Aims: To identify variation in the management of -community-acquired pneumonia between two New Zealand hospitals and the factors that may account for any differences.
Methods: A 12-month, prospective two-centre study was conducted. Between July 1999 and July 2000, 474 adult patients with community-acquired pneumonia were enrolled: 304 in Christchurch Hospital and 170 in Waikato Hospital. The patients were similar in age, sex, prior antibiotic use and comorbidity. There was no significant difference in the clinical outcomes for the patients at the two centres.
Results: The mean duration of i.v. antibiotic therapy was 1.7 versus 3.0 days (P < 0.001) and length of stay (LOS) was 3.0 versus 5.9 days (P < 0.001) for Waikato and Christchurch Hospitals, respectively. Using multivariate analysis, we could account for 61% of the observed variation in LOS. Duration of i.v. antibiotic therapy independently accounted for 16% of variation in LOS compared with age (2%), chronic obstructive pulmonary disease, duration of fever, intensive care unit admission and centre of admission (all <1%). For the duration of i.v. antibiotics, centre of admission, largely reflecting clinician practice at each centre, independently accounted for 13% of variation, compared with duration of fever (5%), admission to the Intensive Care Unit (4%), Pneumonia Severity Index score (3%) and bacteraemia (3%).
Conclusion: Of the identifiable factors, variations in clinician behaviour outweighed the influence of patient factors on the duration of i.v. antibiotic therapy, which in turn was the major determinant of LOS for patients hospitalised with community-acquired pneumonia. An early switch from i.v. to oral antibiotic therapy in conjunction with early discharge planning may significantly reduce LOS without compromising patient outcomes.