Objective: To gather data about available technology, staffing, administrative policies, and bed capacities of ICUs in the United States.
Design and setting: On January 15, 1991, survey instruments were mailed to the administrators of 4,233 hospitals to gather information from the medical director of the institutions' respective ICUs for the purpose of developing a database on ICUs in the United States. The sampling frame for this study was based on all American Hospital Association (AHA) hospitals that stated they have ICUs.
Measurements: Census questionnaires solicited information on types of hospitals, types of ICUs, number of ICU beds open and closed, technology available to the unit, organizational structure and management of the ICU, as well as the staffing and certification of unit personnel.
Main results: Data were obtained on 32,850 ICU beds with 25,871 patients from 2,876 separate ICUs in 1,706 hospitals in the United States. Census responses came from units in all sizes of hospitals within all ten census regions in the country, all states, and all types of hospital sponsorship (federal, state, and local government, private nonprofit and private for profit). The census response rate was 40% of the AHA hospitals that stated that they have ICUs, with specific ICU data on 38.7% of the nation's ICUs. The number of ICUs per hospital increases with overall hospital size. The smallest hospitals (less than 100 beds) usually had only one ICU. As hospital size increased, the single, all inclusive medical/surgical/coronary care units diminished, and in hospitals with greater than 300 beds, specialization of units became prevalent. In absolute terms, hospitals had the following number of ICUs: 1.04 +/- 0.20 (less than or equal to 100 beds); 1.30 +/- 0.65 (101 to 300 beds); 2.37 +/- 1.58 (301 to 500 beds); and 3.34 +/- 2.21 (greater than 500 beds). ICU beds averaged, nationally, 8.09% of hospital-licensed beds with a median of 6.98%. Generally, medical units, pediatric units, coronary care units (CCUs), and medical/surgical/CCUs reported an average of 10 beds per unit. Neonatal units averaged 21 beds, and surgical units averaged 12 beds. The average ICU size, nationally, was 11.7 +/- 7.8 beds per unit. Available technology within hospitals and individual units was increased as hospital size increased; surgical units tended to have more available technology than other unit types. A wide range of organizational arrangements within hospitals determines where the ICU appears in an organizational chart and to whom unit management is accountable. Thirty-six percent of the units were located organizationally within the hospital's department of medicine, while 23% were considered "free standing," having no departmental affiliation. Although units must have a medical director, the perception as to whether this director supervises the day-to-day operation was different in larger vs. smaller hospitals. In hospitals with less than or equal to 100 beds, 72% of the units were perceived to be supervised by the medical director, whereas in larger hospitals (greater than 500 beds), 81% of units were supervised. Study results indicated that medical directors in pediatric, neonatal, and burn units most often were perceived to supervise the unit. Presently, 63% of all ICUs responding are directed by an internist. The next largest group to direct ICUs were surgeons, followed by pediatricians. Pediatrician involvement tended to be exclusive in pediatric and neonatal units. Surgeons directed most surgical and neurologic units and were involved in 21% of mixed medical/surgical units. Internists predominated in medical units and in CCUs, as well as in combined medical/surgical/CCUs. Direction by anesthesiologists, although relatively infrequent, predominated in the surgical unit. Critical care medicine certification of the medical director and attending staff of the ICU increased as hospital size increased, although only 44% of all units stated that thei