WESTPORT, CT (Reuters Health) – A new model more adequately predicts 30-day, all-cause mortality following lower respiratory tract infection among nursing home residents than does the Pneumonia Severity Index (PSI), according to results of a prospective study conducted in Missouri. The model's developers believe that it will help clinicians assess the need for patient hospitalization.

The model is based on 1406 episodes of lower respiratory tract infection among 1044 residents of 36 nursing homes in Missouri. Subjects were at least 60 years old. As reported in the November 21st issue of The Journal of the American Medical Association, 207 patients died within 30 days.

Dr. David R. Mehr of the University of Missouri in Columbia and associates selected 25 categories of variables that could be related to mortality. They then used logistic analysis involving 975 randomly selected patients as a "development sample" to devise an eight-variable model to predict 30-day mortality.

Of the eight variables the most significant were serum urea nitrogen, white blood cell count, body mass index, pulse rate, and scores for activities of daily living. Also included in the model were total lymphocyte counts, male sex, and decline in mood over the previous 90 days.

For patients in the development sample, the ratio of mortality in the highest-risk and lowest-risk quintiles as predicted by the model was 17.2. In a validation set of 431 patients, the ratio of mortality in the highest and lowest quintiles was 13.8.

The model developed by Dr. Mehr and his colleagues identified 52% of the residents as having a low or relatively low risk of 30-day mortality. The PSI, on the other hand, would have classified 85% of the subjects in the highest risk categories. According to the authors, the PSI overstates risk in nursing home residents because it adds points for each year of age and for nursing home residence.

"If someone is at low risk, he or she in most circumstances is probably a very good candidate to treat in the nursing home with a simple oral antibiotic regimen," Dr. Mehr told Reuters Health.

The decision-making becomes more complex at the high-risk end, he added. In this study, for the 65 people in the highest risk category, mortality was 60%. Dr. Mehr said that it may be time to ask, "If these people are close to the end of their lives, do we want to subject them to care in an intensive care unit and putting them on a ventilator?"

He noted that his team is collaborating with investigators in The Netherlands to evaluate 3- or 6-month mortality among patients with pneumonia and identify predictors of longer-term mortality.

"For example, people with severe dementia who get pneumonia have a very high 6-month or 1-year mortality," Dr. Mehr added. "If such a patient survives one episode of pneumonia, perhaps it is time to think about whether treatment with antibiotics would be appropriate in the next episode."

If independently validated, "our prediction rule could aid clinicians and researchers in optimizing care for nursing home residents with lower respiratory tract infections," the investigators conclude.

JAMA 2001;286:2427-2436.