Objectives: To describe the organization and distribution of intensive care unit (ICU) patients and services in the United States and to determine ICU physician staffing before the publication and dissemination of the Leapfrog Group ICU physician staffing recommendations.
Design and setting: Stratified, weighted survey of ICU directors in the United States, performed as part of the Committee on Manpower for the Pulmonary and Critical Care Societies (COMPACCS) study. Using lenient definitions, we defined an ICU as "high intensity" if > or =80% of patients were cared for by a critical care physician (intensivist) and defined an ICU as compliant with Leapfrog if it was both high-intensity and providing some form of in-house physician coverage during all hours.
Subjects: Three hundred ninety-three ICU directors.
Measurements and main results: We obtained a 33.5% response rate (393/1,173). We estimated there were 5,980 ICUs in the United States, caring for approximately 55,000 patients per day, with at least one ICU in all acute care hospitals. The predominant reasons for admission were respiratory insufficiency, postoperative care, and heart failure. Most ICUs were combined medical-surgical ICUs (n = 3,865; 65%), were located in nonteaching, community hospitals (n = 4,245; 71%), and were in hospitals of <300 beds (n = 3,710; 62%). One in four ICUs were high-intensity (n = 1,578; 26%), half had no intensivist coverage (n = 3,183; 53%), and the remainder had at least some intensivist presence (n = 1,219; 20%). High-intensity units were more common in larger hospitals (p = .001) and in teaching hospitals (p < .001) and more likely to be surgical (p < .001) or trauma ICUs (p < .001). Few ICUs had any in-house physician coverage outside weekday daylight hours (20% during weekend days, 12% during weeknights, and 10% during weekend nights). Only 4% (n = 255) of all adult ICUs in the United States appeared to meet the full Leapfrog standards (a high-intensity ICU staffing pattern plus dedicated attending coverage during daytime plus dedicated coverage by any physician during nighttime).
Conclusions: ICU services are widely distributed but heterogeneously organized in the United States. Although high-intensity ICUs have been associated previously with improved outcomes, they were infrequent in our study, especially in smaller hospitals, and virtually no ICU met the Leapfrog standards before their dissemination. These findings highlight the considerable challenge to any efforts designed to promote either 24-hr physician coverage or high-intensity model organization.