Tripp v Ross (MBCA): Delayed Colon Cancer Diagnosis and the Benhaim Causation Framework

Of the cases I follow that bear on causation in Canadian medical malpractice, Benhaim v St-Germain, 2016 SCC 48, is the one I return to most often. It is the modern Supreme Court of Canada articulation of how causation operates in the difficult subset of malpractice cases where the defendant’s negligence itself has impaired the plaintiff’s ability to establish what would have happened with appropriate care. Benhaim sits on top of the older framework from Snell v Farrell, [1990] 2 SCR 311 (the “robust and pragmatic” approach) and Clements v Clements, 2012 SCC 32 (the “but-for” framework), and resolves a specific question that the earlier cases had left open: when a defendant’s negligence creates causal uncertainty, is the trier of fact required to draw an adverse inference of causation against the defendant?

The Supreme Court’s answer in Benhaim was no, the inference is discretionary. Where a plaintiff has led some affirmative evidence of causation and the defendant’s negligence has impaired the proof, the trier of fact may draw an adverse inference but is not obliged to. The reasoning runs through the structure of the law of negligence: a mandatory rebuttable inference would effectively shift the burden of proof onto defendant physicians and, in the Supreme Court’s phrase, would risk “turning defendant professionals into insurers.” The framework retains the orthodox onus on the plaintiff to prove causation on the balance of probabilities while allowing the trier of fact to draw permissible inferences from the available evidence.

Tripp v Ross, 2025 MBCA 25, released by the Manitoba Court of Appeal on March 21, 2025, is one of the clearest recent applications of the Benhaim framework. The case is a malpractice action arising from a colonoscopy in which the defendant gastroenterologist failed to complete the examination and missed a colon cancer that was then diagnosed eight months later. The defendant conceded the standard of care breach prior to trial. The contested question was whether the eight-month delay had caused the patient’s death from metastatic colon cancer. The trial judge held that it had not. The Court of Appeal unanimously affirmed.

The case is doctrinally important for several reasons. It is one of the cleanest recent illustrations of the post-Benhaim causation framework in a cancer-diagnosis-delay case. It addresses the limits of statistical evidence in causation analysis, including the “evidentiary bridge” requirement. It illustrates the loss-of-chance versus causation distinction where the survival probability falls below 50 percent. And it reinforces the appellate deference framework on factual findings of causation. For prospective clients considering similar cases, the framework operates as a significant practical constraint on the viability of delayed-cancer-diagnosis malpractice claims.

This post addresses the appellate decision. For the trial decision, see the paired trial commentary.

The clinical context — colon cancer and the screening colonoscopy

A brief clinical overview is useful for the legal analysis.

Colon cancer. Colorectal cancer is the second leading cause of cancer death in Canada. The disease typically begins as a benign polyp on the inner lining of the colon, which can develop dysplastic features over time and ultimately progress to invasive adenocarcinoma. The natural history is generally slow: the typical progression from a benign polyp to an invasive cancer takes years, often a decade or longer. The slow natural history is the foundation of the screening colonoscopy program: detection of polyps at the pre-cancerous stage allows removal before invasive disease develops.

The staging framework. Colon cancer is staged using the TNM (tumour, node, metastasis) system, which combines into Stage I through Stage IV:

• Stage I: cancer confined to the inner layers of the colon wall
• Stage II: cancer extending through the colon wall but not involving lymph nodes
• Stage III: cancer involving regional lymph nodes (with sub-stages IIIA, IIIB, IIIC reflecting increasing nodal involvement)
• Stage IV: cancer with distant metastases (typically to liver, lung, peritoneum, or other distant sites)

The five-year survival rate decreases with each stage. Modern estimates for colon cancer:

• Stage I: approximately 90 percent five-year survival
• Stage II: approximately 70-80 percent
• Stage III: approximately 50-70 percent depending on sub-stage
• Stage IV: approximately 14-30 percent

The numbers vary by study, time period, and treatment regimen. The general pattern is that earlier-stage disease is substantially more curable than metastatic disease.

Liver metastases. The liver is the most common site of distant metastasis from colon cancer. The portal venous drainage from the colon passes through the liver, which makes the liver the first capillary bed encountered by cells released from a colorectal tumour. Once colon cancer has metastasized to the liver, the disease is staged as Stage IV. Liver metastases can sometimes be resected with curative intent, but the prognosis remains substantially worse than non-metastatic disease.

The doubling-time framework. Cancer cells grow exponentially. The doubling time (the time for the cancer cell population to double in size) is a clinical parameter that varies by cancer type and individual case. For colorectal cancer liver metastases, the typical doubling time is in the range of weeks to months. The clinical implication: a metastatic deposit visible on imaging today has been present at a smaller size for many doubling cycles before reaching detectable size. The framework is well-established in oncology and is routinely used in expert evidence on causation questions in delayed-diagnosis cases.

The screening colonoscopy. A complete colonoscopy involves examination of the entire colon from the rectum through to the cecum, with inspection of the mucosa for polyps, areas of inflammation, or suspicious lesions. The standard of care requires the gastroenterologist to advance the colonoscope to the cecum (the “cecal intubation rate” is a recognized quality metric, with most accreditation programs requiring a rate of at least 90-95 percent), to withdraw the scope slowly while inspecting the mucosa, and to biopsy or remove any suspicious findings. The standard does not permit the procedure to be terminated mid-examination upon finding an apparent explanation for the patient’s symptoms unless the full examination is otherwise contraindicated.

The facts

The patient. Mr. Lindsey Shaun Tripp was 53 years old in December 2017 when he began experiencing abdominal cramps and rectal bleeding. He attended his family physician, who referred him to the defendant gastroenterologist for a colonoscopy.

The January 2018 colonoscopy. The defendant performed a colonoscopy in January 2018. During the procedure, he observed five polyps. Four of the polyps were biopsied; pathology subsequently confirmed that they were at risk for developing adenocarcinoma. The defendant advised Mr. Tripp to return for a follow-up colonoscopy in one year.

The critical clinical fact, as found by the trial judge: the defendant did not complete the colonoscopy. He assumed that the five polyps were the source of Mr. Tripp’s rectal bleeding and did not continue the examination to the rest of the colon. A larger and more advanced lesion was present elsewhere in the colon and was missed by the incomplete examination.

The eight-month delay. Mr. Tripp continued to experience intermittent abdominal pain and rectal bleeding. In August 2018, the symptoms flared. His family physician referred him back to the defendant, who ordered investigations.

The September 2018 diagnosis. A barium enema in September 2018 demonstrated a tumour in the proximal descending colon. A CT scan that same month showed the tumour involving the spleen.

The October 2018 surgery. Mr. Tripp was referred to a surgeon who, in October 2018, removed the tumour. The procedure was extensive: the operation also removed 40 percent of the pancreas, the spleen, and all but a foot of the lower bowel. The cancer was staged as Stage IIIC at surgery: it had extended beyond the colon to involve the pancreas, with six lymph nodes positive for cancer.

Adjuvant chemotherapy. Mr. Tripp received eight cycles of adjuvant chemotherapy.

Post-operative complications. Mr. Tripp suffered substantial complications: a perforation of the surgical anastomosis between the small bowel and the colon, type 1 diabetes (likely related to the partial pancreatectomy), and a C. difficile infection.

The detection of liver metastases. A CT scan in December 2018 suggested possible liver lesions. An MRI confirmed a liver metastasis greater than one centimetre in size. Serial imaging through 2019 tracked the progression of the metastatic disease.

The stage IV diagnosis. In January 2020, Mr. Tripp was advised that he had Stage IV colorectal cancer with liver metastases. He commenced a further course of chemotherapy.

The death. Mr. Tripp died from colorectal cancer in February 2022.

The action. Mr. Tripp’s estate brought a malpractice action against the defendant gastroenterologist. The defendant conceded prior to trial that his performance of the January 2018 colonoscopy fell below the standard of care: he had not completed the examination and had missed a larger and more advanced cancer that was present at the time. The defendant also conceded that the eight-month delay in diagnosis had caused some increased pain and suffering for which compensation was appropriate. The contested question was whether the eight-month delay had caused Mr. Tripp’s death.

The trial decision

The trial judge of the Manitoba Court of King’s Bench (decision at 2023 MBKB 173) heard expert evidence on the causation question. Each side called an oncologist:

• The plaintiff’s expert was Dr. Schipper
• The defendant’s expert was Dr. Moore

The trial judge found both experts impressive and forthright, with no appearance of bias. The substantive content of their evidence was largely consistent, with the differences arising from how the questions were framed rather than from any fundamental disagreement on the science.

The agreed clinical framework. Both experts agreed on the foundational propositions:

• Colon cancer is slow-growing
• The tumour found in October 2018 had been present for a long time
• Once colon cancer metastasizes to a distant site such as the liver, the disease becomes fatal in most cases

The critical disputed question. The key question for causation was whether the liver metastasis that was subsequently detected had already been present in January 2018 at the time of the negligent colonoscopy. If it had, then even if the colon tumour had been detected at the colonoscopy, the underlying disease would already have been metastatic and the prognosis would have been substantially worse.

Dr. Moore’s analysis (defence). Dr. Moore worked from the size of the liver metastasis as it was detected in December 2018 (greater than one centimetre) and the typical doubling time of colorectal cancer liver metastases. He calculated that the metastasis would have required 15 to 30 months to reach its detected size from a single cell. The implication: the metastasis was likely already present in January 2018, before the colonoscopy. Dr. Moore’s opinion on Mr. Tripp’s chance of survival in January 2018 was approximately 30 percent (because the underlying disease was already metastatic).

Dr. Schipper’s analysis (plaintiff). Dr. Schipper did not address the possibility of pre-existing liver metastasis in his written report. On cross-examination, he agreed that the liver metastasis “would have been there for many months if not years.” Even on the most favourable interpretation of the plaintiff’s case (no pancreatic involvement, no metastatic spread, only two or three lymph nodes involved), Dr. Schipper’s estimate of Mr. Tripp’s chance of survival in January 2018 was approximately 42 percent.

The trial judge’s conclusion on causation. The trial judge accepted Dr. Moore’s opinion that the liver metastasis was likely present in January 2018. The basis for accepting Dr. Moore’s evidence over the plaintiff’s case included not only the statistical doubling-time analysis but also the specific clinical features of Mr. Tripp’s disease and the absence of evidence from the plaintiff side establishing that the metastasis was probably not present in January 2018.

The trial judge framed the legal question as whether the plaintiff had shown that Mr. Tripp would have had a greater than 50 percent chance of survival if his cancer had been diagnosed in January 2018. On the evidence, the plaintiff could not meet that threshold. The 30 percent survival probability (on Dr. Moore’s analysis) and the 42 percent best-case probability (on Dr. Schipper’s analysis) both fell below 50 percent. The trial judge accordingly concluded that the defendant’s negligence had not caused Mr. Tripp’s death.

The damages awarded. The trial judge did award damages of $75,000 for the pain and suffering that Mr. Tripp experienced as a result of the eight-month diagnostic delay. The category of harm is distinct from the death itself: even where the cancer was already incurable at the time of the negligence, the patient suffered the conscious experience of cancer for a longer period because of the delay. The harm is recognized and compensable in Canadian malpractice law.

The appellate analysis — the Benhaim framework

The plaintiff appealed to the Manitoba Court of Appeal. The principal argument was that the trial judge had failed to apply the “robust and pragmatic” approach to causation from Snell v Farrell and should have drawn an adverse inference against the defendant because his negligence (the incomplete colonoscopy) had impaired the plaintiff’s ability to prove causation.

The Court of Appeal’s analytical approach traced the Canadian causation framework from its English origins through the leading Supreme Court of Canada decisions.

The “robust and pragmatic” framework — origins. The framework originates in the English House of Lords decisions in McGhee v National Coal Board, [1972] UKHL 7, and Wilsher v Essex Area Health Authority, [1987] UKHL 11. Lord Bridge in Wilsher described the basic principle: the burden of proving causation in medical malpractice cases remains on the plaintiff, but the trier of fact may take a “robust and pragmatic approach to the undisputed primary facts” to draw legitimate inferences from the available evidence.

The Canadian adoption — Snell v Farrell. The Supreme Court of Canada adopted the “robust and pragmatic” framework in Snell v Farrell, [1990] 2 SCR 311. Sopinka J described the English approach as promoting “a robust and pragmatic approach to the facts to enable an inference of negligence to be drawn even though medical or scientific expertise cannot arrive at a definitive conclusion.” The framework allows the trier of fact to draw inferences of causation on the basis of common sense where the medical and scientific evidence does not provide a definitive answer.

The refinement in Benhaim. The Supreme Court in Benhaim clarified several aspects of the framework that had been left open by Snell. The clarifications are doctrinally important:

The adverse inference is discretionary, not mandatory. Where the defendant’s negligence has impaired the plaintiff’s ability to prove causation, the trier of fact may draw an adverse inference of causation but is not obliged to. The decision whether to draw the inference is a question of fact and is owed deference on appeal. The framework operates as part of the ordinary fact-finding process rather than as a categorical rule.

The category of “negligently created causal uncertainty” does not displace the standard burden. The Supreme Court rejected the proposition that cases in which the defendant’s negligence has impaired causation proof should be subject to a rebuttable legal inference that shifts the burden onto the defendant. Wagner J explained: “Shifting the consequences of causal uncertainty in this manner risks turning defendant professionals into insurers.” The framework retains the orthodox onus on the plaintiff.

The legal standard is distinct from the scientific standard. Causation in law requires proof on the balance of probabilities. The scientific and medical standards typically require a higher degree of certainty before drawing conclusions on causation. The trier of fact applies the legal standard, which can be satisfied by evidence that would not establish scientific certainty.

Inferences are drawn from all the evidence. The decision to draw an adverse inference must be based on an evaluation of all the evidence in the case. The trier of fact does not draw inferences mechanically; the framework requires substantive engagement with the specific evidence at hand.

Application to Tripp. The Court of Appeal found that the trial judge had correctly applied the Benhaim framework. The trial judge had specifically considered whether to draw an adverse inference and had declined to do so. The decision not to draw the inference was a fact-finding decision that, in the absence of palpable and overriding error, was owed deference on appeal. The record reasonably supported the trial judge’s key factual findings (that the liver metastasis was present at the time of the negligent colonoscopy; that Mr. Tripp’s chance of survival in January 2018 was approximately 30 percent). The Court of Appeal accordingly found no basis to interfere.

The appellate analysis — statistical evidence

A separate argument on appeal addressed the trial judge’s use of statistical evidence. The plaintiff argued that the trial judge had erred by equating the legal burden of proof with statistics on five-year survival rates of different stages of colon cancer.

The Court of Appeal addressed this argument through the Benhaim framework on statistical evidence. Benhaim had specifically considered the use of statistical evidence in causation analysis:

The “naked statistics” problem. Pure statistical generalizations (for example, the proposition that 78 percent of lung cancers discovered fortuitously are at Stage I) are of limited value in determining causation in individual cases because they represent “accidental groupings” rather than evidence about the specific patient before the court. Without a link to the patient’s actual clinical circumstances, naked statistics do not establish what was probably true of the patient.

The “evidentiary bridge” requirement. Statistical evidence becomes useful when it is linked to the specific clinical circumstances of the plaintiff. Wagner J in Benhaim: “Without an evidentiary bridge to the specific circumstances of the plaintiff, statistical evidence is of little assistance.” The bridge is typically provided by expert evidence about how the statistical generalization applies (or does not apply) to the particular patient’s disease, the specific features of the case, and the differential considerations that affect the individual prognosis.

Application to Tripp. The Court of Appeal found that the trial judge’s reliance on statistical evidence in Tripp met the Benhaim requirement. The statistics on five-year survival rates for the various stages of colon cancer were not used as freestanding generalizations. They were used in conjunction with expert evidence about:

• The size of the liver metastasis as detected in December 2018
• The typical doubling time of colorectal cancer liver metastases
• The specific features of Mr. Tripp’s underlying disease
• The implications of these factors for the disease stage at the time of the negligent colonoscopy

The combination provided the “evidentiary bridge” that linked the general survival statistics to Mr. Tripp’s specific clinical situation. The trial judge’s reliance on the statistics was accordingly permissible.

The doctrinal anchors

Several doctrinal anchors emerge from the case.

The discretionary adverse inference framework. Where a defendant’s negligence has impaired the plaintiff’s ability to prove causation, the trier of fact may draw an adverse inference but is not obliged to. The framework operates as part of the ordinary fact-finding process and is owed deference on appeal. The framework parallels other discretionary inferences in Canadian civil procedure (for example, adverse inferences from failure to call a witness or failure to produce documents).

The rejection of “negligently created causal uncertainty” as a freestanding category. The Supreme Court in Benhaim specifically rejected the proposition that cases in which the defendant’s negligence has impaired causation proof should be subject to a rebuttable legal inference. The framework retains the orthodox onus on the plaintiff. The rejection is doctrinally important because it forecloses a category of argument that plaintiff counsel had been advancing in delayed-diagnosis cases.

The “below 50 percent” loss-of-chance versus causation distinction. Where the survival probability for the patient was less than 50 percent even with appropriate care, the death cannot be said on the balance of probabilities to have been caused by the negligence. The framework is well-established in Canadian malpractice law: the plaintiff must show that the negligence made the outcome worse on a more-likely-than-not basis. Where the underlying disease would probably have killed the patient even with appropriate care, the negligence did not legally cause the death.

The doctrinal point is significant for prospective clients. Many delayed-cancer-diagnosis cases involve underlying disease that was already advanced at the time of the alleged breach. Where the patient’s chance of survival with appropriate care was already less than 50 percent, the causation analysis typically fails even where the standard of care breach is clear.

The statistical evidence and “evidentiary bridge” framework. Statistical evidence in causation analysis must be linked to the specific clinical circumstances of the plaintiff. Naked statistical generalizations are of limited value. Expert evidence that connects the general statistics to the particular patient’s disease, the specific features of the case, and the differential considerations provides the bridge that makes the statistics useful. The framework is generalizable to any causation analysis that involves epidemiological or outcome data.

The legal-versus-scientific causation distinction. The legal standard for causation (proof on the balance of probabilities) is distinct from the scientific standard (often a higher degree of certainty). The distinction is doctrinally important because medical and scientific expert evidence sometimes resists the legal standard’s binary framing. The trier of fact applies the legal standard and is not bound by the scientific framework’s typical caution.

Appellate deference on causation findings. Causation is a question of fact. Trial judge findings on causation are owed deference on appeal in the absence of palpable and overriding error. Housen v Nikolaisen, 2002 SCC 33, is the foundational authority on the standard. Benhaim confirms that the framework applies specifically to the decision whether to draw an adverse inference of causation. Tripp applies it to a delayed-diagnosis fact pattern.

Pain and suffering from delayed diagnosis as a separate compensable harm. Even where causation of death fails, the patient who has experienced cancer for longer because of the delay has a compensable harm. The category includes the conscious experience of additional symptoms, the psychological burden of knowing the diagnosis longer, and the loss of treatment options that earlier diagnosis would have provided. The framework operates wherever a breach has produced harm that is distinct from the ultimate death. The damages in Tripp ($75,000) are modest but illustrate that the recovery is available.

The conceded-breach-with-contested-causation pattern. Where the defendant has conceded the standard of care breach, the contest at trial focuses on causation. The pattern is common in cancer-diagnosis-delay cases because the breach (a missed lesion, an incomplete examination, a failure to follow up) is often plain on the contemporaneous record. The causation question is then determined by expert evidence about the natural history of the disease and the prognosis at the time of the breach.

The expert evidence framework in oncology cases. Cancer-diagnosis-delay cases turn on expert evidence about cancer biology, tumour doubling times, staging at the time of the alleged breach, and survival statistics. The quality of the expert evidence on each side is typically the determinative feature of the case. In Tripp, the trial judge’s assessment of the two experts as both impressive and forthright, with the substantive disagreement being narrow, made the causation finding turn on the specific points of disagreement (most importantly, the timing of the liver metastasis).

Why this case matters

For families considering similar cases. Delayed-cancer-diagnosis cases are among the most common categories of malpractice inquiry. The framework that Tripp illustrates is important for understanding when these cases are viable.

The standard of care breach is often the easier question. Where a colonoscopy was incomplete, where a worrisome finding was not followed up, or where a referral for further investigation was not made, the SOC analysis is often relatively straightforward. The defence sometimes concedes the breach (as in Tripp) or the breach is established through expert evidence with limited contest.

The causation analysis is the harder question. Whether the breach affected the outcome turns on the natural history of the underlying disease. Where the disease was already advanced at the time of the breach (with a survival probability already below 50 percent), the breach typically cannot be shown to have caused the death on the balance of probabilities. The framework operates whenever the underlying disease trajectory was already substantially negative.

The timing of the breach relative to the disease trajectory matters. Where the breach occurred at a stage in the disease where intervention could realistically have changed the outcome (Stage I or early Stage II disease, for example), the causation analysis is more favourable to the plaintiff. Where the breach occurred at a stage where the disease was already metastatic or substantially advanced, the analysis is much more difficult.

Pain and suffering recovery is available even where causation of death fails. A finding that the breach did not cause death does not foreclose damages for the period of additional symptoms, the psychological burden of delayed knowledge, and the loss of treatment options. The recovery is typically modest in absolute terms but is recognized.

Statistical evidence is part of the analysis but requires expert linkage. The general survival statistics for different cancer stages are evidence in these cases, but they are not freestanding proof. The expert evidence must connect the statistics to the particular patient’s disease. Plaintiffs who rely on general statistics without specific evidence about the patient’s clinical trajectory typically face difficulties.

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

For physicians and clinical teams. A few practical observations:

Complete the examination. The trial finding that anchored Tripp was that the defendant had stopped the colonoscopy upon finding what he assumed was the source of the bleeding. The standard of care for colonoscopy requires complete examination to the cecum unless the procedure is otherwise contraindicated. The principle generalizes: where the standard of care contemplates a structured workup, partial completion based on assumptions about the underlying cause can produce a serious breach.

Document the cecal intubation and the inspection findings. Quality metrics for colonoscopy now include documentation of cecal intubation (typically with a photograph of an anatomical landmark) and time spent on withdrawal inspection. The documentation is part of the standard, not just defensive practice. Where the documentation is absent or incomplete, the trier of fact can draw inferences about what was actually done.

The “found the source of bleeding” assumption is dangerous. A patient with rectal bleeding can have multiple sources. The bleeding from polyps does not exclude bleeding from a more advanced lesion elsewhere in the colon. The clinical reasoning that supports stopping the examination requires more than finding an apparent explanation; it requires confidence that the explanation is sufficient, which typically requires completing the examination.

---

Decision Date: March 21, 2025

Jurisdiction: Manitoba Court of Appeal

Citation: Tripp v Ross, 2025 MBCA 25 (CanLII)

Underlying trial decision: Tripp v Ross, 2023 MBKB 173 (CanLII) (Manitoba Court of King’s Bench)

Outcome: Appeal dismissed. The Court of Appeal of Manitoba unanimously held that the trial judge had correctly applied the legal framework for causation, had appropriately exercised her discretion in declining to draw an adverse inference of causation against the defendant under the Benhaim framework, and had reasonably relied on statistical evidence with the requisite “evidentiary bridge” to the plaintiff’s specific clinical circumstances. The trial judge’s factual findings (that the liver metastasis was likely present at the time of the negligent colonoscopy in January 2018, and that Mr. Tripp’s chance of survival at that time was approximately 30 percent) were supported by the record and were owed deference. The plaintiff had not established on the balance of probabilities that the eight-month diagnostic delay had caused Mr. Tripp’s death. The trial judge’s award of $75,000 in damages for the pain and suffering caused by the delayed diagnosis was undisturbed.

Key authorities: Benhaim v St-Germain, 2016 SCC 48 (discretionary adverse inference; rejection of “negligently created causal uncertainty” as freestanding category; statistical evidence and the “evidentiary bridge”); Snell v Farrell, [1990] 2 SCR 311 (robust and pragmatic approach to causation); Clements v Clements, 2012 SCC 32 (but-for causation framework); Housen v Nikolaisen, 2002 SCC 33 (appellate review of factual findings; palpable and overriding error standard); Wilsher v Essex Area Health Authority, [1987] UKHL 11 (origins of the robust and pragmatic approach); McGhee v National Coal Board, [1972] UKHL 7 (origins of the robust and pragmatic approach).