Determining the Basis for a Taxonomy of Mechanical Ventilation

Discussion

Over the last 30 years there has been a rapid evolution of the technologic capabilities of mechanical ventilators. These capabilities have motivated the creation of dozens of names to describe modes of ventilation. This proliferation of names has become counterproductive, making education of end users very difficult and potentially affecting the quality of patient care. Recognizing this dilemma, the International Standards Organization is attempting to create a controlled vocabulary of mechanical ventilation to standardize mode descriptions among manufacturers. That effort has led to the appreciation of the larger need to create a taxonomy, and even an ontology, of mechanical ventilation to support future developments in the semantic web (see the Art and Science of Taxonomy in the online supplemental text). This study was conceived to support such endeavors. Specifically, we sought to determine whether prior publications describing mode terminology and classification were sufficiently disseminated to suggest that an informal consensus currently exists, at least in principle. That consensus was hypothesized to exist among stakeholders in the international respiratory care community, including physicians, nurses, engineers, and manufacturers' representatives working in clinical, educational, and business environments.

This study generated survey results that indicated a significant level of concordance with 10 basic constructs of mechanical ventilation among stakeholders. Not surprisingly, the level of concordance was highest among respiratory therapists, presumably because they are more familiar with the prior art as it was published in journals and textbooks aimed at their profession. Of note was the fact that engineers had the lowest level of agreement with the construct that ventilatory assistance can be achieved “by only 2 means: control of the inspiratory volume/flow waveform, or control of the inspiratory pressure waveform.” Possible explanations for this result include the fact that these engineers were not aware of the construct because it appeared almost exclusively in clinically oriented journals (as opposed to engineering journals) and because engineers typically recoil at the idea of there being “only” a set number of ways to control a system. Hence, the engineers, as well as many others who responded, may not have appreciated the larger context of the survey questions (ie, creating a universal taxonomy) and tended to interpret them within their own narrow perspective.

The fact that there were no differences among responses when grouped according to professional activity is perhaps not surprising. Biases are held by people, not abstract professions. In addition, each professional activity has a mixture of professionals, so any potential differences would tend to be canceled out.

The finding that concordance varied among constructs invites discussion. The least concordant construct was represented by question 1: “A breath is defined as an inspiration paired with an expiration, where the 2 are matched by size or timing, thus allowing small fast breaths on top of large slow breaths during mechanical ventilation.” The comments indicated that the low concordance with this construct was due, in part, to the perceived complexity/ambiguity of the question. Part of the problem is the compound nature of the construct. But simply postulating that a breath is an inspiration paired with an expiration would have been trivial. The issue of matching of size is relevant in the context of some modes that first impose a large machine initiated inspiration, then allow the patient to breathe relatively smaller breaths for a short period, followed by a large machine initiated expiration (eg, airway pressure release ventilation). There is also the alternative possibility that the ventilator imposes small, rapid breaths upon slower, larger, patient initiated breaths (eg, high frequency oscillatory ventilation). Clearly there was no way to explain the full context of the question, yet it had to be asked. Confusion resulted because respondents were not familiar with thinking about the larger context. Nevertheless, if the survey were repeated, this question would have to be better worded.

The question with the second lowest concordance was number 2: “A ventilator can assist a breath by only 2 means: control of the inspiratory volume/flow waveform (ie, volume control), or control of the inspiratory pressure waveform (ie, pressure control).” The lower concordance for this construct may be a direct result of the confusion in the industry about the word “assist.” Some authors and manufacturers have promulgated the notion that any patient-initiated breath is an “assisted” breath. On the contrary, physiologists and physicists tell us that the word assist has nothing to do with what the patient does but rather relies on whether or not the ventilator performs work on the respiratory system. This misunderstanding is tied to a much larger issue affecting the formation of a taxonomy: there is a fundamental disagreement among stakeholders who view the ventilator in terms of how it is used clinically (ie, from the patient's point of view) and those who view it in terms of what it does mechanically (ie, from the operator's point of view). For example, the term “assist/control” has been very popular for decades to name a mode wherein either the patient triggers the breath (assist) or the ventilator triggers the breath (control). Some authors even go so far as to divide ventilatory support into 2 groups: “assisted” modes and “controlled” modes, where “assisted” means “the patient interacts in some way with the ventilator,” and “controlled” means “the machine determines the rate, the inspiratory time, and the tidal volume, so the patient does no work.”⁹ This viewpoint is anachronistic and dates to decades ago when patient initiation of inspiration was a novelty. And this is fundamentally a patient-centric view, because those who use the term “assist/control” are only interested in whether the ventilator responds to the patient's inspiratory effort or whether the breath is imposed on the patient. From the point of view of creating a taxonomy, the term assist/control is less than useless because it elevates a very low level system feature (the trigger variable) to the most important hierarchical position. Indeed, the terms assist and control as used in this fashion relate to no other constructs in a hierarchical way and are in fact names rather than tags (ie, taxonomic attribute groupings). Worse, they conflict with more important meanings of the word assist and control that do relate to fundamental constructs. Thus, we may conclude that the low concordance with construct number 2 indicates where more education needs to be directed. In fact, we could argue that this particular construct indicates the key issue: getting people to stop thinking in terms of names promulgated by manufacturers and start thinking about general terms of classification (when appropriate of course, as in educational programs and operator manuals).

The construct with the next lowest concordance was associated with question number 8: “Being able to identify any mode as one of these 5 basic ventilatory patterns has little practical value: 1. VC-CMV, 2. VC-IMV, 3. PC-CMV, 4. PC-IMV, 5. PC-CSV. VC = volume control. PC = pressure control.” In order to agree with this construct, the respondent would have to be aware of the expanded definitions of the terms CMV, IMV, and CSV. Despite the fact that question 7 (the next higher concordance level) defined these terms, most people probably have a more restricted understanding of them and associate them with specific mode names on specific ventilators rather than viewing them as more generic categories of mode characteristics. Again, this bias was expected, and, indeed, the fact that the majority of responses were concordant is encouraging.

The application of the 10 constructs described in this study to the development of a formal taxonomy of mechanical ventilation is beyond the scope of this paper. However, the confusion surrounding terms related to modes of ventilation has been acknowledged by the editorial board of Respiratory Care. As a result, the Journal publicly endorsed the adoption of standard vocabulary to describe ventilator modes. Preferred nomenclature and a glossary of terms used to describe ventilator modes (consistent with the constructs described in this study) are included in the guidelines for manuscript preparation and published in the Journal and on its Web site (Table 6).¹⁰

Recently, the American Association for Respiratory Care assembled a task force to identify potential new roles and responsibilities for respiratory therapists and to suggest the elements of education, training, and competency-documentation needed to assure safe and effective execution of those roles and responsibilities within the next 5 years and beyond.¹¹ The group identified the important role education and competency assessment in mechanical ventilation will play in our profession's future.¹² Education and competency assessment of the modes and principles of invasive and noninvasive mechanical ventilation are being required by physician-based organizations as well. A multi-society task force identified and defined modes and principles of invasive and noninvasive mechanical ventilation as one of the essential competencies to be addressed for internal medicine-based critical care and pulmonary medicine.¹³

To safely and effectively manage a mechanical ventilator, the respiratory therapist must thoroughly understand ventilator design, classification, and operation in order to properly match ventilator capability with physiologic need. The proliferation of names for modes of ventilation and lack of a standardized nomenclature may contribute to confusion and inappropriate clinical application of ventilatory modes. The effects may impair gas exchange and compromise pulmonary mechanics and patient safety. Lack of understanding of ventilator design can contribute to documentation issues.¹⁴ Documentation errors in the medical record increase the propensity of medical decision errors and adverse patient outcomes.^15,16