Uncertainty in orthopaedics

Clinical Orthopaedics and Related Research, 2016

This study examined whether uncertainty intolerance influences surgical decision-making in clinical orthopedics. We found that surgeons vary significantly in their tolerance for ambiguity, and that this variation affects treatment recommendations.

Journal of the American Academy of Orthopaedic Surgeons, 2021

Cognitive biases are ubiquitous in human reasoning, and orthopedic decision-making is no exception. This study evaluated the prevalence of specific biases—including anchoring, availability, and framing effects—among academic orthopedic surgeons.

PLOS ONE (submitted)

 Bayesian probability is notoriously difficult to apply in clinical settings, yet it remains unclear how often surgeons actually use Bayesian reasoning in practice. This study quantified the prevalence of base rate neglect and other non-Bayesian patterns in clinical decision-making vignettes.

High-frequency experience sampling captures what traditional outcomes research misses: the pathway of recovery, not just the destination. By querying patients daily with a single-item functional score, we construct continuous recovery curves that reveal improvement velocity, total disability burden, and the moment healing stabilizes. 

Fracture healing is a continuous biological process, yet orthopedic research routinely forces it into categorical endpoints—healed or not healed, union or nonunion. This project examines the epistemological problems that arise when continuous phenomena are compressed into binary classifications, and how this mismatch affects the validity of clinical research and treatment decisions.

When estimating probabilities for an individual patient, surgeons must decide which reference class to draw from—but this choice is rarely straightforward and often formally intractable. Building on work by Reichenbach and Hájek, this project explores how expert clinicians navigate the reference class problem in practice.

The determination of when a fracture transitions from "not healed" to "healed" is inherently imprecise—yet it is routinely treated as binary in both research and clinical practice. This project examines how surgeons navigate this vagueness and whether the Sorites paradox (which arises when sharp boundaries are applied to vague predicates) is recognized or tolerated in orthopedic decision-making.