A PROSPECTIVE, MULTICENTER STUDY OF A PNEUMONIA PRACTICE GUIDELINE

Objective: To study the effect of a length of stay practice guideline on patient outcomes. Design: A prospective, nonrandomized, interventio nal trial, Setting: Six geographically distributed hospitals. Patients : Two hundred forty-two consecutively hospitalized ''low-risk'' patien ts with pneumonia, Measurements and results: One hundred fifty-two pat ients (63%) completed the mailed postdischarge survey and were include d in the analysis. Data were prospectively collected for 85 patients f rom the baseline observation period (B) and 67 patients from the inter vention period (I), During the I, case managers provided physicians wi th patient risk information based on guideline recommendations. There was no significant change in guideline compliance (B vs I: 76.5% vs 83 .6%; p=0.32) or length of stay (B vs I: 3.5 days [95% confidence inter val, 3.2 to 3.8] vs 3.6 days [95% confidence interval, 3.3 to 4,0]), A lso, there were no statistically significant effects of the interventi on on patient outcomes, care following hospital discharge, and patient satisfaction scores, Conclusion: Patients in this study often had sho rter lengths of stay than recommended by the practice guideline. This suggests that the external environment may have had a greater effect o n physician behavior and length of stay than the practice guideline it self, Moreover, it demonstrates the importance of continuous assessmen t of physician practices immediately prior to, during, and after appli cation of the clinical practice guideline.