Triage boosts outpatient treatment of CAP

A triage system for grading the severity of community-acquired pneumonia (CAP) helped hospitals treat more patients as outpatients without compromising their medical outcomes. A moderately intensive program for implementing the triage system was the most efficient way to get hospital staffs to use the system in the study with 32 hospitals and 1,874 patients, as reported at the 98th International conference of the American Thoracic Society. The strategy of moderate intensity for implementing a pneumonia-triage system is similar to what most state Peer Review Organizations (PROs) already do. Our findings endorsed what is currently done added by a pulmonologist at Norwalk (Conn.) Hospital. The study compared three strategies for getting the staffs of hospital emergency departments to apply the Pneumonia Severity Index (PSI) when triaging patients. The PSI is a five-point scale for assessing the severity of disease in patients with community-acquired pneumonia. (Results from prior studies have shown that patients with the lowest severity of disease on the PSI, risk classes I, II or III, who are also not hypoxemic, can be safely managed as outpatients. The current study was designed to compare three different strategies for getting the medical staffs of emergency departments to use the PSI to identify patients with CAP who could be treated as outpatients.

The study included 32 emergency departments at hospitals in Connecticut and Pennsylvania. Eight of the hospitals were randomized to use a low-intensity strategy to implement the PSI-based triage scheme. The staffs received a guideline that said that low-risk, non-hypoxemic patients could be treated as outpatients. Twelve hospitals used a moderate-intensity strategy. In addition to receiving this guideline, the state PRO gave the emergency department director data on how many pneumonia patients in the past fell into the low-risk category. Each director also had to devise a quality improvement plan, and emergency-department physicians went to an educational session on implementing the guidelines. The other 12 hospitals received a high-intensity intervention. In addition to the moderate-intensity program, they also received a paper algorithm to calculate the PSI score, and each physician on the emergency department staff received patient-specific audit information.

These emergency departments also used a plan-do-study-act cycle every two months to help improve guideline adherence. During January 15, 2001 through December 31, 2001, 3,450 patients appeared at these hospitals with CAP. Of these, 1,874 were diagnosed as low-risk patients with a PSI of I-III. The fraction of these patients who were treated as outpatients was 38.2 percent at the low-intensity intervention hospitals, 62.4 percent at the moderate-intensity hospitals, and 62 percent at the high-intensity hospitals. When the data were adjusted for PSI risk class, the odds ratio for a patient being treated as an outpatient was one for the low-intensity intervention hospitals, three for the moderate-intensity hospitals, and 2.4 for the high-intensity hospitals.

The death rate among patients was similar in all three groups, ranging from 2.1 percent at the low-intensity-intervention hospitals to 1.7 percent at the high-intensity hospitals. The need for re-hospitalization was also similar among the hospitals in all three intervention groups, ranging from five percent in the moderate-intensity group to 6.4 in the high-intensity group.