On August 19, 2020, Justice Carole J. Brown of the Ontario Superior Court of Justice released her judgment in Hacopian-Armen v Mahmoud, 2020 ONSC 4946. The plaintiffs were the estate of Armineh Hacopian-Armen, who had died of metastatic uterine leiomyosarcoma in August 2011 at the age of 49, her brother Armen Hacopian-Armen as litigation administrator, and the estate of their mother Vrijouhi Casper, who had relied on Armineh as her caregiver. The defendant was the obstetrician-gynaecologist to whom Ms. Hacopian-Armen had been referred in May 2009 for the management of long-standing heavy menstrual bleeding. Justice Brown found that the defendant had breached the standard of care by failing to perform an endometrial biopsy at the first consultation on May 25, 2009, that the breach more likely than not deprived the patient of an effective treatment for her cancer at a stage when she could have been cured, and that the breach caused her death. The damages had been agreed between the parties before trial. The decision was upheld on appeal in Hacopian-Armen v Mahmoud, 2021 ONCA 545, which is treated in a separate post on this site.
The case is doctrinally rich. It contains a careful application of the standard of care for gynecological specialists, a sustained engagement with the use of statistical and epidemiological evidence in malpractice causation, a finding on the inadequacy of the defendant’s clinical documentation as a separate breach of the standard, a credibility analysis that preferred the plaintiff’s experts over the defence’s, and an application of the doctrine that a defendant cannot rely on an evidentiary gap that the defendant’s own negligence created. Paul Cahill represented the family at trial.
The clinical context
The clinical question at the centre of Hacopian-Armen v Mahmoud was the appropriate workup for a 40-something woman presenting with abnormal uterine bleeding. The differential diagnosis for abnormal uterine bleeding is broad. It includes the benign and common (fibroids, polyps, hormonal imbalance, anticoagulation effects) and the rare but consequential (endometrial hyperplasia, endometrial cancer, and uterine sarcoma, of which leiomyosarcoma is the most aggressive subtype). The diagnostic workup is designed to discriminate between these possibilities before deciding on a treatment, because the treatment for a benign cause (medical management, embolization, ablation, or hysterectomy as a quality-of-life choice) is fundamentally different from the treatment for a malignancy (immediate hysterectomy with oncological staging and adjuvant therapy as indicated).
The endometrial biopsy is a simple, brief, office-based procedure in which a thin flexible sampler (a Pipelle catheter) is inserted through the cervix and used to aspirate a sample of the endometrium for histopathology. It can be performed during the same office visit as a Pap smear, requires no special equipment beyond what is already in any gynecologist’s office, and produces a diagnostic tissue sample within days. The published Canadian guidance, the Society of Obstetricians and Gynaecologists of Canada (SOGC) Guidelines on Abnormal Uterine Bleeding, identifies a constellation of clinical indications for the test, including age over 40, abnormal bleeding patterns, and certain risk factors. Where the indications are met, the standard of care expects the biopsy to be performed at the time of presentation.
Uterine leiomyosarcoma (LMS) is a rare and aggressive smooth-muscle cancer of the uterus. It accounts for approximately one percent of uterine cancers. Most cases arise de novo (the published evidence suggests 95% of cases) rather than from a pre-existing fibroid. The disease is staged according to the same FIGO framework as other uterine cancers. At Stage I, the tumour is confined to the uterus; treatment by total abdominal hysterectomy can achieve a five-year survival of 60 to 75 percent, which the gynecological oncology literature treats as effectively a cure for most patients (since recurrences typically occur within one to two years of treatment). At Stage IV, the tumour has metastasized to distant sites such as the lungs; treatment is palliative and the five-year survival is poor.
The patient and the long delay
Armineh Hacopian-Armen was 47 years old in May 2009. She had a long history of fibroids dating to 1999. She had developed heavy, prolonged menstrual bleeding in 2004, five years after the fibroids were first identified. By 2009 she was bleeding for approximately two weeks of each month. She was nulliparous (she had no children). Her family doctor referred her to the defendant gynecologist for management of fibroids.
At the first consultation on May 25, 2009, the defendant performed a Pap smear and a vaginal swab. He inserted a speculum, which is the same instrument used to perform an endometrial biopsy. He did not perform a biopsy. He attributed the bleeding to the fibroids that had been documented since 1999. He offered the patient a choice of treatment options: a uterine fibroid embolization to shrink the fibroids, a myomectomy, or a hysterectomy. The patient chose to think about the options and to consider the embolization. There was no further investigation of whether the bleeding might have a cause other than the fibroids.
What followed was a two-year period of management that focused on the fibroids and the embolization while the underlying disease progressed. In August 2010, the patient was taken to the emergency department of North York General Hospital, weak and palpitating, and required a blood transfusion for severe anemia from her bleeding. She was diagnosed with deep vein thrombosis. A chest CT angiogram performed during a subsequent emergency department visit identified pulmonary nodules. The embolization was eventually performed at the end of September 2010. The pulmonary nodules were not investigated further at that stage. The patient was diagnosed with a second deep vein thrombosis in February 2011, and a CT identified two new lung nodules described as worrisome for metastases. A chest CT confirmed metastatic disease. A biopsy of the lung nodules in March 2011 returned positive for sarcoma. On April 7, 2011, the defendant finally performed the endometrial biopsy. It returned positive for poorly differentiated high-grade sarcoma, favouring leiomyosarcoma. By then the disease was Stage IV. The patient underwent a hysterectomy at Sunnybrook in May 2011 and started chemotherapy. She died on August 24, 2011.
The breach finding
Justice Brown found that the defendant’s conduct on May 25, 2009 fell below the standard of care of a reasonable and prudent gynecologist. Her reasoning rested on several points.
First, the SOGC Guidelines on Abnormal Uterine Bleeding identified the patient’s clinical presentation (age over 40, abnormal bleeding patterns, nulliparity) as indications for endometrial biopsy. The plaintiff’s two main standard-of-care experts, Andrew Browning and Allan Covens, both testified that the guidelines applied. The defendant’s position that the bleeding was not abnormal because it occurred regularly (notwithstanding that it lasted two weeks per cycle and was severe enough to produce a hemoglobin drop and require transfusion within a year) was, on the evidence, unsustainable.
Second, the defendant himself had acknowledged that he could easily have performed the biopsy at the first consult. The speculum was already in place for the Pap smear. The defendant testified that the additional step would have been straightforward. He had declined to do it because, on his clinical judgment, he saw no concern of malignancy. Justice Brown found that this judgment, on the constellation of facts presented, fell outside the range of conduct that a reasonable specialist would have exercised in the same circumstances.
Third, the defendant’s clinical documentation was a separate breach. Justice Brown found that the defendant’s notes were “personal sketchy notes” rather than the kind of consultation reports that a reasonable specialist would produce. They lacked detail, lacked specificity, did not record the history of the present illness in sufficient form, and did not communicate findings and plans to the referring family physician. As Justice Brown put it, the documentation fell below the standard expected of a skilled professional in the field. This finding is doctrinally significant because it identifies clinical documentation as a free-standing component of the standard of care, not merely as evidentiary support for some other clinical breach.
The causation analysis
The harder question, on which the trial focused most of its evidentiary weight, was whether the failure to biopsy in May 2009 caused the patient’s death. The plaintiff had to establish on a balance of probabilities that the biopsy, if performed, would more likely than not have detected a malignancy, that the malignancy would have been treatable at that earlier stage, and that the patient’s outcome would have been substantially better. The defence position attacked each of these links and rested heavily on epidemiological evidence about the sensitivity of endometrial biopsy for detecting LMS in pre-menopausal women.
Justice Brown addressed the causation question by reference to the framework in Clements v Clements, 2012 SCC 32, on the “but for” test, together with the Supreme Court’s earlier guidance in Snell v Farrell, [1990] 2 SCR 311, and Laferrière v Lawson, [1991] 1 SCR 541, on the use of scientific and statistical evidence. Several doctrinal points carried the analysis.
The court is not bound by statistical evidence. As Justice Brown noted, the Supreme Court in Laferrière v Lawson established that the court is not paralyzed by statistical abstraction. Scientific findings are not identical to legal findings, and proof of causation on a balance of probabilities must be made on all the evidence, factual and statistical, that the judge is entitled to consider. Where the defendant’s statistical evidence about the sensitivity of endometrial biopsy in pre-menopausal women rested on study populations that did not match the patient’s specific clinical context (age 47, nulliparous, with rising follicle-stimulating hormone levels suggesting she was moving toward perimenopause), the statistical evidence could not displace the plaintiff’s case.
LMS does not develop overnight. Both the plaintiff’s experts opined that the Stage IV LMS diagnosed in April 2011 must have been at an earlier stage on May 25, 2009. The patient had developed heavy bleeding in 2004, five years after the fibroids were first identified, suggesting that some additional process had supervened. The pattern of slow metastatic progression observed between 2010 (when lung nodules were first noted but not investigated) and 2011 supported a relatively indolent growth pattern from a Stage I tumour in May 2009 to a Stage IV tumour in April 2011.
The defence experts agreed in cross-examination that earlier diagnosis would have improved the outcome. Two of the defence experts, Doctors Vilos and Dodge, conceded on cross that the prognosis at Stage I is substantially better than at Stage IV. Dr. Leyland declined to opine on the point. The credibility analysis (discussed below) supported the plaintiff’s position on this question.
A defendant cannot rely on an evidentiary gap that the defendant’s own negligence created. This is the doctrinally important point of the case for general practice. Justice Brown drew on Goodwin v Olupona, 2013 ONCA 259, and Ghiassi v Singh, 2018 ONCA 764, for the proposition that where a defendant’s negligence is the reason the relevant evidence does not exist, the defendant cannot use the absence of that evidence to defeat the plaintiff’s case. The defence position that “there is no evidence of LMS on May 25, 2009” was, as Justice Brown observed, a direct consequence of the defendant’s failure to perform the biopsy that would have produced the evidence. The defendant could not use his own negligence as a shield.
Justice Brown concluded that the plaintiff had established factual and legal causation on a balance of probabilities. The breach (failure to biopsy) more likely than not produced a delay in diagnosis from May 2009 to April 2011, during which the disease progressed from a curable Stage I to an incurable Stage IV.
The credibility analysis
The case turned in significant part on the trial judge’s assessment of the competing experts. Justice Brown made findings on credibility that are worth noting because they reflect a recognizable pattern in trial-level malpractice litigation.
The plaintiff’s experts (Browning, Covens, and Cheung) were described as forthright, impartial, consistent in chief and on cross-examination, and credible. Justice Brown took particular note of the fact that Dr. Browning had done most of his medical-legal work for the Canadian Medical Protective Association (which represents physician defendants), making him an unusual choice as a plaintiff’s expert and supporting the inference that his opinions were not partisan.
The defence experts (Vilos, Leyland, and Dodge) were described as knowledgeable but inconsistent in their evidence, contradictory in cross-examination relative to their direct examination, less than forthright, and tending to argue or “joust” with plaintiff’s counsel. Justice Brown found that the defence experts had a greater tendency to function as advocates for the defence. Where the testimony of the two sides differed, Justice Brown preferred the plaintiff’s experts.
This kind of credibility finding is not unusual in trial-level malpractice litigation, but it is doctrinally significant for two reasons. First, the credibility finding is largely insulated from review on appeal, which gives the trial judge’s assessment durable weight in the litigation. Second, the credibility finding feeds back into the causation analysis: the trial judge’s preference for the plaintiff’s experts on the standard of care question carries through to the trial judge’s preference for the plaintiff’s experts on the staging, sensitivity, and prognosis questions that drive the causation analysis.
The verdict and the appeal
Justice Brown found that the defendant breached the standard of care by failing to perform an endometrial biopsy on May 25, 2009 and by failing to maintain adequate clinical documentation. She found that the breach caused the patient’s death. The damages had been agreed before trial. The plaintiffs’ claims were allowed.
The defendant appealed to the Ontario Court of Appeal. The Court of Appeal unanimously dismissed the appeal in Hacopian-Armen v Mahmoud, 2021 ONCA 545, affirming the trial verdict in its entirety. The appellate decision is treated in a separate post on this site.
Why this case matters
For patients experiencing abnormal uterine bleeding, the lesson of Hacopian-Armen v Mahmoud is that the workup matters. The presence of fibroids does not exhaust the possible causes of bleeding, even where fibroids have been documented for years. An age over 40, a history of nulliparity, and abnormal bleeding patterns are clinical features that the SOGC Guidelines specifically identify as indications for endometrial biopsy. A patient presenting with these features is entitled to ask whether a biopsy has been performed, and if not, why not. The answer “your bleeding is from your fibroids” is, on the published guidance, not a sufficient reason to skip the diagnostic step.
For gynecologists and gynecological surgeons, the lesson is that the SOGC Guidelines describe a standard of care that the courts will apply. The biopsy is brief, low-risk, and high-yield when its clinical indications are met. The decision to omit it because, in the clinician’s judgment, malignancy is “too remote to justify” the additional sampling is the kind of decision that may not survive appellate review if the clinical features that triggered the indications are not addressed in a documented analysis of the differential. Justice Brown’s separate finding on the inadequacy of the defendant’s notes underlines that clinical documentation is itself a component of the standard of care, not an evidentiary afterthought.
For the broader practice of medical malpractice litigation in Ontario, Hacopian-Armen v Mahmoud is a useful authority on three doctrinal points. The first is the appropriate use of statistical and epidemiological evidence in causation: the court is not bound by statistical abstraction and is entitled to weigh the patient’s individual clinical context against generalized population-level data. The second is the doctrine that a defendant cannot rely on an evidentiary gap that the defendant’s own negligence created. The third is the recognition of clinical documentation as a free-standing component of the standard of care for specialist physicians.
The case sits alongside other Ontario decisions where a delayed diagnosis of cancer produced a wrongful-death claim with substantial damages. The Court of Appeal’s affirmation has secured the doctrinal points for future use. The decision is one of the more thorough trial-level treatments of the causation framework in delayed-cancer-diagnosis litigation, and its analytical structure (sequential application of the four elements of negligence, with particular attention to the “but for” test under Clements) is a clean template for similar cases.
Decision Date: August 19, 2020
Jurisdiction: Ontario Superior Court of Justice
Trial Judge: Justice Carole J. Brown
Citation: Hacopian-Armen v Mahmoud, 2020 ONSC 4946 (CanLII)
Affirmed on appeal: Hacopian-Armen v Mahmoud, 2021 ONCA 545
Counsel for the plaintiffs: Paul J. Cahill
