Yang v Freed: Echocardiography Timing, ECMO, and the Limits of Causation in Critical Care

In the cluster of “breach without causation” cases that comes through Canadian malpractice litigation, Yang v Freed, 2024 ABKB 763, released by the Alberta Court of King’s Bench on December 20, 2024, sits as a clean illustration of one of the more difficult features of the framework. The plaintiff established a breach of the standard of care. The defendants conceded that elements of the workup could have been done differently. The trial judge agreed. But the trial judge also concluded that the breach, taken at its plaintiff-favourable highest, would not have changed the outcome. The condition that ultimately killed Ms. Liu was not present at the time the additional investigation should have been ordered. The earlier investigation would have shown the same things the contemporaneous investigation showed. The surgical intervention that the plaintiff’s theory required would not, on the available evidence, have been completed in time. The causation analysis collapsed despite the breach finding. The action was dismissed.

The case is a defendant trial win in a critical care medicine fact pattern involving extracorporeal membrane oxygenation (ECMO), severe mitral regurgitation, endocarditis, and cardiogenic shock. It is doctrinally important as a clean application of the “counterfactual would not have helped” framework, which operates as a recurring defendant-favourable position in failure-to-investigate cases. It also adds a substantive critical care medicine case to the cluster.

A brief note on terminology. The case is styled Yang v Freed. The patient is Ms. Liu. Yang is a family member who brought the action on behalf of the estate and/or as a dependent family member; the precise plaintiff identity is not relevant to the doctrinal analysis. Throughout this post I refer to the patient as Ms. Liu and to the plaintiff side as the family.

The facts

The patient. Ms. Liu was 51 years old. She presented to the emergency department of the University of Alberta Hospital in Edmonton in December 2014 with severe respiratory distress. The hospital is a major academic centre with full critical care capability including ECMO.

The clinical course. The initial assessment showed potential heart failure together with a respiratory infection. The respiratory infection was confirmed on subsequent testing to be Human Coronavirus NL63 — one of the four “common-cold” human coronaviruses that circulate in the general population and that are distinct from SARS-CoV-2 (the virus that produces COVID-19). NL63 is well-recognized as a cause of upper and lower respiratory tract infection and is particularly significant in patients with underlying cardiac compromise.

Ms. Liu’s condition deteriorated rapidly. She suffered a cardiac arrest. She required ECMO support and a tracheostomy for continued ventilation. ECMO is an advanced form of life support that uses an external device to oxygenate the blood when the heart and lungs are unable to do so. Two forms of ECMO are commonly used in critical care: veno-venous (VV-ECMO) and veno-arterial (VA-ECMO). VV-ECMO supports the lungs alone; VA-ECMO supports both the lungs and the circulation. The clinical decision about which form to use depends on whether the patient’s cardiac function is also compromised. Ms. Liu was initially placed on VV-ECMO.

There was some clinical improvement over the next several days, but the cardiac picture remained complicated. An initial transthoracic echocardiogram (TTE) showed signs of moderate mitral regurgitation (MR) — leakage of blood back through the mitral valve, between the left ventricle and the left atrium. The MR was clinically significant but not, at that point, severe.

Approximately eight days into her admission, Ms. Liu deteriorated abruptly. Further investigation revealed severe mitral regurgitation that had developed since the initial TTE. She required cardiac surgical intervention for the mitral valve issue. She died during the procedure. The final diagnoses listed were endocarditis (infection of the heart valves), cardiogenic shock, and pneumonia.

The action. The family brought a malpractice action against two of the critical care medicine specialists who had been involved in Ms. Liu’s care during the early phase of her ECMO support. The action alleged that these physicians had failed to maintain an adequate differential diagnosis (specifically by not adequately considering a cardiac cause for the deterioration) and that they had failed to arrange timely echocardiographic investigations that would have detected the cardiac pathology in time for surgical intervention.

Damages had been settled prior to trial. The liability-only trial focused on standard of care and causation.

The legal framework — standard of care

The trial judge separated the standard of care analysis into two components.

Component 1: The differential diagnosis approach. The plaintiff alleged that the critical care specialists had failed to maintain an appropriate differential diagnosis and had inappropriately fixed on a viral respiratory illness as the primary explanation. The trial judge rejected this allegation. The findings, distilled:

• The physicians had appropriately prioritized the life-threatening issues: the neurological concerns and the acute respiratory distress syndrome (ARDS)
• Their working diagnosis (viral respiratory illness with secondary cardiac compromise) was supported by multiple factors and was based on sound judgment, care, and skill
• The treatment plan demonstrated a clear understanding of the need to prioritize life-threatening issues
• The differential remained open and the physicians continued to consider the cardiac dimension

The trial judge found the standard of care met on this component.

Component 2: The timing of follow-up echocardiography. The plaintiff alleged that the physicians had failed to arrange timely follow-up echocardiographic investigations. The trial judge accepted this allegation. The findings, distilled:

• After the initial TTE showed moderate mitral regurgitation, a follow-up TTE or TEE should have been ordered relatively quickly
• The follow-up was particularly indicated once Ms. Liu had been placed on ECMO, which complicates the cardiac picture and can itself produce or unmask valvular pathology
• The clinical practice for a diligent intensivist would have been to conduct a diagnostic transesophageal echocardiogram (TEE) while the patient was on VV-ECMO, and to follow this with a repeat TTE within a 24-hour period
• The defendant physicians’ explanations for the delay (concerns about the risks of TEE in the ECMO patient and the potential for poor imaging) were not sufficient to displace the standard

The trial judge accordingly found the standard of care breached on this component.

The split outcome on standard of care — major component met, secondary component breached — is itself an analytically useful feature of the case. It illustrates that standard of care findings are not all-or-nothing. The trier of fact assesses each distinct alleged breach against the relevant standard. Some allegations succeed; others do not. The overall outcome of the case depends on whether the breach that is established also satisfies the causation analysis.

The legal framework — causation

The causation analysis is where the case ultimately turned. The trial judge worked through three distinct propositions.

Proposition 1: The acute mitral regurgitation that killed Ms. Liu was not present at the time of the alleged breach. The trial judge found that the severe acute MR developed during the eight-day period of the ECMO support. The proposed follow-up echocardiogram (which should have been performed shortly after the initial TTE) would have shown what the initial TTE showed — moderate MR. It would not have shown severe MR because severe MR did not yet exist. The earlier investigation would not have detected the catastrophic pathology because the pathology had not yet developed.

This is the most analytically important feature of the case. The breach was real (the follow-up should have been ordered) but the breach did not connect to the harm because the harm was caused by a subsequent pathology that the earlier investigation would not have caught.

Proposition 2: Even with early detection, surgical intervention would not likely have been timely. The trial judge addressed the hypothetical scenario in which the additional investigation had been performed and had produced new information. Even in that scenario, the plaintiff could not establish on the balance of probabilities that surgical intervention would have occurred in time to prevent the eventual death. The cardiac surgical pathway in an ICU patient on ECMO is complex. The decision-making involves multiple teams. The operative window is constrained by the patient’s overall clinical status. The plaintiff did not adduce evidence sufficient to establish that the surgical intervention would have been timely.

Proposition 3: The combination of these two propositions defeats the causation analysis. Even taking the breach at its plaintiff-favourable highest, the family could not establish that the breach caused Ms. Liu’s death. The “but-for” causation framework from Clements v Clements, 2012 SCC 32, requires the plaintiff to establish on the balance of probabilities that the breach made a difference to the outcome. The “robust and pragmatic” approach to causation under Snell v Farrell, [1990] 2 SCR 311, allows the trier of fact to draw inferences where the evidence supports them. Neither framework saves the family’s case here. The breach simply did not connect to the harm.

The trial judge accordingly dismissed the action.

The doctrinal anchors

Several doctrinal anchors emerge from the case.

Breach without causation as a recognized pattern. Yang v Freed joins a substantial line of cases in the cluster where the standard of care is found to have been breached (in whole or in part) but causation cannot be established. The pattern includes Williamson v Y (BC anaesthesia), Papineau v Sharma (Ontario), Lorencz v Talukdar (Saskatchewan failure-to-refer), and now Yang v Freed (Alberta critical care). The framework is now well-anchored in the cluster as a recurring defendant-favourable pattern.

The “counterfactual would not have helped” framework. Yang v Freed is the cleanest illustration of this framework in the cluster. The breach was real — the investigation should have been performed. But the investigation would not have produced different information because the pathology that ultimately killed the patient had not yet developed. The temporal mismatch between the breach and the harm broke the causal chain.

The framework is generalizable. In any failure-to-investigate case, the plaintiff has to establish not just that the investigation should have been performed but that the investigation would have produced information that would have changed the clinical course. Where the relevant pathology was not yet present at the time of the alleged breach, the investigation would not have detected it, and the causation analysis fails.

The principle parallels the Lorencz v Talukdar framework on wait times — there, the breach was real (the referral should have been made), but the referral would not have produced the necessary clinical intervention in time, and causation failed. The two cases together illustrate that the breach analysis and the causation analysis operate independently and that the second can defeat a plaintiff even where the first goes in the plaintiff’s favour.

The “surgical intervention timeline” framework. A separate doctrinal point: even where additional investigation would have produced new information, the plaintiff still has to establish that the resulting surgical intervention would have been timely. The cardiac surgical pathway in a complex critical care patient is constrained by multiple factors: the patient’s overall clinical status, the availability of operating theatres and teams, the decision-making across multiple specialties, the post-operative recovery requirements. The plaintiff who proposes that surgical intervention would have saved the patient has to engage with each of these factors in the evidence.

The framework operates in conjunction with the “counterfactual would not have helped” framework. The plaintiff must overcome both: the additional investigation would have produced relevant new information AND the resulting intervention would have been timely. Where either fails, the causation analysis collapses.

The bidirectional causation principle. The case illustrates that the “robust and pragmatic” approach to causation under Snell v Farrell is not a free pass for plaintiffs. The framework allows the trier of fact to draw permissible inferences where the evidence supports them. Where the evidence does not support the inference, the framework provides no help. The bidirectional operation of Snell is now demonstrated across multiple cluster cases:

• Hasan v Trillium Health Centre (plaintiff success — defendant’s conduct created the gap)
• Noel v Hawrylyshyn (defendant success — operative urgency provided sufficient defence)
• Lorencz v Talukdar (defendant success — wait time evidence speculative)
• Yang v Freed (defendant success — temporal mismatch between breach and harm)

The cluster now has four cases articulating the bidirectional operation of Snell. The framework is settled.

The split standard of care outcome. Yang v Freed illustrates that standard of care findings are not all-or-nothing. The plaintiff alleged two distinct categories of breach: the differential diagnosis approach and the echocardiography timing. The trial judge found the first SOC component met and the second component breached. The split outcome is doctrinally important for two reasons:

1. It demonstrates that plaintiffs can succeed on parts of their case even where the overall claim fails
2. It demonstrates that the trial judge is willing to make granular findings rather than treating the SOC analysis as a single bottom-line verdict

For plaintiff counsel, the practical implication is that pleadings and proof should be calibrated to the strongest grounds. Weak grounds that are unlikely to succeed can undermine the credibility of the stronger grounds. The defendant’s expert who is required to defend three propositions in a multi-ground case can do less work on each than would otherwise be possible.

Critical care medicine — the doctrinal context

A few clinical and doctrinal observations on the critical care medicine context that frames the case.

ECMO and the changed clinical picture. Extracorporeal membrane oxygenation is one of the most advanced forms of life support in modern critical care medicine. The technology supports patients whose cardiopulmonary function has failed beyond the point where conventional mechanical ventilation and vasopressor support can compensate. Patients on ECMO are by definition extremely sick. The clinical management is complex: anticoagulation must be carefully calibrated; bleeding risk is significant; infection risk is elevated; and the underlying cardiac and pulmonary pathology continues to evolve.

In ECMO patients, the echocardiographic picture changes. The ECMO circuit affects the loading conditions on the heart. Cardiac chamber dimensions can change. Valve function can be altered. The clinical interpretation requires expertise in both critical care echocardiography and ECMO physiology. The decision about whether to perform a TEE specifically (which requires advancing a probe down the esophagus, near the ECMO cannulation site) in an ECMO patient is itself a substantial clinical decision. The defendants’ caution about the risks of TEE in this setting was not unreasonable; the trial judge concluded only that the magnitude of those risks was not sufficient to justify delay where the clinical picture indicated the need for follow-up investigation.

TTE vs TEE. The two main forms of echocardiography in adult critical care:

• Transthoracic echocardiography (TTE) uses a probe placed on the chest wall to image the heart. Non-invasive. Acceptable image quality in many patients. Limited image quality in patients with body habitus, prior surgery, or ECMO that interferes with acoustic windows.
• Transesophageal echocardiography (TEE) uses a probe advanced through the mouth and into the esophagus to image the heart from behind. The probe sits much closer to the heart than the TTE probe. Image quality is generally substantially better, particularly for the posterior cardiac structures (including the mitral valve). The procedure is more invasive: the patient must be sedated; the probe must be placed by an experienced operator; bleeding risk and esophageal injury are recognized complications.

In an ECMO patient with suspected mitral pathology, TEE is generally the test of choice for definitive assessment of the mitral valve. TTE may not provide sufficient image quality. The defendants in Yang v Freed were aware of these considerations and were managing the trade-offs in their clinical decisions. The trial judge found that on the available evidence, a diligent intensivist would have completed the TEE within the relevant window despite the technical complexity.

Endocarditis, cardiogenic shock, and the underlying disease trajectory. The final diagnoses listed for Ms. Liu — endocarditis (infection of the heart valves), cardiogenic shock (severe failure of the heart to pump effectively), and pneumonia — together describe a clinical picture in which the underlying disease process was extremely severe. Endocarditis with mitral valve involvement is particularly difficult to manage; surgical intervention is sometimes required, but the operative mortality in this clinical context can be 30-50% or higher. The causation analysis must account for the underlying disease trajectory: even with optimal care, the prognosis in a patient with severe endocarditis, mitral regurgitation, cardiogenic shock, and respiratory failure requiring ECMO is guarded.

The framework parallels Williamson v Y, where the underlying disease severity formed part of the causation analysis. In a patient whose baseline prognosis is poor, the plaintiff has to establish that the breach changed the outcome rather than just that the outcome was bad. Yang v Freed illustrates the framework operating in the critical care medicine context.

Care responsibility transitions

A separate procedural feature of the case warrants brief discussion. The defendants argued that they were no longer the responsible physicians at the time Ms. Liu’s clinical decline made surgical intervention necessary. The point is doctrinally relevant in critical care, where patient care is typically transferred between intensivists at structured intervals (shift changes, weekly handovers, transfers between subspecialty teams).

The trial judge did not need to resolve this question definitively because the causation analysis dispatched the case. But the framework is worth noting. Where a physician’s involvement in a patient’s care ends before the subsequent deterioration that produces the harm, the analysis of that physician’s contribution to the harm has to account for the intervening period and the conduct of the subsequent providers. The framework parallels the multi-defendant analysis discussed in Communication Failures, Continuity of Care, and Medical Malpractice.

Why this case matters

For prospective clients. The case is doctrinally useful as a reminder that the standard of care analysis and the causation analysis operate independently. A finding that the physician should have done something differently does not automatically produce a successful malpractice claim. The plaintiff must also establish that doing it differently would have changed the outcome on the balance of probabilities. Where the breach is real but the counterfactual does not improve the outcome — for example, because the relevant pathology was not yet present, or because the resulting intervention would not have been timely — the case fails despite the breach.

In assessing a potential failure-to-investigate case, several practical questions are relevant:

• What was the clinical picture at the time the investigation should have been performed? If the relevant pathology was not yet present, the investigation would not have detected it, and the causation analysis is at risk.
• What would the investigation have shown? The expert evidence must establish what the additional investigation would have detected. Speculation is not enough.
• What would have followed from the new information? The plaintiff must establish not just that the information would have been obtained but that the resulting clinical pathway would have intercepted the harm in time.
• What is the underlying disease trajectory? Where the patient’s baseline prognosis is poor, the causation analysis must account for that prognosis. The plaintiff must establish that the breach changed the outcome, not just that the outcome was tragic.

For more on the general framework for evaluating these cases, see Suing for Medical Malpractice in Ontario: What You Need to Know.

For physicians and clinical teams. A few observations:

Document the rationale for clinical decisions. Where a follow-up investigation is not performed, document the reason. The trial judge in Yang v Freed considered the defendants’ explanations for not ordering the additional echocardiogram (concerns about TEE risks in ECMO patients, potential for poor TTE imaging) but concluded that the explanations did not justify the delay. The clinical rationale documented at the time would have informed the trial analysis regardless of the ultimate conclusion.

Calibrate technical complexity against clinical need. The technical risks of TEE in an ECMO patient are real but not absolute. Where the clinical picture indicates the need for definitive assessment of the cardiac anatomy, the technical complexity is a feature to be managed, not a reason to defer. Modern critical care echocardiography practice expects intensivists and consulting cardiologists to be willing to perform TEE in ECMO patients with appropriate technique.

Update the differential diagnosis as the clinical picture evolves. The trial judge found the initial differential diagnosis met the standard of care. The breach was at the follow-up stage. The principle: the initial assessment is not the end of the diagnostic process. As the clinical picture evolves, the differential must be reassessed and additional investigations must be considered.

Cluster integration

The breach-without-causation cluster:

• Williamson v Y (BC anaesthesia)
• Papineau v Sharma (Ontario)
• Lorencz v Talukdar (Saskatchewan failure-to-refer)
• Yang v Freed (Alberta critical care)

Four documented cases now anchor this sub-cluster. The pattern is consistent: SOC breached (in whole or in part) but causation defeated. The framework is well-established.

The “Snell v Farrell” symmetrical operation cluster:

• Hasan v Trillium Health Centre (plaintiff success)
• Noel v Hawrylyshyn (defendant success — operative urgency)
• Lorencz v Talukdar (defendant success — wait time evidence speculative)
• Yang v Freed (defendant success — temporal mismatch between breach and harm)

Four cases now anchor this sub-cluster. The bidirectional operation of Snell is settled.

The “counterfactual would not have helped” framework:

• Yang v Freed is the principal cluster authority on the temporal-mismatch version of this framework
• Lorencz v Talukdar is the principal cluster authority on the timeline version (wait times made the counterfactual speculative)
• Together the two cases anchor the “counterfactual reconstruction defeats plaintiff” pattern

Critical care medicine practice area (new for the cluster):

• Yang v Freed is the principal cluster authority
• Connects to the broader cluster on ICU and acute medicine cases

Cardiac and cardiology practice area:

• Lorencz v Talukdar (failure to refer to cardiology)
• Yang v Freed (critical care echocardiography in cardiac patient)
• The cluster now has substantive coverage of the cardiac diagnostic pathway from primary care through critical care

---

Decision Date: December 20, 2024

Jurisdiction: Court of King’s Bench of Alberta

Citation: Yang v Freed, 2024 ABKB 763 (CanLII)

Outcome: Action dismissed. The trial judge found that the defendant critical care medicine specialists had partially breached the standard of care by failing to arrange timely follow-up echocardiography (TTE or TEE) after the initial TTE showed moderate mitral regurgitation in a patient on VV-ECMO. However, the trial judge also found that the family had not established that this breach caused the patient’s death because (a) the acute severe mitral regurgitation that ultimately killed the patient had not developed at the time the additional investigation should have been performed; (b) the additional investigation would not have detected pathology that was not yet present; and (c) even with hypothetical early detection, the plaintiff had not established that surgical intervention would likely have occurred in time.

Key authorities (implicit in the analysis): Clements v Clements, 2012 SCC 32 (but-for causation framework); Snell v Farrell, [1990] 2 SCR 311 (robust and pragmatic causation); the established framework for standard of care in critical care medicine including expert evidence from intensivists and consulting subspecialists.