Special Feature

The Structure of Intensive Care: If You’ve Seen One ICU, You’ve Seen One ICU

By Gordon D. Rubenfeld, MD, MSc

Measuring the quality of a complex service such as critical care that combines the highest technology with the most intimate caring is a challenge. Donabedian proposed a framework for thinking about the quality of medical care that separates quality into three components: structure, process, and outcome.1 There is a debate about which of these measurements is the most important measure of quality. A compelling argument can be made for considering patient outcome as the only important measure of medical care quality. Several attempts have been made at judging the quality of intensive care based on an important outcome—severity-adjusted mortality.2 Nevertheless, structure is an essential aspect of critical care and it, along with process, are the means by which we effect change in outcome.

Studies of ICU structure are complicated by the variable nature and elusive definition of critical care. One must decide whether one is studying the structure of the ICU as a geographic location (as all existing studies have done) or of critical care as a service to a particular patient population. Are patients on prolonged ventilatory support in a nursing home in an ICU? Are patients being weaned in a post-anesthesia care area after cardiac bypass surgery receiving intensive care? Is the emergency department an "intensive care unit" when critically ill patients are cared for there for various periods? The answers to these questions will determine which special care areas and which processes of care constitute critical care.

The Society of Critical Care Medicine (SCCM) has published a number of guidelines on ICU structure.3,4 The American College of Surgery’s criteria for trauma center designation contain explicit recommendations for the organization of critical services.5 This detailed document includes specific structural criteria for human and material resources as well as their organization.

With one or two important exceptions, there is little evidence linking ICU structure to specific outcomes and therefore most recommendations are not evidence based. An understanding of the complex domains of ICU structure is facilitated by breaking it up into components (see Table 1).

Table 1-Domains of ICU Structure
Material Resources
   • Physical layout
   • Technology
Human Resources
   • Physicians
   • Nurses
   • Nonphysician clinicians
Organizational Structure
   • Admission policy—closed vs. open
   • Governance style
   • Specialty units and teams

Material Resources

Physical Layout. There have been two sets of recommendations by the Society of Critical Care Medicine for the layout and resources necessary for an ICU.3,4 Many of the recommendations come from common sense and basic concepts of human performance engineering. Of equal importance to the effect of the ICU on providers is the effect of the room, its layout, and design on the patient and visitors. In fact, ICU designers must consider a variety of users who spend time in the ICU. Therefore, it is essential that design of new units or modification of existing units be a collaborative process.

Technology. Perhaps nothing defines the modern ICU more than its armamentarium of sophisticated technology for monitoring, supporting, and changing human physiology, as well as organizing and displaying the collected data. The SCCM guidelines recommend, at a minimum, "the analysis and display of one or more electrocardiographic leads, at least two fluid pressures, and direct or indirect measures of arterial oxygen levels."6

Technology has a profound effect on the organization and process of care in the ICU. Theoretically, sophisticated "smart" alarms would allow one nurse to care for more patients. Unfortunately, while it seems intuitive that trading technology for human labor will result in lower costs, there is little evidence that substituting technology for nurses can result in significant cost savings. The effect of advanced technology on patient outcome is itself a controversial area.7,8 Some studies suggest that having advanced technology available leads to better outcome, but other studies do not.9 Curiously, and perhaps not unexpectedly, one study found that higher levels of technologic availability were associated with a decreased ability to meet family needs, raising the possibility that a focus on technology may lead ICU providers to neglect the human needs of their patients.9

Human Resources

In addition to physicians and nurses, a large number of nonphysician clinicians, allied health personnel, and support staff contribute to the care of critically ill patients. Workforce issues related to critical care are extremely important.10 Recent health care reforms in the United States have had a profound effect on the deployment of human resources throughout the health care delivery system.10

Physicians. Although there are a great deal of data on the effect of the organization of the ICU on outcome, there are relatively few data on the effect of subspecialty physician training on knowledge, process of care, or outcome in the ICU. Studies evaluating patient outcome and physicians’ knowledge in cardiology and AIDS care show benefit of specialty training and the same might be expected in critical care.11,12

Nurses. In many ways, intensive "care" is intensive "nursing care." The education and allocation of nursing staff in an ICU is an important job for ICU leaders. There are relatively few data available to guide clinicians to the optimal nurse:patient ratio for critically ill patients. In one study, the nurse:patient ratio was not associated with outcome or efficiency of ICU care.9 It may be that the actual nurse:patient ratio has less of an effect on care than how the patient’s needs are assessed, how tasks are assigned, and that this process be collaborative.

Other Nonphysician Clinicians. A number of nonphysician clinicians participate in the care of critically ill patients. These include physicians’ assistants, nurse practitioners, advanced practice nurses, respiratory therapists, and pharmacists. These providers have demonstrated efficacy in a number of practice areas outside of the ICU. There is some experience using nonphysician clinicians to replace housestaff in intensive care.13,14 Experience within the adult ICU is limited and a recent ICU survey suggests that less than 10% of adult ICUs use physician assistants or nurse-practitioners.15 Finally, although virtually every ICU in the United States employs respiratory therapists, relatively little is known about the organization and delivery of respiratory care in ICUs.16

The role of pharmacists in critical care deserves special mention. A recent report on the effect of medical errors on patient outcomes identified adverse drug events as a major component of preventable hospital mortality and morbidity.17 There is a compelling body of evidence that routine use of pharmacists in the ICU and computerization of drug ordering can reduce these negative outcomes.18

Organizational Structure

Organizational structure is a less tangible concept than either human or material resources. It describes how the individual pieces of the ICU interact to form a whole. Business schools recognize the importance of organizational structure, and entire courses and texts are devoted to this topic. Two texts focus on organizational management issues in critical care.19,20 For the purposes of this review, organizational structure in the ICU will be considered in three general categories: admission policies, governance style, and specialization.

Admission Policy. While most ICUs have a physician medical director, the responsibility of this person for the actual care of patients varies considerably in different ICUs.21 Although there is no standard terminology, admission policies to ICUs are described as "open" (physicians admit their own patients to the ICU and direct their care), "closed" (admission to the ICU is restricted, and physicians must transfer primary care of their patients on admission to a designated ICU team that is responsible for order writing), or "semi-closed" (patients receive mandatory consultation from a select group of physicians, but primary care of the patient is not transferred). Obviously, these are points along a continuum and a variety of intermediate models exist. In some hospitals, different units will be run under different admitting policies.21

Because there is no standard set of terms to describe these policies in ICUs, it is difficult to compare results from different studies and therefore difficult to understand what the admitting policies are in various ICUs around the world; however, considerable variation in ICU size and admitting policies has been observed (see Table 2).

Table 2-Structure of U.S. and Western European ICUs
U.S. 199221 U.S. 199815 Western Europe 199227
  University or
  university affiliated 31% 43% 49%
  Community 69% 57% 51%
Size Average 11.7 beds 48% > 10 beds 25% > 10 beds
40% 6-10 beds 57% 6-10 beds
12% < 6 beds 18% < 6 beds
Mixed Medical Surgical 45% 66% 74.4%
Closed unit 22% 29% NR
24-hr critical care 5% NR 71.7%
specialist available

Larger hospitals with larger ICUs, academic hospitals, medical (as opposed to surgical) units, and ICUs with predominantly managed care patients all tend to have or be closed units.15,21 These data suggest that the impediments to an organized critical care service may be due to insufficient ICU activity to support a full-time service, reimbursement strategies that provide incentives to care for one’s own patients in the ICU, and the reluctance of operating surgeons to relinquish control over care of their patients.

At least 16 studies have tackled, in one form or another, the question of the effect of admitting policy and authority on ICU outcome. Although the admitting policies in these studies and the methods are too heterogeneous to allow a formal quantitative meta-analysis, some general themes emerge. Generally, the studies show an improvement in outcome with a reduction in mortality with a closed ICU run by an intensivist. None of the studies show worsening mortality with a move toward a controlled admission policy. Almost all of the studies show some improvement in the way the ICU is used: reductions in monitor-only patients, reduction in length of stay, or reduction in ineffective treatments.

Governance Style. Older, hierarchical models of medical care, where a single physician autocratically determines management, are no longer tenable. As medicine becomes more complex and interdisciplinary, newer, collaborative models of governance are required.22 Modern quality improvement techniques rely on empowering all workers to identify novel solutions and work toward improving the process of care. Critical care epitomizes team-based, multidisciplinary care, and it seems fairly obvious that teams that communicate well, that empower all members to contribute ideas, and where responsibility is shared, will provide better care.

Specialty Units and Teams. One of the most reproducible observations in health services research is called the "volume-outcome" relationship.23 Across a broad range of medical treatments, the more experience a provider or hospital has with a particular disease or procedure, the better the outcomes appear to be. Two models have been explored: 1) bringing the patient to the experienced providers (specialized ICUs) and 2) bringing the experienced providers to the patient (special consult teams within a general ICU). Regionalization of critical care has been proposed as a way to improve outcomes by funneling critically ill patients to large-volume centers.

The Future

Technological innovation and changes in reimbursement may have profound effects on the organization of critical care over the next decade. These may lead to novel structures for ICUs or may lead to the dissolution of the ICU as we currently know it. In most hospitals, patients are moved to the ICU when they become critically ill. In the future, the technology may be brought to the patient or telemedicine may bring a distant doctor to the bedside.24,25

Thus far, we have dealt with the ICU as a structural entity in isolation. However, there are important organizational issues in how the ICU relates to the rest of hospital care and, more broadly, how critical care is incorporated into a regional health care system. Hospitalists, a new specialty of physician, have been proposed in part to meet the specialized needs of an increasingly complex and severely ill population of hospitalized patients.26 How hospitalists and intensivists will share responsibility for critically ill patients as they move from the emergency room, operating room, or hospital ward to the ICU and back again has not been carefully worked out.


The modern ICU is a structural entity. In it are focused specific human and material resources organized to deliver care to critically ill patients. While some improvements in ICU structure have face value and may not merit extensive empiric validation, many questions about the optimal structure of critical care remain. Studies demonstrating an effect of ICU structure on process of care or outcome are challenging because they do not lend themselves to randomized controlled trials; they present hierarchical data analysis problems, and their findings may not generalize beyond the ICU in which they were carried out.

Regardless of these limitations, ample data exist to convince us that the structure of health care has a profound effect on both process of care and outcome. Identifying which structural and organizational improvements apply in different ICU settings, and linking these to specific patient outcomes, are great challenges for future investigation.


1. Donabedian A. The quality of care. How can it be assessed? JAMA 1988;260:1743-1748.

2. Teres D, Lemeshow S. Using severity measures to describe high performance intensive care units. Crit Care Clin 1993;9:543-554.

3. Critical care services and personnel: Recommendations based on a system of categorization into two levels of care. Crit Care Med 1999;27:422-426.

4. Guidelines for intensive care unit design. Crit Care Med 1995;23:582-588.

5. American College of Surgeons, Committee on Trauma. Resources for Optimal Care of the Injured Patient: 1999. Chicago, Ill: American College of Surgeons; 1999.

6. Recommendations for critical care unit design. Crit Care Med 1988;16:796-806.

7. Robin ED. The cult of the Swan-Ganz catheter. Overuse and abuse of pulmonary flow catheters. Ann Intern Med 1985;103:445-449.

8. Robin ED. Death by pulmonary artery flow-directed catheter. Time for a moratorium? Chest 1987;92: 727-731.

9. Shortell SM, et al. The performance of intensive care units: Does good management make a difference? Med Care 1994;32:508-525.

10. Rubenfeld GD. Workforce and organizational change: Implications for cost containment. Seminars in Resp Crit Care Med 1999;20:245-251.

11. Jollis JG, et al. Outcome of acute myocardial infarction according to the specialty of the admitting physician. N Engl J Med 1996;335:1880-1887.

12. Kitahata MM, et al. Physicians’ experience with the acquired immunodeficiency syndrome as a factor in patients’ survival. N Engl J Med 1996;334:701-706.

13. Mitchell-DiCenso A, et al. A controlled trial of nurse practitioners in neonatal intensive care. Pediatrics 1996;98(6 Pt 1):1143-1148.

14. Dubaybo BA, et al. The role of physician-assistants in critical care units. Chest 1991;99:89-91.

15. Future Workforce Needs in Pulmonary and Critical Care Medicine. Cambridge: Abt Associates; 1998.

16. Keenan SP, et al. Ventilatory care in a selection of Ontario hospitals: Bigger is not necessarily better! Critical Care Research Network (CCR-Net). Intensive Care Med 1998;24:946-952.

17. Kohn LT, Corrigan J, Donaldson MS, eds. To Err Is Human: Building a Safer Health System. Washington, DC: National Academy Press; 1999.

18. Leape LL, et al. Pharmacist participation on physician rounds and adverse drug events in the intensive care unit. JAMA 1999;282:267-270.

19. Sibbald WJ, Massaro TA, McLeod DM, eds. The Business of Critical Care: A Textbook for Clinicians Who Manage Special Care Units. Armonk, NY: Futura Pub. Co.; 1996.

20. Fein IA, Strosberg MA. Managing the Critical Care Unit. Rockville, Md: Aspen Publishers; 1987.

21. Groeger JS, et al. Descriptive analysis of critical care units in the United States. Crit Care Med 1992;20: 846-863.

22. Pew Health Professions Commission. Critical challenges: Revitalizing the health professions for the twenty-first century. The Commission: San Francisco. 1995.

23. Hughes RG, et al. Effects of surgeon volume and hospital volume on quality of care in hospitals. Med Care 1987;25:489-503.

24. Maxwell G, Marion L. Flexible monitoring. Bringing technology to the patient. Nurs Manage 1997;28:48B, 48F-48G,48I.

25. Grundy BL, et al. Telemedicine in critical care: Problems in design, implementation, and assessment. Crit Care Med 1982;10:471-475.

26. Wachter RM, Goldman L. The emerging role of "hospitalists" in the American health care system. N Engl J Med 1996;335:514-517.

27. Vincent JL, et al. Organization of intensive care units in Europe: Lessons from the EPIC study. Intensive Care Med 1997;23:1181-1184.

Each of the following is an aspect of ICU structure except:

a. the number of ventilators the ICU owns.

b. the nurse-to-patient ratio.

c. the risk-adjusted mortality of the patients.

d. who is allowed to admit to the ICU.

e. whether the nurses and doctors work collaboratively.

Which of the following statements about U.S. ICUs is true?

a. All ICUs have an on-site 24-hour intensivist.

b. All ICUs have a step-down unit available.

c. All ICUs have at least six beds.

d. All ICUs are capable of performing continuous veno-venous hemofiltration.

e. None of the above