The Unexamined Practice: Five Philosophical Concepts in an Orthopaedic Encounter

Most of us entered orthopaedic surgery drawn to its tangible, mechanical nature—the satisfaction of reduction and fixation, the clarity of anatomic restoration. Philosophical speculation rarely surfaces in a busy clinical practice. Yet the moment a patient walks into the clinic, we find ourselves navigating questions that have occupied philosophers for millennia: What can we truly know? When can we claim one thing caused another? Where do we draw the line between categories that exist in nature and those that we invent? To navigate these questions, we must confront the intellectual inertia that often accompanies technical expertise. The philosopher Bertrand Russell (1872 to 1970) warned that “the man who has no tincture of philosophy goes through life imprisoned in the prejudices derived from common sense…and from convictions which have grown up in his mind without the co-operation or consent of his deliberate reason.”1 Russell’s assertion is a call to intellectual autonomy. It suggests that without the “tincture” of critical examination, our clinical decisions may not be the product of our own reason; rather, they may be merely the reflection of unexamined habits. In an era in which clinical algorithms, evidence-based guidelines, and increasingly sophisticated artificial intelligence systems offer ready-made answers to complex questions, the intellectual courage to examine our own reasoning, rather than simply accept external directives, may be more necessary than ever. This article examines a routine clinical encounter, the displaced clavicle fracture, to illuminate some of the philosophical assumptions embedded in our daily practice. The aim is to demonstrate how philosophy can cultivate epistemic humility—a disciplined awareness of the boundaries of what can be known—which is a safeguard against the complacency of the unexamined practice. Every time we determine that a fracture is “healed,” apply the statistics from a study to a specific individual, or claim that our surgery “caused” improvement, we are relying on a philosophical infrastructure that rarely gets explicit attention. The Vignette A 32-year-old cyclist presents 3 days after a fall. He has a midshaft clavicle fracture, shortened by 2 cm, with a butterfly fragment. Fit and impatient, he asks the questions we hear every day: “How bad is it?” “Will surgery make it heal faster?” “When can I ride again?” From the first glance at his radiograph to his final discharge, this patient’s journey forces us to confront fundamental issues in epistemology, ontology, and causal reasoning. Theory-Laden Observation The first philosophical hurdle begins the moment the clinical encounter starts. As the cyclist walks in—shoulder depressed, head tilted—we do not merely see a man in pain. To the trained eye, his very posture is a diagnosis. Our training imposes a conceptual framework on the raw visual data. This cognitive filtering is what transforms novices into experts, but only when we remain aware that we are interpreting, not merely observing. As philosophers of science Norwood Russell Hanson (1924 to 1967) and Thomas Kuhn (1922 to 1996) argued, there is no such thing as neutral observation; what we see depends on what we know2,3. This concept becomes particularly important when we view the radiograph before examining the patient. We see evidence of comminution, and the subsequent physical examination is filtered through that expectation. We might “find” tenderness exactly where the radiograph suggested it should be, while subtle findings that the radiograph did not predict may go unnoticed. We see the patient through the lens of our expectations. Even measuring displacement is theory-laden. We filter sensory input through conceptual frameworks—in this case, operative versus nonoperative thresholds. This helps explain why 2 experienced surgeons can look at the same radiograph and literally see 2 different things. Expert pattern recognition functions as a cognitive filter, giving more weight to certain features than others. The Reference Class Problem The cyclist looks at us and asks, “What are the odds I need surgery to get back to racing?” To answer, we must first decide which group of patients he belongs to. This is the reference class problem: every individual belongs to multiple plausible groups, and the probability of an outcome depends entirely on which group we choose. The logician John Venn (1834 to 1923) noted that any single patient simultaneously belongs to innumerable overlapping classes, each associated with a different probability of the same event4. Consider our cyclist: In the class of “all clavicle fractures,” the prior probability of nonunion is generally considered low; surgery seems unnecessary. In the class of “displaced midshaft fractures with comminution,” the nonunion rate rises to ∼11%. Surgery seems reasonable5. In the class of “high-demand, competitive, 32-year-old amateur cyclists,” the definition of a “good outcome” changes again, and we have limited data from which to estimate probabilities. The philosopher Hans Reichenbach (1891 to 1953) argued that probability estimates should draw on the narrowest reference class for which reliable statistics are available6. But as we narrow the class to fit our specific patient, the sample size approaches zero. We are left to choose which group membership matters most—and that choice is not determined by the data itself. This problem is compounded by a well-documented cognitive bias: base rate neglect. As the psychologists Daniel Kahneman (1934 to 2024) and Amos Tversky (1937 to 1996) observed, humans consistently ignore prior probabilities when faced with vivid individual information7. The displacement on the radiograph creates a powerful visual impression. The bone looks severely deformed. The vividness tempts us to judge the probability of failure by resemblance rather than by starting with the fact that most clavicles heal. We may ultimately determine that surgery is the right choice; indeed, studies such as the Canadian Orthopaedic Trauma Society trial suggest functional benefits when surgically treating this fracture pattern5. But the philosophical challenge lies in how we arrive there. Are we consciously selecting a reference class and updating our prior probability calculation based on new evidence? Or are we letting the vividness of the image override the base rate entirely? Risk Versus Uncertainty The patient asks about complications. We estimate a nonunion rate of 5% for nonoperative management and an infection rate of <2% for surgery. Here we are dealing with risk, measurable variability in which probabilities are known. The economist Frank Knight (1885 to 1972) distinguished risk from uncertainty, in which probabilities are unknowable8. Surgeons are comfortable with risk; we struggle with uncertainty9. The patient’s return to work, the sensation of the plate under a backpack strap, or the psychological impact of curtailing activity fall into the realm of Knightian uncertainty. These outcomes cannot be quantified. Yet the pressure to provide definitive answers can lead us to treat uncertainty as if it were risk, assigning percentages to outcomes that resist quantification. Acknowledging this distinction allows us to communicate more honestly: while we can cite statistics about fracture healing, the subjective experience of recovery involves variables that cannot be reduced to probabilities. The Sorites Paradox and Ontology Six weeks postoperatively, the patient returns. The radiograph shows callus formation. He asks, “Is it healed?” This raises an old problem: when does a gradual biological process become a categorical clinical fact? Here we encounter the sorites paradox (or the paradox of the heap)10. If you remove 1 grain of sand from a heap, it remains a heap. At what precise point, grain by grain, does it cease to be one? Similarly, fracture healing is a biological continuum, yet surgeons must impose binary thresholds when clearing a patient for weight-bearing or return to sport. This leads us to a deeper ontological question: Are our categories natural kinds (discovered in nature) or constructed categories (created for utility)? The metaphor of “carving nature at its joints”11 used by Plato (circa 428 to 348 bce) has become a touchstone for the idea that some categories reflect real discontinuities. A midshaft clavicle fracture seems to represent such a category because the bone is objectively broken. But delayed union is different: rather than discovering a biological entity, we are applying a label because the calendar has passed 3 months. These categorical boundaries are necessary for utility but are ontologically slippery. Recognizing this prevents us from mistaking our constructed labels for biological destiny. Causal Inference and Counterfactuals Three months later, the cyclist returns with full range of motion and solid union, and he is back to racing. He shakes our hand and says, “Thank you for fixing me.” We feel a surge of professional satisfaction: the surgery worked. But this moment of satisfaction hides a profound and common philosophical problem. We are encountering the logical fallacy of post hoc, ergo propter hoc, meaning “after this, therefore because of this.” Just because the healing followed the surgery does not logically entail that the surgery caused the healing. To claim that “the surgery caused the healing” is to make a claim about a history that never happened. This is the realm of counterfactuals, or reasoning about what would have occurred under different conditions. The philosopher David Lewis (1941 to 2001) defined causation as counterfactual dependence12. Lewis argued that to say “A caused B” is simply shorthand for the claim “If A had not occurred, B would not have occurred.” Here lies the trap: we can never observe that parallel scenario. As the statistician Paul Holland (1940-) famously noted, this is the “fundamental problem of causal inference”: we cannot observe the counterfactual outcome for the same individual13. In orthopaedics, this truth is particularly humbling because the natural history (the counterfactual) is often benevolent. If the cyclist had been treated nonoperatively, he likely would have healed as well. If the outcome—union—would have occurred regardless of treatment, then under Lewis’s definition, the surgery was not a “difference-maker,” even if we took credit for it. This does not mean that surgery never makes a difference—only that we should remain appropriately humble about when it does and when natural history would have sufficed. When we fail to grapple with the unobservable counterfactual, we risk conflating production (e.g., the mechanics of putting in screws) with causation (e.g., altering the destiny of the patient). Epistemic humility asks us to consider that, at times, the surgeon is not the architect of healing, but rather a passenger on the biological journey of the patient. Returning to Epistemic Humility The journey of our cyclist, from theory-laden observation to reference class problems, uncertainty, vagueness, and counterfactuals, reveals the philosophical depth beneath the technical surface of orthopaedic practice. Evidence-based medicine provides frameworks for selecting reference classes, quantifying uncertainty, and testing causal claims, but it cannot eliminate the underlying philosophical problems. The surgeon must still choose which reference class applies, distinguish risk from uncertainty, and acknowledge the unobservable counterfactual. Recognizing the limits of our knowledge is not a new idea; it is the foundational insight of Western philosophy. As recounted by Plato, Socrates (circa 470 to 399 bce) located wisdom not in the accumulation of facts but in the recognition of one’s own ignorance14. This stance (epistemic humility) is not a lack of confidence, but rather a disciplined awareness of the boundaries of what can be known. In the operating room, this humility is a safety mechanism. It prevents the confidence of experience from hardening into the dogmatism that Russell warned against1. When we ignore the philosophical substrate of our profession, we risk remaining “imprisoned in the prejudices derived from common sense,” mistaking our habits for truths and our observations for objective reality. Philosophical literacy does not make the fracture heal faster, the reductions more precise, or our outcomes more predictable. But it may give us a clearer perspective on the limits of knowledge and the boundaries of intervention. It encourages us to question our assumptions, to remain alert to the patient who defies the reference class, and to reconsider our claims regarding cause and effect. It allows us to navigate the fog of clinical uncertainty with a more considered understanding of what we actually know, and, more importantly, what we do not.