This edition of The Mutual's "The Verdict Is In" was written by Dr. Gary Zaid, a Vituity partner and Mutual physician educator, who feels passionate about enhancing patient safety. Answer three short questions about the article to receive CME as well as be entered into The Mutual's quarterly $300 visa card raffle.

Cognitive Errors in Medicine

Medicine as a vocation has several well-known forms of cognitive bias that contribute to diagnostic and treatment errors. Dr. Jerome Groopman's book, "How Doctors Think", looks in depth at the common cognitive errors healthcare providers make. A series of critical reviews over the years have also identified strategies physicians and advanced providers can use to reduce the impact of these cognitive biases in medical decision making. Recognizing and working to prevent the potential for these errors can help avoid untoward outcomes and reduce liability exposure.

These are some of the common cognitive errors:

Anchoring Bias: Anchoring bias refers to the tendency to fix on, or anchor, to the first symptom or bit of data and fail to consider the full range of causes. Anchoring leads to a misdiagnosis and constrains further investigation.

Example: A 55-year-old male from Northern Wisconsin presents with new onset joint pain over the past several weeks. He is an avid mountain biker on local trails. X-rays obtained showed mild DJD and his vitals were normal. He was diagnosed with osteoarthritis and discharged on NSAIDs. Several weeks later the patient is seen by another provider with worsening symptoms and tests positive for Lyme Disease. The fixation on the X-ray led to a failure to further investigate the new onset of arthritis in a patient where Lyme disease was prevalent. Similar errors can occur by anchoring on a single abnormal lab test or study.

Availability Bias: Availability bias is the trap of falling into a recent dramatic or an unusual case that biases our thought process and limits the differential diagnosis and investigation.

Example: A provider recently diagnosed a patient with acute cauda equina syndrome. A new patient presents with acute numbness in both lower legs and thoracic and lower back pain. MRI of the thoracic and lumbar spine were ordered. Other etiologies were not considered including vascular causes. The patient was eventually diagnosed with a Type 2 Aortic Dissection.

Attribution Bias: Attribution bias often arises from both negative and positive stereotypes that can lead to cognitive error.

Example: A suspected homeless patient presented with weakness which was is attributed to poor nutrition and substance abuse. The work up was limited, and an acute adrenal crisis was missed. The patient had been unable to fill their Rx of medications for Addison's disease.

Confirmation Bias: Confirmation bias involves ignoring or rationalizing contradictory data to make the pieces of the puzzle fit neatly into the presumed picture. This often involves cherry picking data that involves selectively accepting or ignoring information.

Example: A 50-year-old dialysis patient presented with atypical chest pain similar to his prior "indigestionc. The EKG appeared unchanged from prior, and a High Sensitivity-Troponin was elevated at 225. The chest pain resolved with a "GI-Cocktail" and the elevated troponin was attributed to "Troponemia" from their renal failure. A second troponin was never ordered, and the patient was discharged. The patient returned 24 hours later in cardiac arrest.

Satisfaction of Search: Satisfaction of search is to try to find a single explanation for a myriad of symptoms. Although a single cause may occur in some circumstances, this is not always the case.

Representative Bias: "Representativeness" or "prototype error" occurs when a case is not typical and may not be a consideration to the clinician's pattern of thinking. Atypical presentations of appendicitis or cardiac ischemia are good examples.

Example: A 30-year-old presented with R flank pain and fever. A urinalysis was positive for only leukocytes and the patient was discharged on antibiotics. Several days later the patient presented with generalized abdominal pain and was diagnosed with a perforated retrocecal appendicitis. Appendicitis was not considered given the flank pain and positive urine leukocytes. The clinician was unaware 20% of patients with appendicitis will have WBCs in the urine from irritation of the ureter. The typical presentation of right lower quadrant pain does not always occur in acute appendicitis.

Triage Bias: Triage bias is a risk especially in Emergency Medicine. It occurs when clinicians rely on limited information provided by prehospital providers or other third parties and can lead to delayed or limited diagnostic evaluation.

Example: It was Saturday night in a busy emergency department and multiple EMS patients arrived in a short period of time. EMS transported a somnolent 40-year-old patient and reported the patient was found "drunk" on the ground outside a local drinking establishment. Glucose was normal, vital signs were reasonable, and the patient appeared intoxicated. After the clinician "eyeballs" the patient on the EMS stretcher the staff moved the patient to an observation area. Three hours later it was noted the patient was obtunded and now had unequal pupils. CT scan identified a large subdural hematoma with midline shift. The clot was evacuated, and the patient had a very prolonged recovery.

The Best Strategy to Incorporate in Practice:

Considered the most successful strategy, guided reflection intervention involves conscious consideration of the information available and other potential outcomes than what automatically comes to mind. For this reason, Vituity and The Mutual recommend incorporating a cognitive pause at key decision points in medical decision making.

Answer three quick questions in SurveyMonkey
via this link for CME and a chance to win our $300 quarterly raffle.

Survey Questions

Congratulations to Dr. Vernell Smith from San Mateo Medical Center who won last quarter’s CME raffle!

References