Original Article
Development of a clinical definition for acute respiratory distress syndrome using the Delphi technique

https://doi.org/10.1016/j.jcrc.2005.03.001Get rights and content

Abstract

Purpose

The objective of this study is to describe the implementation of formal consensus techniques in the development of a clinical definition for acute respiratory distress syndrome.

Materials and Methods

A Delphi consensus process was conducted using e-mail. Sixteen panelists who were both researchers and opinion leaders were systematically recruited. The Delphi technique was performed over 4 rounds on the background of an explicit definition framework. Item generation was performed in round 1, item reduction in rounds 2 and 3, and definition evaluation in round 4. Explicit consensus thresholds were used throughout.

Results

Of the 16 panelists, 11 actually participated in developing a definition that met a priori consensus rules on the third iteration. New incorporations in the Delphi definition include the use of a standardized oxygenation assessment and the documentation of either a predisposing factor or decreased thoracic compliance. The panelists rated the Delphi definition as acceptable to highly acceptable (median score, 6; range, 5-7 on a 7-point Likert scale).

Conclusions

We conclude that it is feasible to consider using formal consensus in the development of future definitions of acute respiratory distress syndrome. Testing of sensibility, reliability, and validity are needed for this preliminary definition; these test results should be incorporated into future iterations of this definition.

Introduction

There is no available diagnostic test to confirm the presence of acute respiratory distress syndrome (ARDS), and it is, therefore, defined by clinical criteria in a similar fashion to diseases such as rheumatoid arthritis and depression. Unlike many rheumatologic and psychiatric diseases, however, very little work has been done to develop and test standardized diagnostic criteria for ARDS [1]. Indeed, for many years after its original description [2], ARDS remained a gestalt diagnosis without specific criteria [3].

Translating the pathophysiologic concepts of ARDS [4], [5], [6], [7], [8], [9] into a working clinical definition has proved challenging. The widely used American-European Consensus Conference (AECC) definition requires the acute onset of hypoxemia and bilateral chest x-ray infiltrates in the absence of left atrial hypertension [10]. This definition has not performed well in limited reliability testing focusing on the chest radiograph [11], [12], and the agreement between this definition and the autopsy findings of ARDS is only moderate [13]. The recognition that rigorously developed and validated diagnostic criteria are vital to the performance of clinical trials in this area has led to calls for a reevaluation of this definition [14], [15], [16], [17].

The use of formal consensus techniques in medicine is becoming increasingly frequent, but in critical care, these have typically taken the form of consensus statements used to help guide clinical care [18], [19], [20], [21], [22], [23], [24]. Formal consensus techniques may be helpful in explicating the necessary judgmental approach to definition development in situations where no gold standard exists, potentially reducing bias and resulting in a definition with improved operating characteristics. The AECC definition, however, was created without the use of formal consensus techniques. We are unaware of any reports documenting the use of such methodologies in the derivation of definitions for ARDS or other clinical entities in critical care. A variety of formal methods is available when conducting a consensus process [19], [25], [26], [27]. One such method is the Delphi technique, an iterative process where experts are polled individually with a series of questionnaires, receiving anonymous group feedback between iterations [19]. In this paper, we describe the use of the Delphi technique in the development of a new ARDS definition. This represents the first phase of a strategy to develop and test a novel ARDS definition using rigorous methodologies.

Section snippets

Design

The Delphi technique was chosen over other consensus techniques because of its ability to overcome temporal and geographic constraints, low cost, and ability to ensure all the panelists an equal voice in the proceedings. Approval from the University of Toronto research ethics board was obtained.

Participants

A systematic method for identifying individuals who were both opinion leaders and clinical researchers was used to avoid bias in the selection of panelists [28]. First, opinion leaders were gathered by

Participants

The first 16 individuals who were approached agreed to participate in the consensus process. However, 11 (69%) of the 16 panelists actually participated in the definition development. Their names are listed in the acknowledgements. Demographic information on the participants is shown in Table 2.

Item generation—Delphi round 1

The first Delphi questionnaire, asking panelists to identify all possible defining characteristics for inclusion in a new ARDS definition, yielded 10 possible defining characteristics (Table 3). The

Discussion

Definitions for difficult-to-define intensive care unit (ICU) clinical entities such as sepsis, ventilator-associated pneumonia, and ARDS have not been approached previously using formal consensus techniques. Papers labeled as “consensus conferences” have been published addressing each of these examples, but in each case, the format used was an informal group discussion. Formal consensus methods, such as the Delphi technique, may be useful in reducing bias and enhancing the transparency of the

Acknowledgments

This study was supported in part by the Canadian Institutes of Health Research/Canadian Lung Association (Ottawa, Ontario, Canada) Post-Doctoral Fellowship (Dr Ferguson) and the Health Career Award from the Canadian Institutes of Health Research (Dr Davis). We thank the following individuals who participated as members of our expert panel during the Delphi process: Drs Antonio Artigas, Jean-Daniel Chiche, Gregory Downey, Margaret Herridge, Waldemar Johanson, Marin Kollef, James Lewis, Neil

References (44)

  • B.G. Garber et al.

    Adult respiratory distress syndrome: a systemic overview of incidence and risk factors

    Crit Care Med

    (1996)
  • O. Lesur et al.

    Acute respiratory distress syndrome: 30 years later

    Can Respir J

    (1999)
  • L.B. Ware et al.

    The acute respiratory distress syndrome

    N Engl J Med

    (2000)
  • G.R. Bernard et al.

    The American-European Consensus Conference on ARDS. Definitions, mechanisms, relevant outcomes, and clinical trial coordination

    Am J Respir Crit Care Med

    (1994)
  • M.O. Meade et al.

    Interobserver variation in interpreting chest radiographs for the diagnosis of acute respiratory distress syndrome

    Am J Respir Crit Care Med

    (2000)
  • A. Esteban et al.

    Comparison of clinical criteria for the acute respiratory distress syndrome with autopsy findings

    Ann Intern Med

    (2004)
  • E. Abraham

    Toward new definitions of acute respiratory distress syndrome

    Crit Care Med

    (1999)
  • E. Abraham et al.

    Consensus conference definitions for sepsis, septic shock, acute lung injury, and acute respiratory distress syndrome: time for a reevaluation

    Crit Care Med

    (2000)
  • K. Lewandowski

    Epidemiological data challenge ARDS/ALI definition

    Intensive Care Med

    (1999)
  • D.P. Schuster

    Identifying patients with ARDS: time for a different approach

    Intensive Care Med

    (1997)
  • A.J. Rotondi et al.

    Consensus conferences in critical care medicine. Methodologies and impact

    Crit Care Clinics

    (1997)
  • A. Fink et al.

    Consensus methods: characteristics and guidelines for use

    Am J Public Health

    (1984)
  • Cited by (111)

    • Acute lung injury

      2020, Current Problems in Surgery
    View all citing articles on Scopus
    View full text