Special Feature

Research Results, Pathophysiologic Reasoning, and Clinical Experience: How Should Critical Care Practice be Guided?

By David J. Pierson, MD, Editor, Professor Emeritus, Pulmonary and Critical Care Medicine, University of Washington, Seattle, is Editor for Critical Care Alert.

A middle-aged man with community-acquired pneumonia complicated by underlying cardiomyopathy develops the acute respiratory distress syndrome (ARDS). He is intubated and ventilated according to the institution's lung-protective ventilation protocol, but as positive end-expiratory pressure (PEEP) is increased in response to worsening hypoxemia he becomes hypotensive and tachycardic. Boluses of intravenous crystalloid produce only transient improvement, and an infusion of norepinephrine is started. On 70% oxygen and 14 cm H2O PEEP, the patient's PaO2 is 70 mmHg. The resident asks the attending intensivist whether a pulmonary arterial catheter (PAC) should be inserted to guide this patient's management. She is told:

• No. Multiple large studies have shown that PACs do not improve outcomes in patients with sepsis or ARDS.


• Yes. The persistent tachycardia, underlying cardiac dysfunction, and high PEEP make accurate clinical assessment of the patient's volume status very difficult, and data from a PAC may clarify the issue. If the filling pressures are low, then giving more fluid may permit further PEEP increases without compromising cardiac function, and allow the FIO2 to be decreased as oxygenation improves. On the other hand, if the wedge pressure turns out to be high, then interventions to improve cardiac contractility are needed, and giving more fluid would be inadvisable.

This hypothetical scenario introduces a topic of fundamental importance in today's practice of critical care: how the results of published studies, pathophysiologic reasoning, and other forms of knowledge should be used in guiding clinical decision-making. The intensivist's first response rests on a seemingly secure foundation of evidence-based medicine (EBM), given the extensive literature on PACs and clinical outcomes indicating that their routine use in the ICU fails to improve mortality and tends to increase complications and costs. But does the available evidence include results from patients demographically or physiologically like the one described, managed according to other aspects of current best evidence? His second response seems very sound, given the stated pathophysiologic rationale along with the mutually exclusive approaches to management that the PAC's data might indicate. But will more numbers and more complete physiologic characterization of the situation really benefit the patient in a way that would matter to him in addition to (presumably) enlightening his physicians?

In addition to choosing one of the options given, the attending intensivist might have chosen any of a number of alternative therapeutic interventions — each of them supported by at least some clinical research and/or pathophysiologic evidence, and each within the "accepted scope of practice" in many institutions. These might include different approaches to ventilatory support — such as recruitment maneuvers, airway pressure release ventilation, or high-frequency oscillatory ventilation — or prone positioning, neuromuscular blockade, or extracorporeal membrane oxygenation. The fact that the benefits and hazards of these and other interventions in critical care remain uncertain, as applied to patients with potentially fatal illness by clinicians whose personal experience often gives them strong but divergent preferences for management, makes reconciling the different types of clinical knowledge and the variation encountered in ICU practice critically important.

This special feature is based largely on the recent report of a multi-society working group on the role of clinical research results in the practice of critical care medicine.1 I will summarize some of the issues and arguments addressed by the group as these relate to the ICU clinician, although the published statement discusses many aspects of these and other things in more detail than can be presented here.


The era of EBM began 20 years ago with a paper by Guyatt and colleagues.2 The goal of EBM is to apply the best available evidence gained from the scientific method to clinical decision making. Ideally, through the use of EBM, clinicians will incorporate all available and relevant medical knowledge in making decisions about how best to manage an individual patient.1 In pursuit of this goal, EBM seeks to assess the strength of the evidence of risks and benefits of treatments and diagnostic tests. Numerous hierarchies of the various forms of clinical evidence have been put forth in pursuit of this assessment. The following is one such hierarchy:3

  • Evidence from randomized controlled trials.
  • Systematic reviews of observational studies.
  • Physiologic studies.
  • Unsystematic clinical observations.

Ranked even higher than individual randomized, controlled trials are meta-analyses of multiple such trials; evidence from personal experience would fall somewhere below the last category on the list.

Criticism of and disagreement with the precepts of EBM have been advanced ever since its introduction. Although its proponents have always stressed the need to place best evidence in the context of the individual patient, and to evaluate the populations, interventions, and practice settings of published studies for relevance to each clinician's circumstances, EBM has been criticized as "cookbook medicine" that robs the practitioner of individual choice and clinical freedom. Its value and applicability in critical care have been hotly debated,4,5 and remain controversial with respect to the importance of other kinds of evidence besides the results of clinical trials in guiding patient management.1

To address this situation, in 2009 the American College of Chest Physicians, the American Thoracic Society, and the Society of Critical Care Medicine assembled the working group that produced the recent official multi-society statement.1 Besides the then-presidents of the three societies and the first author (an intensivist-educator and ethicist), the group included several experts in critical care research and education who were known to represent a spectrum of viewpoints on the issues under discussion, plus a philosopher and a patient advocate. Their purpose was to identify and clarify the issues involved, and then to create a conceptual framework for clinicians with respect to incorporating the results of clinical research into the practice of critical care medicine. Over a 2-year period the group had a series of face-to-face meetings, conference calls, and electronic communications, leading to unanimous consensus on the statements contained in the published document.1 There was no intent to accept or reject the principles of EBM, nor to resolve any of the controversies related to specific interventions or practices in the ICU. Instead, the goal was to provide clinicians with a conceptual framework for evaluating discordant information and conflicting claims when caring for critically ill patients.


A central message of the working group's document is that clinical knowledge comes in several forms, that each of these is important, and that none of them is sufficient by itself for optimal clinical decision-making. The main types of knowledge discussed are the results of clinical research, pathophysiologic reasoning, and knowledge gained from experience.1

Clinical research, via properly designed randomized clinical trials (RCTs), can yield data to address specific clinical questions and has the advantage of controlling and minimizing the bias that inevitably comes with personal experience. However, with the strengths of clinical research also come its weaknesses when it comes to applying the results to the individual patient at hand. Researchers in clinical trials study populations, and in order to focus on the variables of interest, they tend to define those populations narrowly, excluding those with atypical features or comorbidities. The protocols used are necessarily very specific, rigid, and rigorously administered. In addition, the results of RCTs are fixed in time and place, and may become less applicable as technical capabilities, other aspects of practice, and the culture of the ICU change. Knowledge gained in RCTs thus cannot be directly and indiscriminately applied to individual patients without careful examination of the circumstances, populations, and protocols used. Attesting to this problem is the frequently demonstrated difference between efficacy (results obtained in the constrained context of an RCT) and clinical effectiveness (what happens when the intervention is employed in routine clinical practice).6 Apparently similar RCTs can yield contradictory results, and the lack of comparability among different studies leads to disagreement and controversy, even when there is considerable evidence available.7-10

A central — and attractive — feature of our field is the prominent role played by physiology in the presentation and evolution of critical illness. It permits a degree of understanding, as well as opportunities for assessment and intervention that are often not possible in other specialties and practice contexts. However, using pathophysiologic reasoning to guide the management of critically ill patients has several shortcomings. Some patients manifest illness and respond to interventions differently from others. There may be important disconnects between desired physiologic effects and improved clinical outcomes. For example, effective suppression of arrhythmias after acute myocardial infarction using flecainide and similar agents resulted in increased mortality.11 Ventilating patients with ARDS using larger tidal volumes and inflation pressures improved arterial oxygenation but worsened mortality.12 And while prone positioning improves oxygenation in patients with hypoxemic respiratory failure, meta-analysis of RCTs showed no effect on survival.13 In addition, as pointed out by the working group, our understanding of the mechanisms involved in critical illness may not be as accurate or as complete as we think.1

While emphasis has been placed on the results of clinical research and on pathophysiologic reasoning, experiential knowledge is also an irreplaceable component of clinical practice, and a legitimate basis for medical decision-making. The source of experiential knowledge may be the individual clinician but can also be derived from other expert clinicians. The problem with this form of knowledge, besides the limited personal experience of individual clinicians in comparison with what can be found in the literature, is the inevitability of bias and the impossibility of its avoidance when used to guide patient care. Better control of bias is a main advantage of clinical research over experiential knowledge, which is an important reason the latter cannot serve as the only source in medical decision-making. Relying on experience tends to make practice patterns more static. And while experience may increase the clinician's levels of comfort and confidence, it may not guarantee greater expertise.


While much attention has been given to reducing practice variation by means of clinical practice guidelines and protocols, and numerous studies have shown improved outcomes with their implementation, practice uniformity comparable to that achieved by airline pilots and others in industry is unlikely to occur in critical care. The working group offers a number of reasons why, even in the presence of evidence from RCTs and other relevant knowledge, intensivists are likely to demonstrate considerable variation in how they use their knowledge. These reasons include the following:1

  • Whether the evidence is consistent with the clinician's prior knowledge or belief.
  • Biologic plausibility of the evidence in relation to the clinician's prior knowledge.
  • Similarity of patient population and/or intervention to those in the clinician's practice.
  • Size of reported effect or benefit.
  • Whether the benefit or outcome is perceived to be of high value.
  • How rapidly the effect or benefit is likely to occur.
  • Perceived costs and safety.
  • Relative ease or difficulty of implementation in the clinician's practice setting.

Just as they vary with respect to natural caution vs aggressiveness, clinicians also vary in their tendencies to adopt and adhere to protocols and routines. There are also "early adopters" who quickly embrace new technology and interventions, and those who are reluctant to change established practice patterns. Because of these and other factors, different intensivists will inevitably use the available evidence differently and there will be practice variation.


In one of its key and most emphatic statements, the working group emphasizes the importance of explicitness and transparency in the context of this practice variation.1 "While there may be important elements of clinicians' knowledge that are tacit, clinicians should be able to identify and articulate the sources and kinds of knowledge that are being invoked in support of a particular clinical decision." Further, the clinician "ought to be able to concisely outline and justify the process of clinical reasoning, elucidating the facts and reasoning supporting a particular decision, such as in the assessment portion of a clinical note, in a presentation to a colleague or trainee, or in discussion with the patient or family."

"Because I say so" is not an acceptable explanation for a management decision in the care of a critically ill patient, particularly when that decision varies from protocol or when not all members of the team agree. Disagreement is to some extent inevitable in the high-stakes environment of the ICU, but everyone involved in the patient's care needs to know what the plan is and why — hence the importance of explicitness, as described above.


In summarizing their discussion of the various types of knowledge and the factors weighing on their use by clinicians, the working group reached consensus on the following points:

  • Clinical research results, pathophysiologic reasoning, and clinical experience each represent a different kind of medical knowledge, and each is crucial for effective clinical decision-making.
  • When utilized in the care of individual patients, each kind of medical knowledge has different strengths and weaknesses: none is by itself sufficient to guide clinical decisions, and none takes precedence over the others.
  • Patient and/or family preferences and features of the system in which care is delivered also represent important considerations in medical decision-making.
  • Clinical research will be more or less compelling to individual clinicians depending on various factors that are independent of study design and statistical validity.
  • Practice variation may be acceptable when based on different weighting of conflicting medical knowledge or different patient or clinician values.
  • Explicitness is a hallmark of sound clinical reasoning and is necessary in assessing the causes and appropriateness of practice variability.

The accompanying figure depicts graphically the concepts developed in this important publication and summarized in the above list.



  1. Tonelli MR, et al; ACCP/ATS/SCCM Working Group. An official multi-society statement: The role of clinical research results in the practice of critical care medicine. Am J Respir Crit Care Med 2012;185:1117-1124.
  2. Evidence-Based Medicine Working Group. Evidence-based medicine. A new approach to teaching the practice of medicine. JAMA 1992;268:2420-2425.
  3. Montori VM, Guyatt GH. What is evidence-based medicine and why should it be practiced? Respir Care 2001;46:1201-1214.
  4. Paul J, et al. Point: Evidence-based medicine has a sound scientific base. Chest 2008;133:1067-1071, 1074-1075.
  5. Tobin MJ. Counterpoint: Evidence-based medicine lacks a sound scientific base. Chest 2008;133:1071-1074, 1076-1077.
  6. Pierson DJ. Translating evidence into practice. Respir Care 2009; 54:1386-1401.
  7. Marini JJ, Rubenfeld G. Pro/con clinical debate: The use of prone positioning in the management of patients with acute respiratory distress syndrome. Crit Care 2002;6:15-17.
  8. Hubmayr RD. Point: Is low tidal volume mechanical ventilation preferred for all patients on ventilation? Yes. Chest 2011;140:9-11.
  9. Gattinoni L. Counterpoint: Is low tidal volume mechanical ventilation preferred for all patients on ventilation? No. Chest 2011;140:11-13; discussion 14-15.
  10. Rubenfeld GD. Understanding why we agree on the evidence but disagree on the medicine. Respir Care 2001;46:1442-1449.
  11. Epstein AE, et al. Mortality following ventricular arrhythmia suppression by encainide, flecainide, and moricizine after myocardial infarction. JAMA 1993;270:2451-2455.
  12. The Acute Respiratory Distress Syndrome Network: Ventilation with lower tidal volumes as compared with traditional tidal volumes for acute lung injury and the acute respiratory distress syndrome. N Engl J Med 2000;342:1301-1308.
  13. Sud S, et al. Effect of mechanical ventilation in the prone position on clinical outcomes in patients with acute hypoxemic respiratory failure: A systematic review and meta-analysis. CMAJ 2008;178:1174-1176.