Processes of care, illness severity, and outcomes in the management of community-acquired pneumonia at academic hospitals

Background: Prompt antibiotic administration, oxygenation measurement, and blood cultures are generally considered markers of high-quality care in the inpatient management of community-acquired pneumonia (CAP). However, few s tudies have examined the relationship between prompt achievement of process -of-care markers and outcomes for patients with CAP. We examined whether an tibiotic administration within 8 hours of hospital arrival, a blood culture within 24 hours, an oxygenation measurement within 24 hours, or performing blood cultures before giving antibiotics was associated with the following : (1) reaching clinical stability within 48 hours of hospital admission, (2 ) a decreased length of hospital stay, or (3) fewer inpatient deaths. Methods: A retrospective medical record review identified 1062 eligible pat ients discharged from the hospital with a diagnosis of CAP between December 1, 1997, and February 28, 1998, among 38 US academic hospitals. We assesse d the independent relationship between each process marker and the 3 clinic al outcomes, controlling for the Pneumonia Severity Index on admission. We also examined the relationship of pneumonia severity on admission to proces s marker achievement and clinical outcomes. Results: Overall, there was no consistent or statistically significant rela tionship between achieving process markers and better clinical outcomes (P > .40 for all). We did observe that performing blood cultures within 24 hou rs was related to not achieving clinical stability within 48 hours (odds ra tio, 1.62; 95% confidence interval, 1.13-2.33). However, this finding likel y reflects residual confounding by severity of illness, since increasing pn eumonia seventy on admission was associated with blood culture performance (P=.009) and with shorter times to antibiotic administration (P=.04). Conclusions: Achieving process-of-care markers was not associated with impr oved outcomes, but was related to the severity of pneumonia as assessed on admission. Our results highlight the difficulty in demonstrating a link bet ween process-of-care markers and outcomes in observational studies of CAP. Randomized studies are needed to objectively evaluate the impact of process -of-care markers on CAP outcomes.