In July 2021, the Ontario Court of Appeal released its decision in Hacopian-Armen Estate v Mahmoud, 2021 ONCA 545, unanimously dismissing the defendant’s appeal from the trial judgment of Justice Carole J. Brown. Strathy C.J.O. wrote for the panel, with Feldman J.A. and Sossin J.A. concurring. The trial decision had found a Toronto-area gynecologist liable for failing to perform an endometrial biopsy at the first consultation with a 47-year-old patient who later died of metastatic uterine leiomyosarcoma; the trial post on this site describes the underlying facts, the breach finding, and the causation analysis at trial in detail. The defendant did not appeal the standard-of-care finding. The appeal was confined to causation, both factual and legal.
The Court of Appeal’s decision is doctrinally useful on three points. The first is the appellate clarification of the foreseeability analysis in delayed-cancer-diagnosis cases, applying Mustapha v Culligan of Canada Ltd., 2008 SCC 27, and Frazer v Haukioja, 2010 ONCA 249, to hold that the foreseeable harm is the kind, type, or class of injury that flows from the breach rather than the specific disease that the breach failed to detect. The second is the application of the Housen v Nikolaisen standard of review to a trial judge’s preference of one expert over another, with the Court of Appeal affirming the deference owed to first-instance credibility findings. The third is a clarification of the doctrine in Goodwin v Olupona and Ghiassi v Singh that a defendant cannot rely on an evidentiary gap that the defendant’s own negligence created. The Court of Appeal held that the principle does not apply where both parties have adduced evidence on the issue of causation, but found that the trial judge’s reference to it was unnecessary rather than determinative, because she had independently accepted the plaintiff’s expert evidence on the staging and detectability questions. The appeal was argued for the respondents by Christopher Morrison and Paul Cahill of Davidson Cahill Morrison LLP.
A brief recap
Armineh Hacopian-Armen was 47 years old in May 2009 when she was referred to the defendant gynecologist for management of long-standing heavy menstrual bleeding. The patient had a history of fibroids dating to 1999 and a history of bleeding lasting two weeks per month since 2004. At the first consultation, the defendant performed a Pap smear and vaginal swab but did not perform an endometrial biopsy, the routine office procedure that the SOGC Guidelines on Abnormal Uterine Bleeding identified as indicated for a patient with her clinical profile (age over 40, abnormal bleeding patterns, nulliparous). The defendant attributed the bleeding to the fibroids. He did not investigate other causes. Over the following two years, the patient deteriorated. In April 2011, the defendant finally performed the endometrial biopsy. By then the patient had Stage IV uterine leiomyosarcoma with metastases to her lungs. She died in August 2011.
The trial judge found that the defendant breached the standard of care by failing to perform the biopsy at the first consult and that the breach caused the patient’s death. The trial judge accepted the plaintiff’s experts’ opinion that the patient’s leiomyosarcoma had likely been at Stage I in May 2009, that an endometrial biopsy at that time would more probably than not have detected it, and that timely treatment at Stage I would have produced a five-year survival of 60 to 75 percent (effectively a cure for most patients). The trial judge found the defendant’s experts less credible than the plaintiff’s. Damages had been agreed before trial.
The appellant’s grounds
The defendant did not challenge the standard-of-care finding on appeal. The appeal focused on causation in three respects.
First, the defendant argued that the trial judge had erred in her legal causation analysis by reasoning retrospectively. The proper question, the defendant argued, was whether it was foreseeable to a reasonable gynecologist that (a) the patient had uterine leiomyosarcoma in 2009, (b) an endometrial biopsy would have detected the leiomyosarcoma, and (c) the failure to perform the biopsy could lead to delayed diagnosis of the leiomyosarcoma. Because uterine leiomyosarcoma is a rare cancer (approximately one percent of uterine cancers), the defendant argued, the specific harm was too remote and unforeseeable to support legal causation.
Second, the defendant argued that the trial judge had erred in her factual causation analysis by finding that the leiomyosarcoma was likely present in May 2009. The plaintiff’s expert opinion that the cancer had likely been at Stage I in May 2009 was, the defendant submitted, inadequately supported in the expert’s report.
Third, the defendant argued that the trial judge had erred by finding that an endometrial biopsy performed in May 2009 would have detected the leiomyosarcoma at that time, given that the cancer was not visible on ultrasound imaging in the months that followed.
The defendant also raised a discrete criticism of the trial judge’s invocation of the Goodwin v Olupona principle that a defendant cannot rely on an evidentiary gap that the defendant’s own negligence created.
The “same class or kind” foreseeability analysis
The Court of Appeal agreed with the defendant on one narrow point. The trial judge’s foreseeability analysis, as framed in her reasons, had included an element of retrospective reasoning. She had asked whether it was foreseeable that, having failed to conduct an endometrial biopsy, the presence of leiomyosarcoma (if not treated) would result in serious injury or death. That question, as the Court of Appeal observed, contained a retrospective element: it assumed that the biopsy would have detected the cancer that was, in fact, ultimately found.
The Court of Appeal held, however, that the error was harmless. The proper foreseeability framework, drawn from the Supreme Court of Canada’s decision in Mustapha v Culligan, 2008 SCC 27, asks whether the harm suffered is “of a kind, type or class” that was reasonably foreseeable as a result of the defendant’s negligence. The Court of Appeal cited Frazer v Haukioja, 2010 ONCA 249, for the application of this test in the medical context: the foreseeability inquiry is not whether the specific disease that occurred was foreseeable, but whether the kind of harm that occurred was foreseeable.
Applied to the facts of Hacopian-Armen, this framework produced a straightforward conclusion. The endometrial biopsy that the standard of care required at the first consult was a test capable of detecting endometrial and uterine cancers, of which leiomyosarcoma is one. The patient’s clinical presentation (age over 40, abnormal bleeding, nulliparity) gave rise to the indications for the test precisely because the test was capable of identifying serious uterine pathologies that would otherwise be missed. Where a gynecologist fails to conduct a test that would have detected cancers of the same class or character as uterine leiomyosarcoma (including leiomyosarcoma itself), it is foreseeable that such cancers will go undetected and that the patient will suffer the consequences of the delay. The fact that the specific cancer that materialized was rare did not narrow the foreseeability analysis to that specific cancer; the test had been indicated because cancer of some kind was a foreseeable risk.
The Court of Appeal also rejected the defendant’s floodgates argument that this analysis would force physicians to order unnecessary tests to avoid the risk of missing unforeseeable diseases. The Court of Appeal noted that the endometrial biopsy was a necessary test in the circumstances of this patient, on the standard of care that the defendant had not appealed. The foreseeability analysis did not impose any obligation to order unnecessary tests. It applied only to the harm flowing from the failure to perform a test that was clinically indicated.
Factual causation and the standard of review
On factual causation, the Court of Appeal applied the Housen v Nikolaisen, 2002 SCC 33, framework. Findings of fact, including findings of mixed fact and law, are reviewable for palpable and overriding error. A trial judge’s preference of one expert over another is a factual finding that attracts substantial deference on appeal.
The trial judge had made an express credibility finding in favour of the plaintiff’s experts (Andrew Browning and Allan Covens), describing them as forthright, impartial, and consistent in chief and on cross-examination. She had described the defendant’s experts (George Vilos, Nicholas Leyland, and Jason Dodge) as knowledgeable but inconsistent, argumentative, and tending to function as advocates for the defence. Where the testimony of the two sides differed, she had preferred the plaintiff’s experts.
The Court of Appeal found no palpable and overriding error in this assessment. The plaintiff’s expert had reasoned, through reverse chronology from the patient’s death in August 2011 back to May 2009, that the cancer that was Stage IV in April 2011 had taken sufficient time to metastasize that it would have been at Stage I in May 2009. The trial judge was entitled to accept this reasoning, and the Court of Appeal would not reweigh the evidence on appeal. The defence position that the plaintiff’s expert’s report was insufficiently detailed on the timing question did not rise to the level of palpable and overriding error; the crux of the expert’s reasoning was discernible from the full report and from his trial testimony.
On the question of whether an endometrial biopsy in May 2009 would have detected the leiomyosarcoma, the Court of Appeal similarly found no error. The plaintiff’s experts had explained that uterine leiomyosarcoma is typically small in its early stages and may not be visible on ultrasound imaging at that point, but that an endometrial biopsy is capable of detecting cellular abnormalities that imaging cannot capture. The trial judge had accepted this evidence over the defendant’s experts’ contrary view. Where two competing expert opinions are presented, a trial judge is entitled to prefer one over the other; the Court of Appeal would not interfere with that preference absent error in principle or palpable and overriding error in the assessment of the evidence.
The Goodwin v Olupona clarification
The most legally interesting aspect of the appeal, for general practice, was the Court of Appeal’s clarification of the evidentiary-gap doctrine. The trial judge had cited Goodwin v Olupona, 2013 ONCA 259, and Ghiassi v Singh, 2018 ONCA 764, for the proposition that a defendant cannot rely on an evidentiary gap that the defendant’s own negligence created. She had applied this principle to the defendant’s argument that there was no direct evidence of leiomyosarcoma in May 2009, noting that the reason for the absence of evidence was the defendant’s failure to perform the biopsy that would have produced it.
The Court of Appeal observed that the principle expressed in Goodwin and Ghiassi does not apply to a case such as this, where both parties have adduced evidence on the issue of causation. In Goodwin and Ghiassi, the defendant’s negligence had produced a complete absence of relevant evidence (in Goodwin, the failure to monitor a foetal heart rate; in Ghiassi, the failure to order a bilirubin test), with the result that there was no direct evidence at all on the causal question. In Hacopian-Armen, by contrast, both parties had adduced expert evidence on the causation question, including expert evidence drawing on the patient’s clinical history, the published medical literature, and the imaging studies that had been performed. The trial judge’s task was to weigh the competing expert evidence, not to import an inference rule that addressed an entirely different problem.
The Court of Appeal held, however, that the trial judge’s invocation of Goodwin was unnecessary rather than determinative of the outcome. The trial judge had independently accepted the plaintiff’s expert evidence that the patient’s cancer had probably been at Stage I in May 2009. The reference to Goodwin was therefore harmless on the appellate analysis: the causation finding stood on the trial judge’s primary credibility assessment, not on any application of the evidentiary-gap principle.
This clarification is doctrinally useful. The Goodwin and Ghiassi principle is a powerful tool for plaintiffs in cases where the defendant’s negligence produces a true evidentiary vacuum. Hacopian-Armen Estate v Mahmoud clarifies that the principle does not extend to cases where the parties’ competing expert evidence supplies the evidentiary record from which the causation question must be answered. Plaintiffs should not over-rely on the principle in cases where conventional expert evidence is available. The principle, properly understood, addresses a narrower problem.
Disposition
The Court of Appeal unanimously dismissed the appeal. The trial verdict was affirmed in its entirety. Damages, which had been agreed before trial, were entered against the defendant.
Why this case matters
For patients and families navigating a delayed cancer diagnosis, Hacopian-Armen Estate v Mahmoud is a useful authority on what the foreseeability analysis actually requires. The doctrine does not ask whether the specific cancer that materialized was foreseeable to the defendant clinician. The doctrine asks whether the kind of harm that resulted from the breach (here, an undetected uterine malignancy progressing to a stage where it could not be cured) was foreseeable. Cancers of the kind that endometrial biopsies are designed to detect are within the foreseeable harm of failing to perform an indicated biopsy. The fact that the specific cancer was rare does not narrow the analysis to that specific cancer.
For physicians, particularly gynecologists and other specialists who order diagnostic tests under published guidelines, the case clarifies that the foreseeability analysis in negligence does not impose any obligation to order unnecessary tests. The analysis applies only to tests that are clinically indicated under the applicable standard of care. The defence concern that affirming liability here would expand the duty to investigate is, as the Court of Appeal noted, unfounded. The duty to investigate is co-extensive with the standard of care for the specialty.
For the broader practice of medical malpractice litigation in Ontario, Hacopian-Armen Estate v Mahmoud contributes three doctrinal points worth noting. The application of Mustapha v Culligan and Frazer v Haukioja to a delayed-cancer-diagnosis case is a useful authority for plaintiffs in similar cases where the defendant tries to narrow the foreseeability analysis to the specific disease. The application of Housen v Nikolaisen to a trial judge’s preference of one expert over another is a useful authority on the limited scope of appellate review of expert credibility findings. And the clarification of the Goodwin v Olupona and Ghiassi v Singh principle helpfully cabins the evidentiary-gap doctrine to cases where it is genuinely needed.
The decision is now in its sixth year as a published authority. It has been cited in subsequent appellate decisions on the same doctrinal points (see, for example, the citation lists in Farej v Fellows, 2022 ONCA 254, and other delayed-diagnosis cases in the years since 2021). It remains the leading recent appellate authority on foreseeability analysis in failure-to-diagnose litigation.
Decision Date: July 29, 2021 (heard May 20, 2021 by video conference)
Court: Court of Appeal for Ontario
Docket: C68655
Panel: Strathy C.J.O. (writing), Feldman and Sossin JJ.A.
Citation: Hacopian-Armen Estate v Mahmoud, 2021 ONCA 545 (CanLII)
Trial Decision: Hacopian-Armen v Mahmoud, 2020 ONSC 4946 (CanLII). Trial post on this site.
Counsel for the respondents on appeal: Christopher I.R. Morrison and Paul J. Cahill
