Emergency department treatment of community-acquired pneumonia decreases hospital length of stay

Abstract

Study objectives: More than 4 million adults in the United States contract community-acquired pneumonia (CAP) each year. Approximately 1 million are hospitalized. Many reports have documented the success of efforts to decrease time receiving intravenous antibiotics, length of stay, and overall costs for pneumonia patients without adversely affecting outcomes. Early administration of antibiotics is cited as an important predictor of improved outcomes. We evaluate the difference in the inpatient length of stay and costs of emergency department (ED) admissions compared to elective admissions for CAP. Methods: This was a statewide retrospective case-control study for California using patient discharge data for 2002 from the Office of Statewide Health Planning and Development (OSHPD). Patient admissions for pneumonia (diagnostic-related group 8, 9) receiving intravenous antibiotics (Current Procedural Terminology procedure code 9921) were included. Interventions were compared beginning on inpatient admission day comparing ED and elective. The subsequent length of stay for the ED admissions were compared with the elective admits. We excluded from our sample patients with length of stay greater than 60 days, as potential outliers. Because cost of care was an important variable in our regression model to control for case severity, we excluded those cases in which cost data were unavailable. Multiple regression was used to test for statistical significance of the identified difference in the average length of stay between ED and elective admissions, controlling for the type of admission (ED versus elective) and patient demographics (age, sex, race, ethnicity, payer status). We used estimated cost of care to control for case severity and derived estimated cost of care by multiplying patient level cost-to-charge ratio to patient charges. To control hospital practice patterns that may influence length of stay, we incorporated OSHPD hospital IDs into the model to absorb the hospital effect in the model. We used natural logarithms to control for left-hand-side tail to the length of stay data. Results: On subjecting the sample to multiple regression, we found that the observed differences in the inpatient length of stay (ED versus elective) were statistically significant (P<.0001) for all the categories in the Table. No major statistically significant difference was observed among demographic subsets of patient categories. Table, abstract 19Inpatient length of stay (days) for pneumonia ED admissions compared with elective admission. No. Mean SD ED admission Pneumonia patient >receiving IV antibiotic 271 4.1 3.4 Elective admission Pneumonia patient receiving IV antibiotic 78 6.3 5.0 IV, Intravenous. Open table in a new tab Study objectives: More than 4 million adults in the United States contract community-acquired pneumonia (CAP) each year. Approximately 1 million are hospitalized. Many reports have documented the success of efforts to decrease time receiving intravenous antibiotics, length of stay, and overall costs for pneumonia patients without adversely affecting outcomes. Early administration of antibiotics is cited as an important predictor of improved outcomes. We evaluate the difference in the inpatient length of stay and costs of emergency department (ED) admissions compared to elective admissions for CAP. Methods: This was a statewide retrospective case-control study for California using patient discharge data for 2002 from the Office of Statewide Health Planning and Development (OSHPD). Patient admissions for pneumonia (diagnostic-related group 8, 9) receiving intravenous antibiotics (Current Procedural Terminology procedure code 9921) were included. Interventions were compared beginning on inpatient admission day comparing ED and elective. The subsequent length of stay for the ED admissions were compared with the elective admits. We excluded from our sample patients with length of stay greater than 60 days, as potential outliers. Because cost of care was an important variable in our regression model to control for case severity, we excluded those cases in which cost data were unavailable. Multiple regression was used to test for statistical significance of the identified difference in the average length of stay between ED and elective admissions, controlling for the type of admission (ED versus elective) and patient demographics (age, sex, race, ethnicity, payer status). We used estimated cost of care to control for case severity and derived estimated cost of care by multiplying patient level cost-to-charge ratio to patient charges. To control hospital practice patterns that may influence length of stay, we incorporated OSHPD hospital IDs into the model to absorb the hospital effect in the model. We used natural logarithms to control for left-hand-side tail to the length of stay data. Results: On subjecting the sample to multiple regression, we found that the observed differences in the inpatient length of stay (ED versus elective) were statistically significant (P<.0001) for all the categories in the Table. No major statistically significant difference was observed among demographic subsets of patient categories. IV, Intravenous.