systems. TRIAL REGISTRATION: ClinicalTrials.gov NCT01743755.