|Year : 2021 | Volume
| Issue : 4 | Page : 227-231
|Protocol failure detection: The conflation of acute respiratory distress syndrome, SARS-CoV-2 pneumonia and respiratory dysfunction
Lawrence A Lynn1, Emily Wheeler2, Russel Woda3, Alexander B Levitov4, Stanislaw P Stawicki5, David P Bahner6
1 Department of Pulmonology, OhioHealth Doctors Hospital, Columbus, Ohio, USA
2 Department of Emergency Medicine, Florida State University, Sarasota Memorial Hospital, Sarasota, Florida, USA
3 Department of Anesthesiology, Wexner Medical Center, The Ohio State University, Columbus, Ohio, USA
4 Department of Pulmonary and Critical Care Medicine, Eastern Virginia Medical School, Norfolk, Virginia, USA
5 Department of Research and Innovation, St Luke's University Health Network, Bethlehem, Pennsylvania, USA
6 Department of Emergency Medicine, The Ohio State University, Columbus, Ohio, USA
Click here for correspondence address and email
|Date of Submission||04-Jun-2021|
|Date of Acceptance||25-Aug-2021|
|Date of Web Publication||24-Dec-2021|
| Abstract|| |
In medicine, protocols are applied to assure the provision of the treatment with the greatest probability of success. However, the development of protocols is based on the determination of the best intervention for the group. If the group is heterogeneous, there will always be a subset of patients for which the protocol will fail. Furthermore, over time, heterogeneity of the group may not be stable, so the percentage of patients for which a given protocol may fail may change depending on the dynamic patient mix in the group. This was thrown into stark focus during the severe acute respiratory syndrome-2 coronavirus (SARS-CoV-2) pandemic. When a COVID-19 patient presented meeting SIRS or the Berlin Criteria, these patients met the criteria for entry into the sepsis protocol and/or acute respiratory distress syndrome (ARDS) protocol, respectively and were treated accordingly. This was perceived to be the correct response because these patients met the criteria for the “group” definitions of sepsis and/or ARDS. However, the application of these protocols to patients with SARS-CoV-2 infection had never been studied. Initially, poor outcomes were blamed on protocol noncompliance or some unknown patient factor. This initial perception is not surprising as these protocols are standards and were perceived as comprising the best possible evidence-based care. While the academic response to the pandemic was robust, recognition that existing protocols were failing might have been detected sooner if protocol failure detection had been integrated with the protocols themselves. In this review, we propose that, while protocols are necessary to ensure that minimum standards of care are met, protocols need an additional feature, integrated protocol failure detection, which provides an output responsive to protocol failure in real time so other treatment options can be considered and research efforts rapidly focused.
Keywords: Acute respiratory distress syndrome, COVID-19, protocol failure, SARS-CoV-2
|How to cite this article:|
Lynn LA, Wheeler E, Woda R, Levitov AB, Stawicki SP, Bahner DP. Protocol failure detection: The conflation of acute respiratory distress syndrome, SARS-CoV-2 pneumonia and respiratory dysfunction. J Emerg Trauma Shock 2021;14:227-31
|How to cite this URL:|
Lynn LA, Wheeler E, Woda R, Levitov AB, Stawicki SP, Bahner DP. Protocol failure detection: The conflation of acute respiratory distress syndrome, SARS-CoV-2 pneumonia and respiratory dysfunction. J Emerg Trauma Shock [serial online] 2021 [cited 2022 Jan 26];14:227-31. Available from: https://www.onlinejets.org/text.asp?2021/14/4/227/333693
| Introduction|| |
Early in the COVID-19 pandemic, researchers and frontline workers began to notice that an excessive percentage of patients were failing to recover despite the application of standard protocols. Moreover, it was difficult to predict who would and would not respond to these evidence-based treatment protocols. An examination of the morbidity and mortality data from the CDC and WHO revealed this was a universal problem. COVID-19 patients frequently did not respond to standard treatment modalities in expected ways. While a treatment protocol may work in a statistically significant percentage of the studied population, we cannot always extrapolate that to the individual patient. In addition, the pathophysiology associated with infection with a novel pathogen that causes sepsis or acute respiratory distress syndrome (ARDS) can vary widely, and sometimes, these physiologic differences are vitally important. For these reasons, it was our perception that, given the introduction of a massive number of patients infected with a novel pathogen into the mix of the group of patient in the ICU with ARDS, the problem causing the excess mortality could be protocol failure due to amplification of the heterogeneity of treatment effects (HTE) wherein the mix of patients for which the protocols had been tested was no longer representative of the new mix of patients under care during the pandemic. These themes and further literature review through Google Scholar and PubMed as well as expert discussion informed the following review.
| Lessons Learned from COVID: One Size Does Not Fit All|| |
The severe acute respiratory syndrome-2 coronavirus (SARS-CoV-2) pandemic has dramatically altered nearly all human activity. The lessons learned will resonate throughout our science, and our lives as the Earth's populace comes to grips with existential biological threats that may only be a variant or an airplane trip away.
The strategic ways clinicians evaluate information and misinformation must now be reexamined, as existing data show a slow learning curve during this pandemic relative to the rate at which lives were being lost. This exposes a number of important blind spots in our general approach to medical science and emerging threats. Perhaps the most important of these lessons is: We must have routine protocol failure mode detection integrated into the protocols themselves. It is not enough to identify that a high mortality rate has occurred over many months before adjusting or officially relaxing protocols, the system must be able to detect protocol failure in real time and adapt accordingly.
The second lesson is that when a new disease is new, its behavior is not reliably predicted by the behavior of similar past diseases. The assumption that COVID pneumonia could be treated as if it was ARDS because it met the 2012 consensus Berlin Criteria comprises an example of institutionalized anchor bias. Many institutions have clinical protocols designed specifically to manage ARDS, and those protocols are embedded in the EMR and thus the ecosystem. As a result, it was relatively easy to anchor and apply the existing ARDS protocols to COVID-19. After all, adherence to clinical protocols may be linked with improved care quality and better patient care. However, biology is unpredictable, and COVID-19 was a remarkably novel disease. Consequently, even the most experienced intensivists were challenged, humbled, and exhausted with long hours, surges in volumes, and competing recommendations from experts and the media who tried to lead despite the lack of evidence upon which to base their opinions.,,,,,
The third lesson is that, during the emergence of a new disease, it's beneficial for the expert scientists in the field to be humble and say, “We don't know,” because none of us do entirely. No one is an expert in the management of a new disease as it unfolds. Existing protocols might fail in a catastrophic way when applied to a new disease, and there is potential for great harm if protocol failure detection is not integrated into the system.
For the sake of discussion, let us emphasize that new organisms generate new stressors in our ecosystem, and individuals generate unique responses to each new stressor. Individuals exhibit their unique set of proteins as their external and internal self because of genetics and behavior that results in the biology of human life. Consequently, the individual recovery of each patient must become our primary focus. Learning how best to care for COVID-19 patients, their respiratory needs, the inflammatory storm, the association with coagulopathy, and organ failure, took a significant amount of time and effort to decipher. An important follow-up question emerges, “What about the next time?”
| The Emergence of Counter Arguments Against Standard ARDS Protocol for COVID-19 Pneumonia|| |
It has become increasing clear that some of the 20th century syndromes, such as Sepsis and ARDS, are actually comprised of a remarkably heterogeneous mix of very different diseases - subphenotypes - which fall within the criteria of the syndrome.,,, The idea that ARDS is a cognitive construct rather than a disease in and of itself has been well understood and studied since the 1980s. However, more recent studies involving lung histology and response to treatment have clearly elucidated that the effect of treatment and the prognosis are affected by the presence of these sub-phenotypes within the ARDS umbrella., While the histopathology of ARDS may vary, its clinical manifestations are perceived to be the final common pathway for many disorders, including sepsis, viral infections, trauma, and pancreatitis. While the Berlin Criteria provides a framework in which the majority of ARDS cases can be diagnosed and managed, it is commonly understood that there is a significant subset of patients who present with the clinical finding of ARDS and will not respond well to early intubation and the lung protective ventilatory techniques proposed by the Berlin consensus., It should be no surprise then that the development of lung disease in COVID-19 patients also presents with histologic as well as phenotypic (clinical) heterogeneity. However, the magnitude of the difference between many cases of COVID-19 pneumonia and the “typical” case of ARDS was unexpected and exposes the weakness of combining these subphenotypes under a common name and protocol.
Since the early 2000s, the advent of ventilation with lower tidal volumes for critically ill patients brought about a paradigm shift in the management of acute lung injury and ARDS. Pooled review of both initial and subsequent trial data demonstrated that this novel approach reduced 28-day mortality by 26%, with a trend toward reducing long-term mortality. Consequently, early intubation and the lung-protective ventilation approach is considered the best care for ARDS. This observation of improved survival represents the average treatment effect for the mix of diseases under test in the randomized controlled trial (RCT). In fact, the mortality benefit for different entities under the ARDS umbrella may fluctuate significantly, both above and below the 26% average., This baseline (and the associated protocol failure rate) can markedly change if the disease population mix changes. Within a few months of the onset of the COVID-19 pandemic, an outcry on social media and a cluster of articles began to appear which questioned the application of the standard early intubation as well as the standard P/F based PEEP titration., In response, other articles defended the standard approach.,,,
Without taking sides, given the lack of data, it is important to call out potential areas of protocol failure which may pose particular risk, if ineffective in COVID-19. The first area of possible protocol failure to consider is early intubation. This is the hallmark of the ARDS protocol. Mechanical ventilation is recommended for even moderate ARDS as defined by P/F. It is not clear that intubation based on ARDS severity determination using P/F is indicated in COVID-19, and there is potential for harm. A second area of potential failure is the titration of PEEP-based solely on P/F tables. This was always a compromise as a low PaO2 may be caused by many factors unaffected, or even made worse, by upward titration of PEEP. Upward titration of PEEP is based on P/F with the perceived goal of recruiting collapsed lung units due to worsening compliance and mitigation of hypoxemia. However, if the hypoxia was not caused by de-recruitment, increasing PEEP may not result in improvement in oxygenation. A clinician following the guidelines may progressively increase the PEEP in a potentially dangerous titration cycle. Along with the known complications of high PEEP, its induction of humoral release of Interleukin-1 and other inflammatory mediators may be of particular detriment in COVID-19 where the hyper-reactivity of the immune system is a large component of the pathology.
Given the limitations of the extant evidence, optimization of individualizing care requires frequent reassessment in real time, both at the bedside and at the systemic level. All must be open to recognizing our mistakes, accepting feedback, growing, and improving our methods or strategy. Disciplined and formalized introspection is the key to overcoming anchor bias at both the individual and institutional levels.
| Consensus Syndromes and Amplification of the Heterogeneity of Treatment Effects|| |
Syndromes are often cognitive constructs of administrative origin defined by consensus and comprised of a mix of disease with similar characteristics. Humans tend to operate by creating cognitive constructs to improve the efficiency of thought and operational capacity. This has been called object thinking. In this regard, many believe that labeling COVID lung disease as ARDS could be detrimental in that it allows us to forget that it is a separate entity with unique pathophysiology. “Definitions beget a sense of finality (often unjustified) and can confine the mind rather than liberate it.”
However, for clinicians, object thinking can be a pitfall. Clinicians cannot simply think in object terms and conflate a new disease with an old syndrome for treatment application just because the new disease may meet the criteria of a prior cognitive construct. When a new disease emerges and is included within a syndrome, only anchor bias would lead one to believe that the treatment which was effective for the original mix studied in a RCT will be reliably effective for the new disease.
Although clinicians and researchers may initially recognize the administrative origin of a broadly defined cognitive construct and therefore perceive its limitations, over time a specific syndrome gains status as a real pathophysiologic entity., When a RCT is used to test the efficacy of a protocol applied to a syndrome comprised of a mix of diseases (or sub-phenotypes), the result only provides evidence of the average treatment effect on the mix under test as a whole, not whether the treatment used in the RCT will be beneficial or harmful for any particular disease or subphenotype within the mix. In this regard, a problem posed by the construction of syndromes, which may not have been evident to those selecting the criteria, is that if the criteria are broad enough to include markedly different diseases, an RCT applied in the study of treatment of the syndrome may suffer from amplification of the HTEs. This is a fragile state, highly dependent on the mix of the diseases present. In 2020, the pandemic changed that mix of ARDS cases in a dramatic way and exposed the weakness of generalization of RCT results to disease populations which were not sufficiently represented in the RCT and worse, were not represented at all.
| The Missing Component of Protocol Application|| |
Traditional consensus protocols applied to syndromes such as sepsis and ARDS may have three major components:
- The detection component (usually comprised of a set of clinical criteria)
- The treatment titration component
- The compliance testing component.
If the treatment protocol is the right treatment for the instant disease under care and the best treatment can be delivered based on the evidence, then these three components, detection, treatment titration, and compliance testing, would seem to be enough. However, we have already discussed because of the potential of significant HTE, we cannot be sure this is the right treatment for an individual patient, so there is something missing: Protocol failure detection.
| Integrated Protocol Failure Detection|| |
Mortality over the first 6 months of the COVID-19 pandemic greatly exceeded expected mortality from ARDS. This triggered the reassessment of the treatment of COVID-19 as ARDS. However, why did this take so long to fully recognize? The answer may be due to the fact that the fourth (missing) component, protocol failure detection, was not a standard part of applied protocols.
The existence of HTE predicts the presence of a baseline number of patients for which the protocol is not the right treatment. These patients may exhibit protocol failure. Protocol failure due to unknown misalignment of the protocol and the clinical condition is a potentially reversible cause of death and therefore must be differentiated from those cases which failed to recover despite the application of a protocol which was properly aligned with their condition.
Routine, early, and reliable recognition that a protocol is not working is possible only if the protocol includes integrated failure detection and reporting as a fourth component of the protocol. In contrast, the assumption that only three components are needed is based on the idea that that the ARDS protocol was evidence-based and therefore the best treatment which could be applied. Under that approach, there is no need to auto-detect protocol failure because failure is not perceived as protocol failure but rather failure of the patient to recover despite the best treatment. The separation of these is two very different states is pivotal to optimize survival at the bedside and global level.
| Developing Real-time Protocol Failure Detection|| |
Protocol failure during the management of syndromes such as ARDS or sepsis is a complex event. For the purpose of bedside protocol failure detection, the acute timeline of a syndrome can be divided into phases of syndrome onset, worsening, and recovery. For each syndrome (or sub phenotype within a syndrome), worsening is associated with a grouping of perturbations directed away from a baseline range. When the protocol is effective, these perturbations will be interrupted by recovery vectors directed back toward the baseline range within a specified range of time intervals. The first step in the development of integrated protocol failure detection is to study and define the range of perturbations, the range of recoveries, and the associated time intervals in cases wherein the protocols were effective and in cases wherein the protocols were not effective. Failure of a given perturbation or set of perturbations to timely recover comprises a protocol failure mode.
At the department and institution level, the protocol failure modes and mortality are tracked and reported separately and contrasted. For example, with COVID-19, excess incidence of protocol failure, and in particular failure of recovery from perturbations of P/F within the expected range of time intervals, would be routinely detected along with the excess mortality. This could be contrasted with the incidence of protocol failure associated with other failure modes.
At the systemic level, protocol failure modes should be studied and monitored in large prospective observational trials. Using these data, expert consensus can then be defined, which identify the temporal components and features of protocol failure modes for each syndrome or sub phenotype and for the parameters defining the detection of the protocol failure modes. New COVID-19 variants may be associated with new protocol failure modes. The identification of excess incidence of protocol failure can provide early warnings to guide expert consensus that responds to a rapidly evolving threat. Upon the detection of a given protocol failure mode, alternative treatment or testing may be rapidly studied and if found effective, specified in a secondary protocol responsive to the detection of that failure mode.
Adding failure detection into the protocol itself provides two main benefits: It requires the developers of the protocol to study recovery vectors, identify the modes of failure, and the means to detect those failures, and it gives scientists (clinicians) a natural checkpoint at which to pause and if needed, to adjust their treatment strategy.
| Summary|| |
Throughout the history, pandemics periodically (and powerfully) shaped our culture, influenced our behaviors, and changed what we value. The 2020 pandemic will take the long view to fully comprehend. Data are still evolving, yet the observed variations in outcomes between countries, regions, and various populations are becoming increasingly apparent. There are well-established steps in a root cause analysis of any pandemic-related data, but that takes too long and provides no real-time feedback. Recovery failure mode analysis must be integrated into the protocols so that protocol failure is routinely identified.
Challenging dogma is the hallmark of science. Parsing out the nuances of RCTs, protocol failures, limitations, bias, and human capacity are difficult but important tasks that can lead to insight. Deep introspection, bold and open discussion about what happened in 2020 is required, especially on the front lines battling COVID. Perhaps the most important lesson was that consensus criteria and RCTs are not clairvoyant. It is a pitfall to assume that previously developed protocols will be effective for novel diseases. Specifically, we should consider existing protocols in the same way, we approach pharmaceuticals: They cannot be dogmatically applied as a treatment for a new emergent disease. Perhaps the most important lesson is that any application of an existing protocol requires detecting the presence of protocol or recovery failure with a recommendation for the clinician to consider adjusting care of the individual in response to that detection and if possible, as part of a clinical trial. Initiating treatment with the protocols most likely to be effective based on similarity with past diseases is rational but cannot be performed without integrated recovery failure mode detection and analysis. The pandemic exposed the need for the incorporation of real-time protocol failure mode detection into protocols, and that lesson will likely provide benefits long after the pandemic has passed.
Financial support and sponsorship
Conflicts of interest
There are no conflicts of interest.
| References|| |
Chauhan V, Galwankar SC, Yellapu V, Perez-Figueroa IJ, Stawicki SP. State of the globe: The trials and tribulations of the COVID-19 pandemic: Separated but together, telemedicine revolution, frontline struggle against “silent hypoxia,” the relentless search for novel therapeutics and vaccines, and the daunting prospect of “COVIFLU”. J Glob Infect Dis 2020;12:39-43.
Stawicki SP, Jeanmonod R, Miller AC, Paladino L, Gaieski DF, Yaffee AQ, et al.
The 2019-2020 novel coronavirus (severe acute respiratory syndrome coronavirus 2) pandemic: A joint American college of academic international medicine-world academic council of emergency medicine multidisciplinary COVID-19 working group consensus paper. J Glob Infect Dis 2020;12:47-93.
Papadimos TJ, Soghoian SE, Nanayakkara P, Singh S, Miller AC, Saddikuti V, et al.
COVID-19 blind spots: A consensus statement on the importance of competent political leadership and the need for public health cognizance. J Glob Infect Dis 2020;12:167-90.
Sheikh S, Baig MA. Silent hypoxia in COVID-19: What is old is new again! J Coll Physicians Surg Pak 2020;30:70-1.
Gattinoni L, Coppola S, Cressoni M, Busana M, Rossi S, Chiumello D. COVID-19 does not lead to a “typical” acute respiratory distress syndrome. Am J Respir Crit Care Med 2020;201:1299-300.
Galwankar SC, Paladino L, Gaieski DF, Nanayakkara KD, Somma SD, Grover J, et al.
Management algorithm for subclinical hypoxemia in coronavirus disease-2019 Patients: Intercepting the “silent killer”. J Emerg Trauma Shock 2020;13:110-3. [Full text]
Sinha S, Sardesai I, Galwankar SC, Nanayakkara PW, Narasimhan DR, Grover J, et al.
Optimizing respiratory care in coronavirus disease-2019: A comprehensive, protocolized, evidence-based, algorithmic approach. Int J Crit Illn Inj Sci 2020;10:56-63. [Full text]
Janzwood, Scott and Michelle Lee. Behind the mask: Anti-mask and pro-mask attitudes in North America. Inter-Systemic Cascades Brief #6 v.3, 2020. Cascade Institute: pp. 1-19.
Leyfman Y, Erick TK, Reddy SS, Galwankar S, Nanayakkara PW, Di Somma S, et al.
Potential immunotherapeutic targets for hypoxia due to COVI-Flu. Shock 2020;54:438-50.
Calfee CS, Delucchi K, Parsons PE, Thompson BT, Ware LB, Matthay MA, et al
. Latent class analysis of ARDS subphenotypes: Analysis of data from two randomized controlled trials. Lancet Respir Med 2014;2:611-20.
Famous KR, Delucchi K, Ware LB, Kangelaris KN, Liu KD, Thompson BT, et al.
Acute Respiratory distress syndrome subphenotypes respond differently to randomized fluid management strategy. Am J Respir Crit Care Med 2017;195:331-8.
Bos LD, Schouten LR, van Vught LA, Wiewel MA, Ong DS, Cremer O, et al.
Identification and validation of distinct biological phenotypes in patients with acute respiratory distress syndrome by cluster analysis. Thorax 2017;72:876-83.
Delucchi K, Famous KR, Ware LB, Parsons PE, Thompson BT, Calfee CS, et al.
Stability of ARDS subphenotypes over time in two randomised controlled trials. Thorax 2018;73:439-45.
Rice TW, Janz DR. In defense of evidence-based medicine for the treatment of COVID-19 ARDS. Ann Am Thorac Soc 2020;17:787-9.
Robba C, Battaglini D, Ball L, Patroniti N, Loconte M, Brunetti I, et al.
Distinct phenotypes require distinct respiratory management strategies in severe COVID-19. Respir Physiol Neurobiol 2020;279:103455.
Aublanc M, Perinel S, Guérin C. Acute respiratory distress syndrome mimics: The role of lung biopsy. Curr Opin Crit Care 2017;23:24-9.
Ferguson ND, Fan E, Camporota L, Antonelli M, Anzueto A, Beale R, et al.
The Berlin definition of ARDS: An expanded rationale, justification, and supplementary material. Intensive Care Med 2012;38:1573-82.
Petrucci N, Iacovelli W. Ventilation with lower tidal volumes versus traditional tidal volumes in adults for acute lung injury and acute respiratory distress syndrome. Cochrane Database Syst Rev. 2004;(2):CD003844. doi: 10.1002/14651858.CD003844.pub2. Update in: Cochrane Database Syst Rev. 2007;(3):CD003844. PMID: 15106222.
Deans KJ, Minneci PC, Cui X, Banks SM, Natanson C, Eichacker PQ. Mechanical ventilation in ARDS: One size does not fit all. Crit care Med 2005;33:1141-3.
Epelbaum O, Galperin I. In defence of extrapolation but not improvisation in SARS-CoV-2 lung disease. Breathe (Sheff). 2020 Jun;16(2):200113. doi: 10.1183/20734735.0113-2020. PMID: 33304409; PMCID: PMC7714550.
Roesthuis L, van den Berg M, van der Hoeven H. Advanced respiratory monitoring in COVID-19 patients: Use less PEEP! Crit Care 2020;24:1-4.
Tsolaki V, Siempos I, Magira E, Kokkoris S, Zakynthinos GE, Zakynthinos S. PEEP levels in COVID-19 pneumonia. Crit Care 2020;24:1-2.
Perchiazzi G, Pellegrini M, Chiodaroli E, Urits I, Kaye AD, Viswanath O, et al.
The use of positive end expiratory pressure in patients affected by COVID-19: Time to reconsider the relation between morphology and physiology. Best Pract Res Clin Anaesthesiol 2020;34:561-7.
Haudebourg AF, Perier F, Tuffet S, de Prost N, Razazi K, Mekontso Dessap A, et al.
Respiratory mechanics of COVID-19- versus Non-COVID-19-associated acute respiratory distress syndrome. Am J Respir Crit Care Med 2020;202:287-90.
Chen L, Xia HF, Shang Y, Yao SL. Molecular mechanisms of ventilator-induced lung injury. Chin Med J (Engl) 2018;131:1225-31.
Tobin MJ. Does making a diagnosis of ARDS in patients with coronavirus disease 2019 matter? Chest 2020;158:2275-7.
Canna SW, Behrens EM. Not all hemophagocytes are created equally: Appreciating the heterogeneity of the hemophagocytic syndromes. Curr Opin Rheumatol 2012;24:113.
Lynn LA. Artificial intelligence systems for complex decision-making in acute care medicine: A review. Patient Saf Surg 2019;13:6.
Kent DM, Steyerberg E, van Klaveren D. Personalized evidence based medicine: Predictive approaches to heterogeneous treatment effects. BMJ 2018;363:k4245.
Dr. Emily Wheeler
Department of Emergency Medicine, Florida State University, Sarasota Memorial Hospital, Emergency Medicine Residency Program, 1700 South Tamiami Trail, Sarasota, Florida 34239
Source of Support: None, Conflict of Interest: None
| Article Access Statistics|
| Viewed||428 |
| Printed||8 |
| Emailed||0 |
| PDF Downloaded||141 |
| Comments ||[Add] |