From
Dr. Daniel Ostlie
Overcoming cognitive bias in evidence-based surgical practice
With Dr. Daniel Ostlie
Chapter 1 of 8
Introduction
Introduction: Evidence-Based Care vs. Personal Bias
Expert statements on this page
No expert statements were drawn from this page.
Host summaries · secondary, not cited in answers
Educational content from recorded physician discussions — not medical advice. Talk to your (or your child's) care team about your situation.
More from this expert
Dr. Daniel OstliePodcast
Management of spontaneous pneumothorax in pediatric patients
Daniel Ostlie · 33 min · Published Jul 2019
Video
Panel discussion on fibrinolytic therapy versus VATS for pediatric empyema management
Daniel Ostlie · 31 min · Published Apr 2017
Video
Choosing a clinical science and teaching career in academic medicine
Daniel Ostlie · 52 min · Published Sep 2012
What the experts said
In 1959, 1.4 million tonsillectomies were performed in the United States for pharyngitis.
There was never a study that showed tonsillectomy was an effective treatment for pharyngitis.
By 2006, 535,000 tonsillectomies were performed in the US, while the population had almost doubled since 1959.
Experienced clinicians perceived their patients did better after tonsillectomy and trained the next generation accordingly, perpetuating the practice.
Humans have an innate weakness in inductive reasoning and prefer deterministic thinking over probabilistic reasoning.
Experience is probably much less valuable than clinicians believe, and improving patient care requires rising above personal experience expectations through conscious effort.
By age 50, 50% of all men will have carcinoma in situ or prostate cancer; by age 80, 100% of US men will have prostate cancer, but most die of other causes.
Prostatectomy for prostate cancer was championed by thousands of urologists based on overwhelming single-institution case series (level three/four data), with randomized trials considered unethical.
Hundreds of thousands of patients underwent radical prostatectomy that never should have, with terrible effects on quality of life and no benefit in long-term survival.
When limited to pediatric surgery-specific journals, the rate of randomized controlled trials improved to 0.4%.
Hand washing hygiene to prevent disease progression was first published in 1843, yet hand hygiene compliance remains a daily challenge in hospitals.
The number of research articles has almost doubled annually over the last decade, and the number of clinical trials has increased fourfold.
From trial design to manuscript publication takes 59–65 months (approximately five years) on average.
There is a broad body of evidence that people do not learn from experience when it consists of a series of cases (level three or four evidence).
People will develop a deterministic rule to explain any series of observations rather than look for probabilism.
Increasing the complexity of a problem does not decrease the likelihood that people will apply a deterministic rule.
Even when specifically told an outcome is probabilistic, people still persistently develop deterministic rules.
A 2005 study of 62 groups of physicians and 59 studies found that 52% showed negative association between experience and knowledge, adherence to diagnostic standards, adherence to treatment standards, and patient outcomes.
In a study of 39,000 acute MI patients treated by 4,500 physicians, mortality increased 0.5% for every year since the doctor graduated from medical school.
A 1999 cardiac surgery study of 275 cardiologists and 83,000 operations found that more years in practice was significantly associated with higher mortality rates (P=0.001).
In a study of 1,033 hepatic resections, there was no difference in mortality based on surgeon volume or advanced training.
A New England Journal of Medicine RCT of 695 men with newly diagnosed prostate cancer (mean follow-up 6.2 years) found radical prostatectomy was 50% less likely to result in death from prostate cancer but showed no difference in overall survival.
Patients who had radical prostatectomy had much worse quality of life: erectile dysfunction was twice as high, urinary leakage three times as high, and dependence on diapers 15 times higher compared to watchful waiting.
A British Medical Journal study comparing Seattle and Connecticut Medicare data found PSA screening was five times more likely in Seattle and radical prostatectomy was done five times more often, but age-adjusted prostate cancer mortality showed no difference between the two regions.
A 1999 review of pediatric surgery journals from 1966–1998 found 34 randomized controlled trials out of 9,300 articles (0.3%).
A broader review of 80,000 articles in all journals over the same period found 134 randomized controlled trials, with only 16 (12%) studying surgical therapy.
A PubMed review from 1999–2012 identified 79 randomized controlled trials out of over 8,000 pediatric surgery publications, representing less than 1%.
After publication, it takes six years to get 25 citations in the literature.
Adoption into practice occurs at only 3.2% per year, taking at least 10–12 years to reach clinical guideline status and 17 years to reach 50% adoption. Virtually nothing ever reaches 80% adoption.
A New England Journal of Medicine study of surgical safety checklists in eight hospitals across eight countries found that a simple checklist decreased death rate from 1.5% to 0.8%.
The surgical safety checklist included: correct patient, mark the site, introduce team members, give antibiotics, have X-rays in the room, and state the operative plan.