Cognitive Biases and Heuristics in Surgical Settings - A Systematic Review
Systematic Review of Cognitive Biases and Heuristics in Surgical Settings
Executive Summary
This briefing document synthesizes findings from a 2025 systematic review published in Annals of Surgery regarding the prevalence and impact of cognitive biases and heuristics in surgical environments. Cognitive biases—systematic errors in thinking—and heuristics—mental shortcuts—are pervasive in surgical practice, contributing significantly to medical errors, "never events," and negative patient outcomes, including mortality.
The analysis identified 38 distinct biases across 21 primary research studies. The most frequently represented biases include confirmation bias, anchoring, risk aversion, and overconfidence. Notably, a quality improvement study of over 5,000 operations found that 21% of adverse events were attributable to cognitive biases. While surgical expertise provides some protection, it does not immunize practitioners against these mental shortcuts, particularly under stressful or high-pressure conditions. The document concludes by outlining essential debiasing strategies, ranging from mindfulness and deliberate reflection to the implementation of "forcing functions" like surgical checklists.
Methodological Overview
The systematic review followed PRISMA guidelines and searched five major databases (Cochrane, Embase, MEDLINE, Web of Science, and SCOPUS) from inception through January 2024.
Study Composition: The 21 included papers comprised 11 surveys (52%), 6 experiments (29%), and 4 retrospective or prospective data analyses (19%).
Specialties Covered: General surgery and pan-specialty studies were most common (23%), followed by orthopedics (19%), neurosurgery, and colorectal surgery.
Risk of Bias: 33% of studies were low risk, 48% moderate risk, and 19% high risk. The high-risk studies often utilized non-validated surveys.
--------------------------------------------------------------------------------
Taxonomy of Cognitive Biases in Surgery
In the absence of a universally accepted taxonomy, the review categorizes biases based on how they mishandle information across three domains: Availability of Information, Information Processing, and Memory.
Detailed Analysis of Key Biases and Impacts
1. Availability of Information (AoI)
These biases lead to flawed reasoning due to inaccurate judgments of information relevance.
Anchoring: Identified as the second most common bias. It is heavily associated with management and therapeutic errors. One analysis of 655 general surgery cases linked anchoring to severe complications and increased patient harm; another study of 736 cases associated it with fatal outcomes.
Overconfidence Bias: Frequently identified in cases with complications. In a survey of 242 orthopedic surgeons, 83% rated themselves as "above average" diagnosticians, while 0% rated themselves as "below average." This bias has a statistically significant association with management errors and mortality.
Ascertainment Bias: Results were mixed; while one study found it to be the most common bias in cases with complications, another did not find a statistically significant probability of it causing severe harm.
2. Information Processing (IP)
These biases arise from the brain's reliance on automatic processes during high-stakes, rapid decision-making.
Risk Aversion: The most common IP bias. It significantly influences surgical decisions, such as the choice to create a defunctioning stoma in rectal surgery. Interestingly, surgeons over age 50 were found to be less risk-averse than their younger counterparts.
Commission and Omission Bias: Commission bias (the tendency toward action) was associated with a >20% increase in the incidence of severely harmed patients. Omission bias (the tendency toward inaction) was correlated with statistically significant increases in management errors and instances of anastomotic failure.
Inattention Blindness: Two experimental studies demonstrated that surgeons often miss information outside their immediate "tunnel of focus." Increased cognitive load further exacerbates this blindness.
Perceptual Bias: Both novice and experienced surgeons were found to underestimate the size of incisions in simulated tasks, though experience offered some protection as visual challenges increased.
3. Memory Biases
These systematic distortions affect the recollection of past events, leading to inaccuracies in current decision-making.
Confirmation Bias: The most frequently identified bias across all categories (7 studies). In a prospective analysis of 1.5 million procedures, confirmation bias accounted for 82% of all identified perceptual errors and was a major contributor to "never events."
Self-Evaluation Bias: General surgery interns were found to generally underestimate their technical skills, with high performers assessing themselves more negatively than their actual scores warranted.
Hindsight Bias: Frequently identified in general surgery case analyses, though not statistically linked to a higher probability of severe patient harm in the reported data.
Impact on Patient Safety
The review establishes a clear link between cognitive biases and compromised patient safety.
Adverse Events: Cognitive biases were identified as the cause of 21% of adverse events in a large-scale quality improvement study.
Mortality: Anchoring and overconfidence bias are specifically cited as being associated with fatal outcomes.
Never Events: Confirmation bias is a primary driver of "never events"—serious incidents that are entirely preventable through standard protocols.
Mitigation and Debiasing Strategies
The review highlights that simply acknowledging bias is insufficient. A multi-pronged approach involving workplace and educational strategies is required.
Workplace Strategies
Effective Teamwork: Building and maintaining cohesive teams helps mitigate the impact of anchoring and confirmation bias.
Surgical Pauses: Implementing deliberate pauses can reduce the errors associated with rapid information processing and fatigue.
Artificial Intelligence (AI): Machine learning and deep learning models can augment decision-making by using intraoperative data to decrease human cognitive bias. However, care must be taken to ensure AI does not inherit or replicate human biases.
Educational and Metacognitive Strategies
Awareness Training: Promoting the message that biases affect all surgeons, regardless of skill level, is essential. The Royal College of Surgeons of England currently offers modules on this topic.
Metacognition: Strategies such as "thinking aloud," deliberate reflection, and "devil’s advocate" exercises can reduce errors.
Mindfulness: One prospective pilot study showed that neurosurgeons who completed an 8-week mindfulness intervention made significantly fewer errors related to inattention blindness.
Forcing Functions
Checklists: Validated and standardized surgical checklists, adapted for specific procedures, act as a critical safety net to prevent biases from progressing into errors.
Thinking Prompts: Using specific prompts during the decision-making process to encourage the consideration of alternative diagnoses or risks.
Conclusions and Future Directions
Cognitive biases and heuristics are not specialty-specific; they consistently impact surgical decision-making across preoperative, intraoperative, and postoperative phases. Because these biases contribute to fatal outcomes and preventable never events, the integration of debiasing strategies into routine training is a clinical necessity.
Future research must move toward more experimental studies and objective measures to quantify the causal links between specific biases and negative outcomes. Additionally, the development of standardized, validated checklists across all stages of surgical care is recommended to mitigate the inherent limitations of human cognition in high-pressure medical environments.
