Performance measurement in pneumonia care: beyond report cards

To compare the medical management of bacteremic pneumococcal pneumonia at a university-based and a community-based teaching hospital and evaluate strategies for performance measurement and subsequent improvement.

We conducted a retrospective cohort study involving a 450-bed university hospital in the inner city and a 400-bed private hospital in a rural community.

The medical records of all adults with bacteremic pneumococcal pneumonia admitted to a university and a community hospital during a 5-year period were reviewed. Information about patient age, sex, underlying medical condition, severity of disease, health-care insurance, management, and outcome was collected and analyzed.

Patients at the two hospitals were similar in underlying illnesses and severity of disease. In comparison with the community hospital, resource expenditure was greater at the university hospital, where all 11 identified diagnostic measures and treatment resources were used more often. This difference was statistically significant for sputum cultures, all cultures, and lumbar punctures. Despite the greater intensity of care, in-hospital mortality was higher at the university hospital (26%) than at the community hospital (12%) (P>0.1).

The outcome of bacteremic pneumococcal pneumonia did not differ significantly at a university hospital in comparison with a community teaching hospital, even though resource expenditure at the university hospital was greater. Our findings suggest that hospital "report cards" based solely on outcome comparisons provide inadequate information. In contrast, examination of variations in profiles of resource utilization can detect important differences in hospitals and can be used to guide continuous quality improvement efforts and ultimately improve hospital care.