Comparison of processes and outcomes of pneumonia care between hospitalists and community-based primary care physicians.

Objective: To compare medical care provided by hospitalists and primary care physicians to patients with community-acquired pneumonia in order to identify specific practices that might explain the improved efficiency of care provided by hospitalists.

Patients And Methods: We retrospectively reviewed the medical charts of 455 patients hospitalized with pneumonia at a community-based tertiary care center between January 1, 1998, and January 1, 1999. Exclusion criteria included human immunodeficiency virus infection, lung cancer, active tuberculosis, hospitalization within 7 days, length of stay (LOS) more than 14 days, and requirement of mechanical ventilation. All patients were cared for by either a full-time hospitalist or a primary care physician. Data collected included patient insurance status, variables to calculate each patient's Pneumonia Severity Index score, initial antibiotic selection, door-to-needle time, time to patient stability for switch to oral antibiotics, time to actual switch, unstable variables at discharge, and subspecialty consultation rate. Each patient's initial chest x-ray film was reviewed and classified as diagnostic of pneumonia, indeterminate, or clear. Outcomes measured via administrative database were mortality, LOS, costs, and readmission rate.

Results: Primary care physicians cared for 270 patients, and hospitalists cared for 185. Primary care physician patients were older, and this group had a higher proportion of the highest-risk patients. The mean time to stability was 3.2 days for hospitalists and 3.3 days for primary care physicians, and the mean time from stability to actual switch from intravenous to oral antibiotics was 1.6 days and 23 days, respectively (P=.003). The mean adjusted LOS was 5.6 days for hospitalists and 6.5 days for primary care physicians. Similarly adjusted costs were $594 less per patient treated by hospitalists. A difference in door-to-needle time of 0.9 hour favoring primary care physicians did not contribute to LOS. No significant differences were noted in adjusted inpatient mortality or the appropriateness of initial antibiotics used. Primary care physicians were more likely to prescribe clindamycin and ceftazidime, and they requested infectious disease consultations more often. At discharge, 14% of hospitalist patients and 7% of primary care physician patients had at least 1 unstable variable. Differences in hospital readmission rates at 15 and 30 days were not statistically significant in combined or risk-stratified analyses.

Conclusions: Inpatients with community-acquired pneumonia cared for by hospitalists had a shorter adjusted LOS than those seen by primary care physicians primarily because of earlier recognition of stability and more rapid conversion from intravenous to oral antibiotics. Adjusted costs were likewise reduced. However, patients seen by hospitalists were discharged with an unstable clinical variable more often. Other than earlier switch to oral antibiotics, less use of clindamycin and ceftazidime, and fewer infectious disease consultations, hospitalists' processes of care were similar to those of primary care physicians.