Ascertainment bias: occurs when a physician’s thinking is shaped by prior expectation such as stereotyping or gender biases.

Example: A patient known to be a drug addict is found lying on the hospital floor unconscious; the staff assumes that he has had a drug overdose and does not consider other possibilities. He is in fact a diabetic who had insulin-induced hypoglycemia.

Availability: the disposition to judge things as being more likely, or frequently occurring, if they readily come to mind. Thus, recent experience with a disease may inflate the likelihood of its being diagnosed. Conversely, if a disease has not been seen for a long time (is less available), it may be underdiagnosed.

Example: A resident sees an elderly patient with mild joint pain having just returned from a noon conference on Systemic Lupus Erythematosis (SLE). The resident orders an extensive batter of tests for this disease despite the fact that it is far more likely that the patient has osteoarthritis.

Base rate neglect: the tendency to ignore the true prevalence of a disease, either inflating or reducing its base rate, and distorting Bayesian reasoning. However, in some cases, clinicians may (consciously or otherwise) deliberately inflate the likelihood of disease, such as in the strategy of “rule out worst case scenario” to avoid missing a rare but significant diagnosis.

Example: in the previous example, the resident worries that he or she may “miss something important” and pursues the SLE diagnosis despite the unlikelihood that the patient has the disease.

Fundamental attribution error: the tendency to be judgmental and blame patients for their illnesses (dispositional causes) rather than examine the circumstances (situational factors) that might have been responsible. In particular, psychiatric patients, minorities, and other marginalized groups tend to suffer from this bias. Cultural differences exist in terms of the respective weights attributed to dispositional and situational causes.

Example: a schizophrenic patient who is delusional thinks that he is being poisoned and refuses to drink any liquids. In the ED the staff members “blame” him saying that he “did this to himself.” Their attitude is reflected in their care; Rather than put in an IV, which they would normally do to a typically dehydrated patient, they keep telling the patient to drink, which, of course, he is not willing to do. They do not consider that his failure to drink is a reflection of a medial condition rather than a volitional act of defiance.

Gambler’s fallacy: the belief that if a coin is tossed ten times and is heads each time, the 11th toss has a greater chance of being tails (even though a fair coin has no memory). That is, the physician allows the pretest probability that a patient will have a particular diagnosis to be influenced by preceding but independent events.

Example: a physician sees a series of patients with chest pain in the emergency department and (correctly) diagnoses all of them with an acute coronary syndrome. When the next patient walks in complaining of chest pain, the physician assumes that the sequence (of acute coronary syndrome) will not continue.

Playing the odds: to opt for a benign diagnosis even though it may be less likely. In equivocal or ambiguous presentations, providers may focus on a benign diagnosis on the basis that it is more likely than a serious one (which it may or may not be). This tendency may be compounded by the fact that the signs and symptoms of many common and benign diseases are mimicked by more serious and rare ones. The strategy may be unwitting or deliberate and is diametrically opposed to the rule out worst case scenario strategy (see base rate neglect).

Example: A patient who is a long-time smoker develops a cough which is productive of blood-streaked sputum without any fever, and weight loss. The provider pursues the diagnosis of pneumonia (benign) thinking that it is more common than lung cancer. Given the absence of fever and weight loss, the post-test probability is much higher for cancer than pneumonia.

Posterior probability error: when a physician’s estimate for the likelihood of disease is unduly influenced by what has gone on before for a particular patient. It is the opposite of the gambler’s fallacy in that the physician is gambling on the sequence continuing.

Example: After a patient presents to a physician five times in a row with a headache that is correctly diagnosed as migraine on each visit, there is a tendency to diagnose migraine on the sixth visit. The fact is that the headache could represent a different condition which, though less common than a migraine (such as a tumor) should be considered. Common things for most patients continue to be common, and the potential for a non-benign headache being diagnosed is lowered through posterior probability.

Mr. S. is a 33 year old homeless man with a history of frequent, intractable migraine headaches and chronic opioid abuse who presents to the ED with moderately intense headache for the past 2 hours. Upon arrival to the ED, he is alert and oriented, has elevated BP but otherwise stable vital signs, has a "non-focal" neurological exam, and is demanding "oxycontin.” As Mr. S. is a "frequent flyer" in the ED, the housestaff believe he is (yet again) drug seeking, starts him on a non-opioid medication for his pain, wheels him to a back hallway in the already overcrowded ED, and begins coordinating with social work for placement in a homeless shelter. When housestaff returns to see him four hours later, he is found unconscious and unresponsive with unequal and unreactive pupils. Subsequent non-contrast CT scan shows a suspected intracranial bleed. The patient expires within the next 30 minutes. [Contributed by Michael Cutalo, P&S Class of 2009]

Anchoring: This is the tendency to lock onto salient features in the patient’s initial presentation too early in the diagnostic process and failing to adjust this initial impression in the light of later information. This CRB may be severely compounded by the confirmation bias.

Example: A patient starts an interview by complaining of abdominal pain which directs the physician to the gastrointestinal system. Only at the end of the interview does the patient complain of claudication (suggesting an aneurysm) which the physician discounts because it does not fit into a diagnosis related to the gastrointestinal tract.

Premature closure: the tendency to apply premature closure to the decision-making process, accepting a diagnosis before it has been fully verified. The consequences of the bias are reflected in the maxim: “When the diagnosis is made, the thinking stops.”

Example: A college football player comes to the ED complaining of chest pain. He notes in the history that he played in a game the day before and was tackled many times in the chest. He is sent for an x-ray which the radiologist notes “questionable rib fracture” which is not definitive, though. The ED resident is satisfied with diagnosis of rib fracture and does not pursue other possibilities. He fails to ask the patient about other symptoms such as hematuria (which the patient has) and misses the kidney contusion (with referred pain to the chest).

A 50 year-old woman with no past surgical history arrives in the ER with severe right upper quadrant pain and is diagnosed with acute cholecystitis. She is rushed to the OR where the anesthesiologist performs a rapid sequence induction of general anesthesia (with succinylcholine and propofol). The anesthesiologist switches on desflurane for maintenance of anesthesia. The procedure runs smoothly for 15 minutes, when the anesthesiologist notices that the patient is becoming tachycardic and hypertensive, the expiratory CO2 is rising, and the core body temperature is heating up. Additionally the patient can be seen to have some muscle tone. The anesthesiologist sees the elevated sympathetic drive as a sign of pain and the presence of muscle tone as the succinylcholine wearing off, both routine occurrences, and administers fentanyl and vecuronium accordingly. The elevated CO2 and body temperature are ignored because they o not fit his diagnoses. In 15 minutes the patient's temperature has risen to 110 degrees, has become tachycardic to 130 and exhibits fully rigid skeletal muscle tone. Finally desflurane is shut off and dantrolene administered, though the diagnosis of malignant hyperthermia should have been identified and managed earlier. [Contributed by Kaliq Chang, P&S 2009]

Representativeness restraint: the tendency to look for prototypical manifestations of disease: “If it looks like a duck, walks like a duck, and quacks like a duck, then it is a duck.” Yet, restraining decision making along these pattern recognition lines leads to atypical variants being missed.

Example: A patient with coronary ischemia comes to the ED complaining of retrosternal chest pain. The ED physician asks the patient about the presence or absence of other symptoms usually associated with coronary ischemia such pain radiating down the arm, the presence of risk factors, or a prior history of exertional chest pressure. The patient denies all of the associated symptoms steering the resident to explore a gastrointestinal basis of the chest pain when in fact the patient is experiencing a myocardial infarction.

Search satisfying: reflects the universal tendency to call off a search once something is found. Co-morbidities, second foreign bodies, other fractures, and co-ingestants in poisoning may all be missed. Also, if the search yields nothing, diagnosticians should satisfy themselves that they have been looking in the right place. [Contributed by Marc Manseau P&S 2009]

Example: A 5 year-old girl with no reported past medical history is brought into the emergency room by her mother on a Sunday evening for evaluation of a "black eye." The mother reports that on Saturday afternoon, her daughter was running through their apartment, tripped over the family dog, and hit her right eyebrow on the corner of the coffee table. On physical exam, the girl is withdrawn, and hesitant to allow you to touch her anywhere, including her right eyebrow. Her right eyebrow is swollen and exquisitely tender to palpation, and her right orbit is swollen, with multi-colored purpura throughout. No hematomas or bone fragments are appreciated. Her right eyelid is swollen shut. A CT scan shows a non-displaced fracture in her right frontal bone, at the supra-orbital notch. She is given a prescription for acetaminophen with codeine for pain control, told to apply ice to her right eyebrow, and given an appointment for outpatient follow-up on Wednesday. The next day (Monday), the girl is sent to the emergency room directly from kindergarten, after she complained of back pain, and the school nurse found multiple bruises in various stages of healing on her back and legs. Child abuse is suspected, and an x-ray skeletal survey shows multiple fractures in all four extremities in multiple stages of healing.

Feedback sanction: a form of ignorance trap and time delay trap. Making a diagnostic error may carry no immediate consequences, as considerable time may elapse before the error is discovered, if ever, or poor system feedback processes prevent important information on decisions getting back to the decision maker. This is especially true when a patient is incorrectly treated and goes to another healthcare provider for a second opinion. The first provider does not learn from the misdiagnosis experience.

Example: A patient is treated by his provider with a new antihypertensive. Upon taking the drug at home, the patient has a serious hypotensive episode, collapses and is brought to the local ED. This same response is seen in the providers next several patients without his knowing it because the patients keep being taken to the local ED and they never return. Eventually, someone in the ED calls the provider to inform him of the recurring hypotensive complications.

Hindsight bias: knowing the outcome may profoundly influence the perception of past events and prevent a realistic appraisal of what actually occurred. In the context of diagnostic error, it may compromise learning through either an underestimation (illusion of failure) or overestimation (illusion of control) of the decision maker’s abilities. In “reconstructing the past” there is a tendency to create a narrative that is the most favorable rendition and compare the actual outcome to the idealized outcome.

Example: A patient with atrial fibrillation is given standard (evidence-based) therapy, amiodarone, and has a very rare reaction (Stevens-Johnson syndrome, less than 2%) requiring hospitalization. The provider thinks “If I had given propranalol instead, the patient would have been fine.” What the provider does not consider is that the “experiment” of placing the patient on propranalol had not been performed and that this drug has its own risks. For example, it is possible that the patient, placed on propranalol would have suffered a very low ventricular rate (AV block) and become hypotensive requiring hospital admission.

Aggregate bias: when physicians believe that aggregated data, such as those used to develop clinical practice guidelines, do not apply to individual patients (especially their own), they are invoking the aggregate fallacy. The belief that their patients are atypical or somehow exceptional may lead to errors of commission, e.g., ordering x rays or other tests when guidelines indicate none are required.

Conflicts of interest (industry-wide): evidence from industry-sponsored clinical studies is, by definition, biased. Research has demonstrated that conclusions of these studies overestimate benefit and underestimate risk.

Conflicts of interest (personal): when a physicians own interests, either financial, scientific or medical, interfere with judgement.

Commission bias: results from the obligation toward beneficence, in that harm to the patient can only be prevented by active intervention. It is the tendency toward action rather than inaction. It is more likely in over confident physicians. Commission bias is less common than omission bias. [submitted by Nancy Chen, P&S 2009]

Example: A 65 year old previous marathon runner presents with 2 weeks of pain in his right knee that worsens with activity. He reports running approximately 3-5 miles a day and this knee pain is associated with weakness making it impossible to run. Physical exam reveals a mildly swollen, non-erythematous, right knee with point tenderness just inferior to the patella and pain with leg extension. The renowned Sports Medicine physician he visits recommends an MRI of his right knee and an arthroscopy.

Framing effect: how diagnosticians (and patients) see things may be strongly influenced by the way in which the problem is framed, e.g., physicians’ perceptions of risk to the patient may be strongly influenced by whether the outcome is expressed in terms of the possibility that the patient might die or might live. In terms of diagnosis, physicians should be aware of how patients, nurses, and other physicians frame potential outcomes and contingencies of the clinical problem to them.

Omission bias: the tendency toward inaction and rooted in the principle of non-malfeasance. In hindsight, events that have occurred through the natural progression of a disease are more acceptable than those that may be attributed directly to the action of the physician. The bias may be sustained by the reinforcement often associated with not doing anything, but it may prove disastrous. Omission biases typically outnumber commission biases.

Overconfidence bias: a universal tendency to believe we know more than we do. Overconfidence reflects a tendency to act on incomplete information, intuitions, or hunches. Too much faith is placed in opinion instead of carefully gathered evidence. The bias may be augmented by both anchoring and availability, and catastrophic outcomes may result when there is a prevailing commission bias. [Contributed by Maya Hughes, P&S 2009]

Example: The patient was a 5 month old ex FT baby girl who was transferred from an OSH for further work up of failure to thrive, chronic cough, and new oxygen requirement. The patient was born without complications to consanguineous parents. She grew at an appropriate rate and had no known infections until age 4 months when she developed a dry cough, blood streaked stool and decreased po intake. She was evaluated by her PMD who presumed the cough was secondary to a viral URI, the blood streaked stool was secondary to a minor anal laceration and the poor po intake was the result of her acute infection. However, 2 weeks later, the patient’s cough persisted, her po intake remained poor, and her weight decreased by one lb. Her parents were considered that a more serious illness may have been afflicting their daughter; however, the PMD continued to defend his original diagnosis. During the following 2 weeks, the patient’s cough worsened and she exhibited increased work of breathing. Her po intake remained poor and she lost an additional 1 lb. She was subsequently evaluated by another PMD who was suspicious of failure to thrive and dehydration. As a result, she was transferred to an OSH for further work up. Extensive studies were performed including bronchoscopy (due to persistent cough, increased work of breathing and new oxygen requirement), which revealed bronchial lining replete with macrophages and foamy debris c/w PCP (Silver stain was positive for PCP, too). She was transferred to CHONY, where a T cell panel was performed and notable for 0% T cells. (Patient HIV negative). Patient was diagnosed with Severe Combined Immunodeficiency Disease (SCID) and underwent bone marrow transplant.

Outcome bias: the tendency to opt for diagnostic decisions that will lead to good outcomes, rather than those associated with bad outcomes, thereby avoiding chagrin associated with the latter. It is a form of value bias in that physicians may express a stronger likelihood in their decision making for what they hope will happen rather than for what they really believe might happen. This may result in serious diagnoses being minimized.

Sunk costs: the more clinicians invest in a particular diagnosis, the less likely they may be to release it and consider alternatives. For the diagnostician, the investment is time and mental energy and, for some, ego may be a precious investment. Confirmation bias may be a manifestation of such an unwillingness to let go of a failing diagnosis.

Example: A patient is admitted to the hospital for a work-up of hematuria. The provider arranges for imaging studies (CAT scan) which identifies a possible small mass in one of the kidneys. Additional imaging (MRI) shows the same and he mobilizes the urology team to schedule removal of the kidney. The patient reads about minimally invasive surgery on the Web and pressures the primary provider to seek numerous second opinions to perform the surgery. After much ado, the surgery is schedule. On the morning of surgery, the resident reminds the primary provider that the patient had heavy proteinuria which is not consistent with a renal mass and justifies performing a renal biopsy instead of a partial nephrectomy. The provider’s initial response is to try to explain away the proteinuria (which he cannot). With much courage he cancels the operation until the biopsy is performed.

Million dollar workup: when patients have been subjected to exhaustive and unavailing diagnostic investigations, they are said to have received the million dollar workup (MDW). The MDW is the tendency to believe that nothing further can be done to throw light on the dark place where, and if, any definitive diagnosis resides for the patient, i.e., the physician is let out of further diagnostic effort. This may prove ultimately to be true, but to adopt the strategy at the outset is fraught with the chance of a variety of errors.

Confirmation bias: the tendency to look for confirming evidence to support a diagnosis rather than look for disconfirming evidence to refute it, despite the latter often being more persuasive and definitive.

Example: a young man complains of retrosternal chest pressure on exertion and is found to have a sore rib physical examination which does not quite match the pain. He is sent for an x-ray of the ribs (confirmation of the tentative diagnosis) rather than for an ECG (confirm or eliminate an alternative diagnosis).

You are covering calls from a college health clinic at 11:30 pm when you receive a call. A student is calling regarding his friend, V.C., who is a 22 year old male in his final semester of college. The student states that V.C. seems “weird” and “like, out of it.” Upon further questioning, you find out that V.C. has had five drinks in the last 2 hours with his football buddies, whom he lives with in a dorm. His friend says this is strange because he usually drinks much more. Though he does not drink every day, when he does drink he consumes 10-15 drinks on average. You continue to question the friend, and find out that V.C. is lying on a couch, somnolent, irritable, and has a slight headache. He has vomited once. At this point, you conclude that given the fact that he binge drinks, has vomited, and he probably drank too much, he is simply drunk. You consider that he might have meningitis because of a recent outbreak in a different location in the state. You send of for an alcohol level from the lab to confirm your diagnosis. You ignore the irritability and somnolence, attributing it to the alcohol. You schedule the student to come in the following week for alcohol counseling. The student did have meningitis, in the early stages. Although this may not have been 100% obvious, red flags that were ignored include somnolence, irritability, headache, and college dorm. The test that should have been ordered is one that rules out meningitis, a lumbar puncture. [Contributed by Cameron Marshall, P&S 2009] Note, this is also an example of “premature closure.”

Sutton’s slip: takes its name from the apocryphal story of the Brooklyn bank robber Willie Sutton who, when asked by the Judge why he robbed banks, is alleged to have replied: “Because that’s where the money is!” The diagnostic strategy of going for the obvious is referred to as Sutton’s law. The slip occurs when possibilities other than the obvious are not given sufficient consideration.

Diagnosis momentum: once diagnostic labels are attached to patients they tend to become stickier and stickier. Through intermediaries (patients, paramedics, nurses, physicians), what might have started as a possibility gathers increasing momentum until it becomes definite, and all other possibilities are excluded.

Example: a patient is admitted in the emergency room complaining of chest pain on inspiration and fever and diagnosed with pneumonia. The resident in the ED looks at the chest x-ray and though seeing only non-specific changes (not unambiguously indicative of pneumonia), treats the patient for pneumonia and does not pursue additional diagnostic possibilities. The resident transfers the patient to the floor where the new resident accepts the diagnosis of the ED resident without considering other possibilities. The next morning, in the light of day, it is obvious that the patient is jaundiced and that the pain on inspiration was coming from the gall bladder (which is infected).

Order effects: information transfer is a U function: we tend to remember the beginning part (primacy effect) or the end (recency effect). Primacy effect may be augmented by anchoring. In transitions of care, in which information transferred from patients, nurses, or other physicians is being evaluated, care should be taken to give due consideration to all information, regardless of the order in which it was presented.