Cognitive bias in veterinary clinical decision‐making: Antemortem misdiagnosis of suspected giant kidney worm ( Dioctophyme renale ) in Australia

Abstract

A 10‐year‐old male neutered Rottweiler was referred for evaluation of suspected Dioctophyme renale infection following haematuria, lethargy, coughing and passage of a worm‐like structure in the urine. The presumptive diagnosis was strongly influenced by the referring veterinarian's prior experience with D. renale in an endemic region. Diagnostic imaging identified left renal abnormalities including hydronephrosis, a dilated ureter containing a tubular intraluminal structure and multiple tubular structures within the urinary bladder; these observations further reinforced the initial clinical impression. However, several key findings, including the absence of ova on urinalysis and the presence of a discrete renal mass with disseminated nodules, were inconsistent with dioctophymosis; 'these were underweighted in clinical reasoning. Rapid clinical deterioration and evidence of extensive multisystemic disease led to euthanasia. Postmortem examination revealed metastatic renal haemangiosarcoma (epithelioid variant), and the presumed nematodes' were confirmed to be ureteral blood casts. This case illustrates how cognitive biases can influence diagnostic reasoning, particularly in unfamiliar epidemiological contexts. It highlights the importance of systematic reassessment of differential diagnosis, integration of local disease prevalence and critical appraisal of initial impressions to mitigate diagnostic error in veterinary practice.

Clinical reasoning underpins veterinary decision‐making but is inherently complex and susceptible to error. ¹ , ² , ³ Diagnostic error is increasingly recognised as a major contributor to patient harm in both human and veterinary medicine and arises from an interplay of cognitive factors, uncertainty inherent to clinical practice, and system‐related influences. ¹ , ² , ³ , ⁴ , ⁵ , ⁶ , ⁷ , ⁸ , ⁹ , ¹⁰ , ¹¹ , ¹² , ¹³

Heuristics are essential cognitive shortcuts that enable clinicians to make rapid decisions in time‐pressured environments. ¹² , ¹³ , ¹⁴ , ¹⁵ However, when insufficiently interrogated, they may lead to systematic errors in reasoning. ¹² , ¹³ , ¹⁴ , ¹⁵ , ¹⁶ Dual‐process theory provides a framework for understanding these processes, distinguishing intuitive, experience‐based System 1 thinking from slower, analytical System 2 reasoning. ¹⁵ , ¹⁷ , ¹⁸ While System 1 thinking promotes efficiency, it is more vulnerable to bias, whereas System 2 thinking enables more deliberate evaluation but may be underutilised in acute or high‐pressure clinical settings. ¹³ , ¹⁴ , ¹⁵ , ¹⁶

Anchoring bias, defined as the disproportionate weighting of initial diagnostic impressions, is particularly influential in complex or unfamiliar cases and is often reinforced by confirmation bias, whereby evidence supporting a preliminary hypothesis is preferentially interpreted while discordant information is discounted. ⁴ , ¹⁹ Together, these biases increase the risk of ‘premature diagnostic closure’ and misinterpretation of clinical data. ¹³ , ¹⁹ Additional cognitive biases further contribute to diagnostic error, including diagnostic momentum, ‘bandwagoning’, and framing effects. ¹⁵ , ¹⁶ , ²⁰ In parallel, clinician‐related factors such as overconfidence and intolerance of uncertainty have been associated with reduced diagnostic accuracy, particularly in situations characterised by incomplete information or time pressure. ¹⁵ , ¹⁶

Certain clinical scenarios are particularly susceptible to bias, especially when striking clinical features align with a recognisable diagnostic pattern or when prior clinician experience strongly informs interpretation. In such contexts, rare diseases may be disproportionately weighted over more probable alternatives. ¹⁴ , ²¹ , ²² One condition that exemplifies these diagnostic challenges is Dioctophyme renale (Goeze, 1782), the ‘giant kidney worm’, a large, blood‐red nematode that infects a range of piscivorous mammals, including domestic dogs, mustelids, wild canids and rarely, humans. ²³ , ²⁴ , ²⁵ , ²⁶ , ²⁷ , ²⁸ In dogs, infection is associated with progressive renal destruction, most often affecting the right kidney. ²³ , ²⁵ The parasite has a broad but geographically variable global distribution and is most commonly reported in South America; its occurrence in Australia is poorly documented, and confirmed veterinary cases are lacking. ²³ , ²⁴ , ²⁵

The notable gross appearance of this parasite, combined with its distinctive pathological effects, may create a compelling illness script when suspected clinically. However, reliance on such pattern recognition in the absence of confirmatory evidence increases susceptibility to cognitive bias, particularly in non‐endemic settings. ¹² , ¹⁴ , ¹⁵ , ²⁶ , ²⁹ , ³⁰

This case report describes the antemortem misdiagnosis of suspected D. renale infection in a dog in Australia that was ultimately diagnosed with metastatic renal haemangiosarcoma. The case is used as a framework to examine how multiple interacting cognitive biases influence clinical reasoning and contribute to diagnostic error. By analysing the diagnostic process in this case, this report aims to highlight practical strategies to improve reflective practice, promote analytical reasoning and reduce the impact of cognitive bias in veterinary clinical decision‐making.

Case report

A 10‐year‐old male neutered Rottweiler (46 kg) was referred to a specialist hospital in Brisbane for evaluation of suspected Dioctophyme renale infection. One week earlier, the dog had presented to a primary care clinic with acute haematuria. Vitamin K and amoxicillin–clavulanic acid were prescribed empirically due to concern for anticoagulant rodenticide intoxication, supported by an elevated activated clotting time and suspected urinary tract infection based on clinical assessment and in‐house urinalysis. Six days later, worsening clinical signs, including persistent haematuria, lethargy, inappetence, tachypnoea, abdominal discomfort and a non‐productive cough, led to re‐presentation. The owner reported passage of a 15 cm dark red, worm‐like tubular structure in the urine. Based on this finding and prior clinical experience in Brazil, the referring veterinarian made a presumptive diagnosis of dioctophymosis. Environmental history included exposure to green tree frogs (Litoria caerulea).

On referral examination, the patient was quiet but responsive, tachypnoeic (44 breaths/min; reference interval [RI] 10–30) and mildly underweight (body condition score 4/9). Thoracic auscultation revealed mildly increased bronchovesicular sounds with an occasional non‐productive cough. The abdomen was tense but non‐distended. Thoracic radiographs demonstrated a diffuse structured interstitial pattern. Point‐of‐care thoracic ultrasonography identified multiple small peripheral pulmonary nodules. Abdominal ultrasonography revealed a moderately distended urinary bladder containing a dependent heteroechoic intraluminal structure composed of multiple tubular hypoechoic elements with hyperechoic walls, an enlarged left kidney with reduced corticomedullary definition and severe pyelectasia, and a markedly dilated left ureter containing an intraluminal soft tissue structure extending toward the bladder (Figure 1). Additional findings included a well‐defined, heterogeneous, encapsulated mass at the caudal pole of the left kidney (approximately 50 × 30 mm), multiple diffuse hypoechoic hepatic nodules and mild splenic parenchymal heterogeneity.

Figure 1.

Ultrasonographic findings in a dog with suspected Dioctophyme renale infection. (A) Transverse view of the left ureter showing marked dilation (8.1 mm in diameter). (B) Transverse view of the left ureter demonstrating a segmental, tubular, non‐shadowing soft‐tissue structure within the ureteral lumen. (C) Longitudinal view of the urinary bladder showing an irregular, heteroechoic, non‐shadowing, dependent tubular structures within the bladder lumen. (D) Sagittal view of the left kidney demonstrating decreased corticomedullary definition and marked hydronephrosis.

Haematology revealed moderately severe regenerative anaemia (HCT 0.22 L/L; RI 0.37–0.55 L/L) and leucocytosis (19.1 × 10⁹/L; RI 5.3–14.9 × 10⁹/L), characterised by neutrophilia (15.1 × 10⁹/L; RI 3.1–10.3 × 10⁹/L) and monocytosis (1.3 × 10⁹/L; RI 0.2–1.0 × 10⁹/L). Serum biochemistry demonstrated mild azotaemia, with increased urea (9.9 mmol/L; RI 2.5–9.0 mmol/L) and creatinine (158 μmol/L; RI 44–125 μmol/L). Urinalysis confirmed adequate urine concentration (USG 1.030) with marked haematuria and proteinuria; no microorganisms or D. renale ova were detected, and aerobic culture was negative.

Diagnosis and clinical reasoning

The constellation of imaging findings was initially interpreted as supportive of a presumptive diagnosis of D. renale infection. Tubular intraluminal structures within the bladder and left ureter were considered consistent with expelled adult nematodes originating from the left kidney, and severe renal hydronephrosis and ureteral dilation were attributed to luminal obstruction by the presumptive parasite. The left renal mass was interpreted as a parasitic granuloma, while pulmonary and hepatic nodules were considered consistent with ectopic parasitic localisation or aberrant larval migration. The presumptive diagnosis of D. renale was maintained across both the referring and specialist institutions and was accepted as the leading explanation for the patient's clinical presentation at the time of antemortem decision‐making.

Notably, several clinical, imaging and laboratory findings present at the time of assessment were discordant with this diagnosis (Table 1). ²³ , ²⁴ , ²⁵ , ²⁶ As summarised in Table 1, several high‐specificity features were discordant with dioctophymosis, including the absence of characteristic D. renale ova on urinalysis, predominant left‐sided renal involvement, the presence of a discrete renal mass and widespread pulmonary and hepatic nodules. Despite their diagnostic significance, these features were underweighted in clinical reasoning and in ordering of differential diagnoses.

Table 1.

Key discrepancies between findings in this case and expected features of Dioctophyme renale infection

Clinical feature	Findings in this case	Expected in D. renale infection	Diagnostic implication
Renal involvement	Left kidney	Predominantly right kidney; left‐sided or bilateral involvement is reported but is uncommon	Atypical for dioctophymosis
Urinalysis (ova)	Not detected	Often present in patent infections; may be absent with intermittent shedding, male‐only infections or ectopic localisation	Reduces diagnostic support for parasitic infection
Renal morphology	Discrete renal mass	Progressive parenchymal destruction without focal mass formation	Reduces diagnostic support for parasitic infection. Strongly suggests alternative pathology (e.g., neoplasia)
Pulmonary and hepatic findings	Multiple nodules	Extrarenal involvement reported but rare and not typically characterised by discrete nodular lesions	Reduces diagnostic support for parasitic infection. Strongly suggests alternative pathology (e.g., neoplasia)
Tubular structure in urinary tract	~20‐cm red tubular structure in the left ureter	Adult worms typically occupy the renal pelvis; spontaneous expulsion is rare	Misleading gross appearance with high risk of morphological misinterpretation
Clinical progression	Rapid, multisystemic disease	Typically chronic, insidious or incidental; acute presentations are uncommon	Reduces diagnostic support for parasitic infection

From a probabilistic perspective, the working diagnosis was inconsistent with expected disease prevalence. Metastatic neoplasia represents a relatively common condition in geriatric dogs, whereas D. renale infection in Australia has an extremely low prior probability in native populations. Failure to appropriately weight these competing probabilities likely contributed to persistence of the initial diagnosis despite conflicting evidence. Differential diagnoses considered included metastatic renal neoplasia, primary pulmonary or hepatic neoplasia, disseminated mycoses, sterile granulomatous disease and other parasitic infections; however, these alternatives were not prioritised during antemortem assessment.

Given progressive clinical decline, extensive multiorgan involvement and a poor overall prognosis, euthanasia was elected. This decision was based on the patient's clinical deterioration and disease burden rather than the presumptive diagnosis.

Postmortem examination

With owner consent, postmortem diagnostics were performed to further investigate the suspected D. renale infection. Whole‐body computed tomography (CT) was performed in ventral recumbency without contrast. Diffuse, well‐defined, multifocal soft‐tissue nodules were identified throughout all lung lobes, predominantly peripheral in location (Figure 2), in addition to multifocal ground‐glass opacities, particularly in ventral lung lobes, considered most consistent with postmortem atelectasis. Abdominal imaging revealed marked left renomegaly with severe hydronephrosis and a well‐defined ovoid mass (41.6 × 26.9 mm) at the caudal pole of the left kidney (Figure 3), exhibiting heterogeneous attenuation with hypodense regions suggestive of cystic or necrotic change. Multiple hypoattenuating hepatic nodules were also present.

Figure 2.

Postmortem computed tomography (CT) images of the thorax. Images are shown in (A) dorsal, (B) transverse and (C) sagittal planes. Diffuse, multifocal, well‐defined soft‐tissue nodules are present throughout all lung lobes with a predominantly peripheral distribution, consistent with a structured or nodular interstitial pattern. Ground‐glass attenuation, most evident in the ventral lung lobes on sagittal and transverse planes, is considered consistent with postmortem atelectasis associated with ventral recumbency.

Figure 3.

Postmortem dorsal plane CT image of the abdomen. (A) liver, (B) right kidney, (C) left kidney showing marked left renomegaly with severe renal pelvic dilation is evident, consistent with hydronephrosis. (D) A well‐marginated, ovoid soft‐tissue mass (41.6 × 26.9 mm) arises from the caudal pole of the left kidney and lies adjacent to the dilated ureter. The mass displays heterogeneous attenuation with hypodense regions suggestive of cystic or necrotic change.

A necropsy was subsequently performed. The lungs, were diffusely congested and contained multifocal nodules (4–20 mm), with larger lesions exhibiting cystic and haemorrhagic change (Figure 4). The liver and spleen were mildly enlarged and contained multiple nodules. The left kidney was enlarged (107 mm) with severe hydronephrosis and a markedly ectatic ureter with mucosal ulceration (Figure 5A). An encapsulated, friable mass (45 mm) was present at the caudal pole (Figure 5B). No grossly identifiable nematodes were observed. The right kidney was normal. A 20 cm red tubular structure was recovered from the left ureter (Figure 6), and the urinary bladder contained fragmented tubular material.

Figure 4.

Gross postmortem photograph of the left cranial and caudal lung lobes, showing diffuse dark red discolouration of the pulmonary parenchyma with multifocal cystic and haemorrhagic nodules measuring approximately 4–20 mm in diameter.

Figure 5.

Gross postmortem photographs of the left kidney (A) and excised caudal pole mass (B). The left ureter is markedly dilated and opened longitudinally, revealing multiple dark red mucosal foci up to 8 mm in diameter (white arrows), consistent with ulceration. A well‐circumscribed, encapsulated mass is present at the caudal pole of the kidney (blue arrow). (B) Sectioned caudal pole mass from the left kidney demonstrating friable solid tissue without cystic components.

Figure 6.

Gross photographs of a structure initially suspected to represent a nematode and removed from the left ureter at postmortem examination. The red, tubular structure measured approximately 20 cm in length and was grossly consistent in appearance with Dioctophyme renale. Panels (A) and (B) show the specimen immediately following removal from the ureter, and panel (C) shows the specimen after preservation in ethanol prior to transport.

Pathological findings

The tubular structure recovered from the left ureter was submitted for parasitological evaluation. Stereomicroscopy did not support nematode morphology, and histological assessment confirmed organised blood clots (tubular casts) formed secondary to ureteral haemorrhage. Histopathology of the renal mass identified a highly invasive malignant neoplasm infiltrating the ureter and perirenal tissues, with widespread pulmonary and hepatic metastases. Immunohistochemistry demonstrated diffuse CD31 expression and patchy positivity for factor VIII‐related antigen, supporting a diagnosis of renal haemangiosarcoma (epithelioid variant). No parasitic structures or ova were identified in any examined tissue.

Discussion

This case illustrates the antemortem misdiagnosis of Dioctophyme renale infection in a dog in Australia and demonstrates how interacting cognitive biases can dominate clinical reasoning, resulting in diagnostic error despite the presence of incongruent clinical evidence. Although D. renale was considered the most likely diagnosis during antemortem assessment, several clinical, imaging and laboratory findings were inconsistent with this interpretation and were more consistent with metastatic neoplasia. Failure to appropriately integrate and weight these discrepancies represents a breakdown in analytical reasoning. Importantly, the discordant findings in this case were present at the time of antemortem decision‐making, indicating that the diagnostic error reflects a failure of real‐time analytical reasoning rather than retrospective reinterpretation.

Anchoring bias played a central role in this case, with the initial diagnostic impression formed following observation of a worm‐like structure disproportionately influencing all subsequent interpretations. ¹³ , ¹⁹ This early diagnostic impression influenced interpretation of diagnostic imaging findings, with tubular intraluminal structures assumed to represent parasitic material rather than alternative processes such as haemorrhage or clot formation. Confirmation bias further reinforced this interpretation, as findings perceived to support parasitic infection were preferentially weighted, while discordant features were under‐recognised or discounted. ⁴ , ¹² , ¹³ , ²² , ²³ , ²⁴ , ²⁵ Together, these processes contributed to premature closure, a recognised contributor to diagnostic error. ²⁹ , ³¹

Diagnostic momentum further contributed to persistence of the presumptive diagnosis. Once established, the diagnosis of D. renale was propagated amongst the responsible clinicians and became entrenched as the leading explanation for the patient's presentation, a process well described in the clinical reasoning literature. ²² , ²⁹ Bandwagoning may also have contributed, with repeated acceptance reinforcing the perceived validity of the diagnosis. ¹⁵ , ²²

Framing effects influenced how the case was conceptualised at a broader level. The initial presentation, centred on the reported passage of a worm‐like structure, established a diagnostic context in which parasitic infection was prioritised over alternative explanations, directing subsequent reasoning toward parasitic infection rather than alternative diagnoses. In this context, visually conspicuous findings acted as cognitive distractors, promoting reliance on intuitive (System 1) reasoning and reducing engagement with analytical (System 2) processes. ¹³ , ¹⁷ , ¹⁸ , ²⁰ The referring clinician's initial diagnostic impression, informed by prior experience with D. renale in endemic regions, established the initial diagnostic frame and set the conditions for subsequent cognitive bias, influencing interpretation of findings across clinicians.

In addition to cognitive biases, clinician‐related factors likely contributed to diagnostic error. Overconfidence in an initial diagnosis and a low tolerance for diagnostic uncertainty have been associated with reduced diagnostic accuracy, particularly in complex or ambiguous cases. ¹⁵ , ¹⁶ Conversely, fear of missing a rare or emerging disease may also influence clinical reasoning, particularly in nonendemic settings where unfamiliar conditions are perceived as novel or diagnostically significant. ³ , ²⁹ This pattern is consistent with a recognised cognitive tendency to overvalue rare or novel diagnoses (“zebra hunting”), leading to under‐recognition of more likely conditions. ²² In this case, the possibility of identifying a previously unreported or emerging parasitic infection in Australia may have increased index of suspicion for D. renale as a diagnosis, reinforcing anchoring and reducing willingness to prioritise more common possibilities. Such dynamics highlight how both overconfidence and uncertainty‐related factors can contribute to diagnostic error.

Although D. renale infection is an important differential diagnosis in endemic regions, its diagnosis is inherently challenging, as definitive confirmation requires identification of adult worms or characteristic ova, and reliance on imaging or gross morphology alone may lead to misdiagnosis. ²³ , ²⁴ , ²⁵ , ³² , ³³ This limitation is illustrated by the only reported Australian human case (1983), in which intrarenal structures initially interpreted as D. renale ova were later considered more consistent with Liesegang rings. ³⁰ , ³² , ³³ Similar diagnostic uncertainty has been reported in other cases where diagnoses relied solely on putative ova morphology. ³² , ³³

The striking appearance of tubular red material in this case, both reportedly passed antemortem within urine and retrieved postmortem from the left ureter, strongly reinforced the presumptive diagnosis. However, intraluminal blood clots, fibrin casts and necrotic tissue can closely mimic parasitic structures. ²³ , ³⁴ Comparable diagnostic challenges have been reported in human medicine, where worm‐like structures have led to presumptive diagnoses of D. renale infection in the absence of definitive confirmation, including cases in which patients were treated empirically with anthelmintics prior to recognition of an underlying malignancy. ²⁶ , ²⁷ , ²⁸ These reports highlight the importance of confirmatory diagnostics and cautious interpretation of visually compelling findings, particularly in non‐endemic settings. Systematic evaluation of such discrepancies through deliberate, objective, analytical reasoning is a key component of diagnostic accuracy. ¹³ , ¹⁴ , ¹⁵

Cognitive bias is intrinsic to clinical practice and independent of individual experience or expertise. Its impact in veterinary medicine may be amplified by time pressures, financial constraints, diagnostic uncertainty and the breadth of species and disease presentations encountered in routine practice. ¹ , ⁴ , ⁶ , ⁷ , ⁸ , ⁹ , ¹⁰ , ¹¹ , ¹² Diagnostic errors arising from cognitive bias are increasingly recognised as a major source of patient harm in both human and veterinary medicine. ¹ , ² , ³ , ⁴ , ⁵ , ⁶ , ⁷ , ⁸ , ⁹ , ¹⁰ , ¹¹ , ¹² , ²⁹ , ³¹ Over the past two decades, human medicine has developed explicit diagnostic taxonomies and safety frameworks and error‐reporting systems to characterise and reduce diagnostic errors. In contrast, veterinary medicine remains comparatively unstandardised in this area. ⁴ , ⁷ , ⁸ , ⁹ , ²¹

Although cognitive biases cannot be eliminated, their impact may be mitigated through structured approaches grounded in both individual cognition and clinical systems. These include explicit generation of alternative diagnoses, deliberate search for disconfirming evidence and incorporation of pretest probability and local epidemiology into diagnostic reasoning. ¹ , ¹³ , ¹⁴ , ¹⁵ , ³¹ Practical approaches include structured “diagnostic time‐outs”, systematic identification of discordant findings and cognitive forcing techniques that may help counter premature closure and diagnostic momentum. ¹³ , ¹⁴ , ¹⁵ , ¹⁷ , ¹⁹ , ³¹

Development of metacognitive skills, defined as awareness and regulation of one's own thinking processes, has been increasingly recognised as a core debiasing strategy. ¹ , ¹³ , ¹⁴ , ³⁵ Metacognition enables clinicians to recognise situations of high cognitive risk and deliberately shift from intuitive (System 1) to analytical (System 2) reasoning. ¹³ , ¹⁷ Incorporation of probabilistic approaches, including Bayesian reasoning, may further improve diagnostic calibration by explicitly integrating prior probability with evolving clinical data. ³⁶ Emerging technologies, including artificial intelligence, may assist in reducing diagnostic error by augmenting clinical reasoning, prompting consideration of alternative diagnostic possibilities, highlighting discordant findings and incorporating epidemiological context to recalibrate diagnostic probability. As a cognitive “backstop,” these systems may help counteract anchoring and confirmation bias, provided they are critically appraised and appropriately integrated into clinical workflows. ³⁷ At a systems level, fostering a clinical culture that supports reflective practice and open discussion of diagnostic reasoning is essential. Incorporating cognitive error analysis into morbidity and mortality rounds, together with explicit discussion of diagnostic uncertainty, may help normalise reconsideration of the initial diagnosis, thereby reducing the impact of diagnostic momentum. ⁶ , ⁷ , ⁸ , ⁹ , ³⁵

Primary renal haemangiosarcoma with distant metastasis carries a poor prognosis, with limited evidence supporting meaningful benefit from chemotherapy. ³⁸ , ³⁹ No standardised treatment pathway exists, and management is largely extrapolated from other visceral forms, with outcomes remaining poor. ³⁸ , ³⁹ Although cognitive bias contributed to delayed diagnosis, the extent of multisystemic disease suggests that the clinical outcome would likely not have differed with an accurate antemortem diagnosis. Additionally, had the aetiology been D. renale, surgical management would not have been feasible in the presence of disseminated disease, and anthelmintic therapy would have been unlikely to provide meaningful benefit given limited efficacy and poor tissue penetration. ²³ , ²⁴ , ²⁵

Conclusion

This case demonstrates how cognitive biases can contribute to diagnostic error, even in the presence of clear and compelling discordant clinical evidence. Such biases are inherent to human decision‐making and affect clinicians across all levels of experience, from students to general practitioners and specialists, in both primary care and referral settings. The case further demonstrates how visually compelling findings and the perceived significance of a rare or novel diagnosis may disproportionately influence clinical reasoning.

Mitigation of bias‐related errors requires integration of cognitive debiasing strategies, including structured reflection, metacognitive awareness, probabilistic reasoning and deliberate evaluation of discordant findings, together with emerging decision‐support tools to improve diagnostic accuracy and patient safety in veterinary practice.

Conflicts of interest and sources of funding

The authors declare no conflicts of interest or sources of funding for the work presented here.

Data availability statement

The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.