Objectives: To prospectively evaluate a clinical algorithm for the diagnosis of pneumonitis and pneumonia in nursing home residents.
Design: Prospective cohort study.
Setting: Inpatient geriatrics unit.
Participants: Nursing home residents admitted to the hospital with suspected pneumonia.
Measurements: Identification of pneumonitis and pneumonia using the algorithm; medical record review and abstraction of clinical data; hospital outcome and length of stay.
Results: One hundred seventy episodes of suspected pneumonia were screened with the algorithm and classified into four groups: 25% pneumonia, 28% aspiration pneumonitis of 24 hours or less duration, 12% aspiration pneumonitis of more than 24 hours' duration, and 35% an aspiration event without pneumonitis. Presenting symptoms and signs, laboratory tests, severity of illness measures, or serum C-reactive protein levels did not distinguish between the four groups. Those with an aspiration event without pneumonitis tended to be treated less often with antibiotic therapy after admission (P=.004) and after discharge (P=.01). Of those who survived, there was no significant difference in mean hospital length of stay between the four groups. There was no significant difference in the percentage of case fatality between the four groups, but those with aspiration pneumonitis of 24 hours or less duration and with an aspiration event without pneumonitis had a lower mortality than the other two groups.
Conclusion: Distribution of episodes of suspected pneumonia by clinical category as determined using the algorithm was similar to that of the derivation study, as were case fatality rates in each category. These findings suggest that the algorithm may be useful for making the distinction between pneumonitis and pneumonia in nursing home residents; further studies are warranted.