POINT: Should Computerized Protocols Replace Physicians for Managing Mechanical Ventilation? Yes

Abstract

We begin with the premise that the objective of mechanical ventilator management is to use modes and settings for ventilation, oxygenation, and weaning that increase—based on valid evidence—the likelihood of favorable outcomes and avoidance of patient harm. Fundamental to this objective is for critical care physicians, nurses, and respiratory therapists (hereafter, clinicians) to operate the ventilator according to evidence-based practices. Unfortunately, extensive evidence documents that clinicians often make well-intended decisions that are inconsistent with evidence-based practices,1Morris A.H. Decision support and safety of clinical environments.Qual Saf Health Care. 2002; 11: 69-75Crossref PubMed Scopus (55) Google Scholar leading to needless and likely injurious variability in ventilator management. For example, low tidal volume lung protective ventilation improves mortality in patients with ARDS,2ARDS NetworkVentilation with lower tidal volumes as compared with traditional tidal volumes for acute lung injury and the acute respiratory distress syndrome. The Acute Respiratory Distress Syndrome Network.N Engl J Med. 2000; 342: 1301-1308Crossref PubMed Scopus (10124) Google Scholar is recommended in international guidelines for mechanical ventilation of patients with ARDS,3Fan E. Del Sorbo L. Goligher E.C. et al.An Official American Thoracic Society/European Society of Intensive Care Medicine/Society of Critical Care Medicine Clinical Practice Guideline: mechanical ventilation in adult patients with acute respiratory distress syndrome.Am J Respir Crit Care Med. 2017; 195: 1253-1263Crossref PubMed Scopus (787) Google Scholar and improves outcomes in patients on mechanical ventilation without ARDS,4Neto A.S. Simonis F.D. Barbas C.S. et al.Lung-protective ventilation with low tidal volumes and the occurrence of pulmonary complications in patients without acute respiratory distress syndrome: a systematic review and individual patient data analysis.Crit Care Med. 2015; 43: 2155-2163Crossref PubMed Scopus (164) Google Scholar, 5Serpa Neto A. Cardoso S.O. Manetta J.A. et al.Association between use of lung-protective ventilation with lower tidal volumes and clinical outcomes among patients without acute respiratory distress syndrome: a meta-analysis.JAMA. 2012; 308: 1651-1659Crossref PubMed Scopus (570) Google Scholar including patients undergoing abdominal surgery,6Futier E. Constantin J.M. Paugam-Burtz C. et al.A trial of intraoperative low-tidal-volume ventilation in abdominal surgery.N Engl J Med. 2013; 369: 428-437Crossref PubMed Scopus (885) Google Scholar but is inconsistently implemented in the ED7Allison M.G. Scott M.C. Hu K.M. Witting M.D. Winters M.E. High initial tidal volumes in emergency department patients at risk for acute respiratory distress syndrome.J Crit Care. 2015; 30: 341-343Crossref PubMed Scopus (9) Google Scholar, 8Fuller B.M. Mohr N.M. Dettmer M. et al.Mechanical ventilation and acute lung injury in emergency department patients with severe sepsis and septic shock: an observational study.Acad Emerg Med. 2013; 20: 659-669Crossref PubMed Scopus (58) Google Scholar and in the ICU.9Bellani G. Laffey J.G. Pham T. Epidemiology, patterns of care, and mortality for patients with acute respiratory distress syndrome in intensive care units in 50 countries.JAMA. 2016; 315: 788-800Crossref PubMed Scopus (2651) Google Scholar, 10Weiss C.H. Baker D.W. Tulas K. et al.A critical care clinician survey comparing attitudes and perceived barriers to low tidal volume ventilation with actual practice.Ann Am Thorac Soc. 2017; 14: 1682-1689Crossref PubMed Scopus (30) Google Scholar, 11Duan E.H. Adhikari N.K. D'Aragon F. et al.Management of ARDS and refractory hypoxemia: a multicenter observational study.Ann Am Thorac Soc. 2017; 14: 1818-1826Crossref PubMed Scopus (40) Google Scholar Guidelines also recommend a high positive-end expiratory pressure (PEEP) strategy for moderate and severe ARDS,3Fan E. Del Sorbo L. Goligher E.C. et al.An Official American Thoracic Society/European Society of Intensive Care Medicine/Society of Critical Care Medicine Clinical Practice Guideline: mechanical ventilation in adult patients with acute respiratory distress syndrome.Am J Respir Crit Care Med. 2017; 195: 1253-1263Crossref PubMed Scopus (787) Google Scholar, 12Briel M. Meade M. Mercat A. et al.Higher vs lower positive end-expiratory pressure in patients with acute lung injury and acute respiratory distress syndrome: systematic review and meta-analysis.JAMA. 2010; 303: 865-873Crossref PubMed Scopus (1058) Google Scholar but this too is inconsistently implemented.9Bellani G. Laffey J.G. Pham T. Epidemiology, patterns of care, and mortality for patients with acute respiratory distress syndrome in intensive care units in 50 countries.JAMA. 2016; 315: 788-800Crossref PubMed Scopus (2651) Google Scholar, 11Duan E.H. Adhikari N.K. D'Aragon F. et al.Management of ARDS and refractory hypoxemia: a multicenter observational study.Ann Am Thorac Soc. 2017; 14: 1818-1826Crossref PubMed Scopus (40) Google Scholar Even when clinicians intend to use a low tidal volume or a high PEEP strategy for ARDS, and think they are compliant, implementation is still variable.10Weiss C.H. Baker D.W. Tulas K. et al.A critical care clinician survey comparing attitudes and perceived barriers to low tidal volume ventilation with actual practice.Ann Am Thorac Soc. 2017; 14: 1682-1689Crossref PubMed Scopus (30) Google Scholar, 11Duan E.H. Adhikari N.K. D'Aragon F. et al.Management of ARDS and refractory hypoxemia: a multicenter observational study.Ann Am Thorac Soc. 2017; 14: 1818-1826Crossref PubMed Scopus (40) Google Scholar Current, clinician-driven models clearly fail to achieve the objective of mechanical ventilation. What then are our options? Protocols, whether paper or electronic, may help to standardize ventilator management.13Morris A.H. Developing and implementing computerized protocols for standardization of clinical decisions.Ann Intern Med. 2000; 132: 373-383Crossref PubMed Scopus (191) Google Scholar Paper protocols are currently available, including the ARDS Network ventilation protocol,2ARDS NetworkVentilation with lower tidal volumes as compared with traditional tidal volumes for acute lung injury and the acute respiratory distress syndrome. The Acute Respiratory Distress Syndrome Network.N Engl J Med. 2000; 342: 1301-1308Crossref PubMed Scopus (10124) Google Scholar or they could be created from published guidelines. Paper protocols, however, can be confusing and difficult to follow at the bedside. More detailed, and hence more reproducible, protocols poorly accommodate the multiple dimensions of ventilator management (eg, tidal volume and rate, fraction of inspired oxygen and PEEP, weaning and support), resulting in an unwieldy proliferation of forms. In our experience, clinicians quickly create simplifying (and inaccurate) heuristics—shortcuts to circumvent protocol logic—in response to complex paper protocols. Computerizing mechanical ventilation protocols to create electronic clinical decision support (CDS) tools assures accurate interpretation of protocol logic even in crisis situations, standardizes medical decisions, and reduces unnecessary variation for ventilator management, while providing an easy and interpretable audit trail for implementation science approaches to further improve ventilator management.13Morris A.H. Developing and implementing computerized protocols for standardization of clinical decisions.Ann Intern Med. 2000; 132: 373-383Crossref PubMed Scopus (191) Google Scholar, 14Sorenson D. Grissom C.K. Carpenter L. et al.A frame-based representation for a bedside ventilator weaning protocol.J Biomed Inform. 2008; 41: 461-468Crossref PubMed Scopus (14) Google Scholar, 15East T.D. Morris A.H. Wallace C.J. et al.A strategy for development of computerized critical care decision support systems.Int J Clin Monit Comput. 1991; 8: 263-269Crossref PubMed Scopus (48) Google Scholar Simple electronic CDS tools for mechanical ventilation (eg, notification of injurious tidal volumes) can supplement basic management strategies. Such simple CDS tools may improve compliance with low tidal volume ventilation,16Eslami S. Abu-Hanna A. Schultz M.J. de Jonge E. de Keizer N.F. Evaluation of consulting and critiquing decision support systems: effect on adherence to a lower tidal volume mechanical ventilation strategy.J Crit Care. 2012; 27 (e421-e428): 425Crossref PubMed Scopus (12) Google Scholar, 17Eslami S. de Keizer N.F. Abu-Hanna A. de Jonge E. Schultz M.J. Effect of a clinical decision support system on adherence to a lower tidal volume mechanical ventilation strategy.J Crit Care. 2009; 24: 523-529Crossref PubMed Scopus (25) Google Scholar but do not offer comprehensive, adequately explicit direction for all aspects of mechanical ventilation. A more complete approach to electronic CDS tools for mechanical ventilation is to create adequately explicit algorithms for ventilation, oxygenation, and weaning. Such CDS tools for mechanical ventilation provide treatment instructions defined in patient-specific terms after input of patient-specific data: they therefore provide replicable instructions for equivalent patient states. Computerized protocols must be developed and iteratively refined using a methodologically rigorous approach. At Intermountain Healthcare, we have developed a process for computerized ventilator protocol development using a knowledge engineering team collaborating with clinicians.15East T.D. Morris A.H. Wallace C.J. et al.A strategy for development of computerized critical care decision support systems.Int J Clin Monit Comput. 1991; 8: 263-269Crossref PubMed Scopus (48) Google Scholar, 18Blagev D.P. Hirshberg E.L. Sward K. et al.The evolution of eProtocols that enable reproducible clinical research and care methods.J Clin Monit Comput. 2012; 26: 305-317Crossref PubMed Scopus (17) Google Scholar The process includes developing protocol logic, computerization, batch processing using real or simulated data, iterative refinement with prospective validation in a controlled clinical environment, and clinical deployment with development of supportive educational resources and continued monitoring for adverse events and outcomes. This paradigm supported development of computerized CDS tools for mechanical ventilation with explicit instructions for oxygenation, ventilation, weaning, and extubation19East T.D. Heermann L.K. Bradshaw R.L. et al.Efficacy of computerized decision support for mechanical ventilation: results of a prospective multi-center randomized trial.Proc AMIA Symp. 1999; : 251-255PubMed Google Scholar, 20East T.D. Bohm S.H. Wallace C.J. et al.A successful computerized protocol for clinical management of pressure control inverse ratio ventilation in ARDS patients.Chest. 1992; 101: 697-710Abstract Full Text Full Text PDF PubMed Scopus (83) Google Scholar that was exported to 10 clinical sites in a prospective randomized trial of 200 patients with ARDS that compared CDS tool use with usual care.19East T.D. Heermann L.K. Bradshaw R.L. et al.Efficacy of computerized decision support for mechanical ventilation: results of a prospective multi-center randomized trial.Proc AMIA Symp. 1999; : 251-255PubMed Google Scholar The tool was active 96% of the time and 38,546 instructions were generated, with 94% compliance. Patients in the CDS tool arm had reduced organ dysfunction and barotrauma. This trial suggests that CDS tool export is feasible, without extensive local adaptation. The key to tool export, however, is modularity and interoperability. With attention to these principles, replicable, adaptive mechanical ventilation CDS tools can be widely disseminated.14Sorenson D. Grissom C.K. Carpenter L. et al.A frame-based representation for a bedside ventilator weaning protocol.J Biomed Inform. 2008; 41: 461-468Crossref PubMed Scopus (14) Google Scholar The computerized CDS tools evaluated to date are open loop (ie, the clinician must decide whether to follow the instruction). At Intermountain Healthcare, we have used open-loop computerized CDS tools for mechanical ventilation for > 25 years on a wide array of patients with and without ARDS.14Sorenson D. Grissom C.K. Carpenter L. et al.A frame-based representation for a bedside ventilator weaning protocol.J Biomed Inform. 2008; 41: 461-468Crossref PubMed Scopus (14) Google Scholar, 15East T.D. Morris A.H. Wallace C.J. et al.A strategy for development of computerized critical care decision support systems.Int J Clin Monit Comput. 1991; 8: 263-269Crossref PubMed Scopus (48) Google Scholar, 20East T.D. Bohm S.H. Wallace C.J. et al.A successful computerized protocol for clinical management of pressure control inverse ratio ventilation in ARDS patients.Chest. 1992; 101: 697-710Abstract Full Text Full Text PDF PubMed Scopus (83) Google Scholar This adequately explicit open-loop mechanical ventilation CDS tool automatically loads patient-specific data and generates instructions for the respiratory therapist to review and implement at clinically relevant time intervals. The respiratory therapists remain intellectually engaged, consult with physicians as needed, and implement CDS tool instructions, with rare exceptions. We track declined instructions to identify systematic problems or patient-specific contexts that may require iterative refinement of the logic in the computerized CDS tool, a further level of oversight for the computerized algorithm. We recently ported the entire mechanical ventilation CDS tool to Cerner, using our previously described method for development and implementation of computerized CDS tools.15East T.D. Morris A.H. Wallace C.J. et al.A strategy for development of computerized critical care decision support systems.Int J Clin Monit Comput. 1991; 8: 263-269Crossref PubMed Scopus (48) Google Scholar, 18Blagev D.P. Hirshberg E.L. Sward K. et al.The evolution of eProtocols that enable reproducible clinical research and care methods.J Clin Monit Comput. 2012; 26: 305-317Crossref PubMed Scopus (17) Google Scholar Among our five tertiary hospitals, the two hospitals with the highest mechanical ventilation CDS tool use had the highest on target tidal volumes of ≤ 6.5 mL/kg predicted body weight (PBW) of 88%, a dramatically better performance than has been reported in recent studies.9Bellani G. Laffey J.G. Pham T. Epidemiology, patterns of care, and mortality for patients with acute respiratory distress syndrome in intensive care units in 50 countries.JAMA. 2016; 315: 788-800Crossref PubMed Scopus (2651) Google Scholar, 10Weiss C.H. Baker D.W. Tulas K. et al.A critical care clinician survey comparing attitudes and perceived barriers to low tidal volume ventilation with actual practice.Ann Am Thorac Soc. 2017; 14: 1682-1689Crossref PubMed Scopus (30) Google Scholar, 11Duan E.H. Adhikari N.K. D'Aragon F. et al.Management of ARDS and refractory hypoxemia: a multicenter observational study.Ann Am Thorac Soc. 2017; 14: 1818-1826Crossref PubMed Scopus (40) Google Scholar There seems little reason to doubt that computerized CDS tools improve compliance with evidence-based mechanical ventilation strategies. Closed-loop protocols—which drive the ventilator without clinician involvement in each instruction—are the new frontier in electronic CDS, increasingly being offered by mechanical ventilator manufacturers.21Burns K.E. Lellouche F. Lessard M.R. Automating the weaning process with advanced closed-loop systems.Intensive Care Med. 2008; 34: 1757-1765Crossref PubMed Scopus (53) Google Scholar Closed-loop protocols are already used clinically to adapt ventilator support in patients dependent on mandatory breaths, to transition patients from controlled to support modes, and to automate ventilator weaning. Closed-loop protocols for weaning have advantages, if only for allowing weaning to proceed quickly, even when clinicians are busy caring for other patients who are unstable. The next step in development will be extension of closed-loop mechanisms to other ventilator tasks. Such systems require high-fidelity, continuous data for all the relevant data streams, which may be difficult to achieve in many clinical realms. Some human oversight with specific triggers will likely remain important. For example, patient ventilator dyssynchrony represents a complex problem that may be difficult to model deterministically. In the validated ARDS Network low tidal volume ventilation protocol, if severe dyspnea is associated with frequent double stacking of ventilator breaths, then the tidal volume may be increased from 6 up to 7 or 8 mL/kg PBW, and increased sedation (and/or temporary paralysis) may be required to keep plateau pressure < 30 cm H2O.2ARDS NetworkVentilation with lower tidal volumes as compared with traditional tidal volumes for acute lung injury and the acute respiratory distress syndrome. The Acute Respiratory Distress Syndrome Network.N Engl J Med. 2000; 342: 1301-1308Crossref PubMed Scopus (10124) Google Scholar It may be counterproductive to standardize ventilator decisions for rare patient scenarios of dramatic ventilator dyssynchrony.1Morris A.H. Decision support and safety of clinical environments.Qual Saf Health Care. 2002; 11: 69-75Crossref PubMed Scopus (55) Google Scholar We prefer an open-loop CDS tool for mechanical ventilation to allow the clinician to intervene and tailor ventilator adjustments for these complex and uncommon clinical situations, while supplementing ventilator management with other interventions. This state of affairs may change with improvements in machine learning algorithms to characterize and subtype dyssynchrony and other difficult ventilator management situations. Even as more and more tasks prove susceptible to computerized control, human involvement that plays to human strengths will remain important and should likely constitute the target of new training and curricula for clinicians operating ventilators. Safety in the clinical environment depends on structures that reduce the probability of harm, evidence that enhances the likelihood of actions that increase favorable outcomes, and explicit directions that implement the actions dictated by this evidence.1Morris A.H. Decision support and safety of clinical environments.Qual Saf Health Care. 2002; 11: 69-75Crossref PubMed Scopus (55) Google Scholar The use of adequately explicit computerized CDS tools can serve all three purposes in improving mechanical ventilation. We do not mean thereby to devalue the role of clinicians. Even as computerized protocols replace the cognitive work of most decisions to be made in ventilator management, we think that by replacing the aspects of ventilator management that humans do poorly, we create new opportunities for clinicians to play to their strengths—dealing with crises or unusual patient conditions, evaluating the work of automated systems, providing the human touch to suffering patients and families, and assuring that patients are correctly integrated into relevant protocols. We think that future training of clinicians—likely with extensive use of simulations—should emphasize precisely those higher-level, nonroutine tasks, while having the routine tasks performed by adequately explicit computerized CDS tools. This hybrid model, which largely replaces flawed human management of key aspects of ventilator management, in our view is most likely to improve compliance with evidence-based mechanical ventilation practices and improve patient safety and outcomes. /cms/asset/1ae6d2b7-8ee0-4b12-bbf6-1245e4892767/mmc1.mp3Loading ... Download .mp3 (32.65 MB) Help with .mp3 files Audio Rebuttal From Drs Grissom and BrownCHESTVol. 154Issue 3PreviewWe appreciate MacIntyre’s1 argument against closed-loop computerized mechanical ventilator management. We agree with several of his comments and respectfully disagree with others. We emphasize the difference between open-loop and closed-loop computerized protocols. In open-loop computerized ventilation protocols, all instructions are reviewed by a clinician before implementation, whereas closed-loop protocols do not require clinician involvement. In our pro argument, we focused on open-loop computerized protocols implementing evidenced-based algorithms to create instructions that are reviewed by a clinician before implementation. Full-Text PDF COUNTERPOINT: Should Computerized Protocols Replace Physicians for Managing Mechanical Ventilation? NoCHESTVol. 154Issue 3PreviewThe first generation of positive-pressure mechanical ventilators were simple high-pressure gas regulators on which clinicians could set the circuit pressure and the breathing frequency. In the middle of the 20th century, more sophisticated devices appeared, which allowed direct clinician control of flow and volume along with breath timing and expiratory pressure. As ventilator design improved and microprocessors became available, feedback mechanisms appeared that could provide automatic adjustments in these set variables depending on a variety of conditions. Full-Text PDF Rebuttal From Dr MacIntyreCHESTVol. 154Issue 3PreviewThe title of our pro-con discussion is “Should computerized protocols replace physicians for managing mechanical ventilation?” To me, the word replace is the key word, and the focus of my argument was that replacement of a clinician by a computer for ventilator management is simply not feasible today. I base this argument on the fact that our mechanical ventilation management evidence base is quite limited, and we lack consensus on how to best balance the often competing goals of gas exchange support vs ventilator-induced lung injury and oxygen toxicity, and comfort/synchrony vs muscle dysfunction and sedation. Full-Text PDF