Objective: Clinical practice guidelines recommend standardized diagnostic microbiological testing for community-acquired pneumonia on hospital admission, although evidence of its impact on quality is limited. This study evaluated the relationship between guideline-concordant microbiological testing (GCMT) and both in-hospital mortality and length of stay.
Design: /st> Retrospective cohort study using a multicenter claims-based inpatient database linked to a government hospital census database in Japan.
Setting and participants: /st> Patients who were diagnosed with and treated for pneumonia, and were discharged between 1 July 2010 and 30 September 2011 (n = 65 145).
Main outcome measures: /st> GCMT was defined to include sputum tests, blood cultures and urine antigen tests conducted on the first day of hospitalization. We examined the association between 30-day in-hospital mortality and both the performance of each test and the number of tests performed using multivariable logistic regression analysis, adjusting for patient demographics, pneumonia severity and hospital characteristics. Length of stay was analyzed using a Cox proportional hazards model.
Results: /st> Simultaneous conduct of all three tests was significantly associated with reduced 30-day mortality (odds ratio: 0.64; 95% confidence interval (CI): 0.56-0.74) and with increased likelihood of discharge (hazard ratio: 1.04; 95% CI: 1.00-1.07), after adjusting for patient and hospital characteristics. The association was more marked as the level of disease severity increased.
Conclusions: /st> Performance of GCMT was significantly associated with lower mortality and shorter length of stay. These results suggest that hospitals should assure performance of GCMT in patients with severe community-acquired pneumonia.
Keywords: community-acquired infections; diagnostic testing; guidelines; pneumonia.