The trusted source for
healthcare information and
A patient with a chief complaint of back pain also reported leg weakness to the emergency physician (EP) evaluating him, but the EP assumed the weakness was related to the back pain. As the patient was about to be discharged, a second EP learned that the patient was incontinent, but failed to question the previous assessment.
"Had he put two and two together, this now would have been back pain plus weakness plus incontinence — an easy call for cord compression," says Robert M. Wachter, MD, professor and associate chairman of the Department of Medicine at University of California — San Francisco, who reviewed the case as an expert witness.
The patient ended up paralyzed from the waist down, and sued the hospital and the involved EPs. "The case settled for hundreds of thousands of dollars," says Wachter.
Wachter says this case is a prime example of "anchoring" — a cognitive error that occurs when a physician latches on to a diagnosis and fails to consider alternatives.
"The primary techniques to avoid anchoring bias are known as meta-cognition’ — literally, thinking about your thinking," he says. This involves EPs asking themselves: "What is the worst thing this could be? If I come in tomorrow and find out I got the diagnosis wrong and the patient did poorly, what would the correct diagnosis be?"
Computerized decision-support tools can be helpful in alerting the EP to consider alternative diagnoses, says Wachter, "but the EP has to recognize his or her uncertainty."
To avoid anchoring bias, EPs must review the differential diagnosis and consider other possible diagnoses before coming to any conclusions, says Ken Zafren, MD, FAAEM, FACEP, FAWM, EMS medical director for the state of Alaska and clinical associate professor in the Division of Emergency Medicine at Stanford (CA) University Medical Center.
Here are some cases involving EPs that involved cognitive errors, including "anchoring," reviewed by Zafren:
• A 15-year-old boy was diagnosed with a cerebral arteriovenous malformation that was apparently an incidental finding on a non-contrast CT obtained to evaluate sore throat with neck pain. He was referred to a neurosurgeon, who saw him three days later and documented a normal neurologic examination.
Three days after seeing the neurosurgeon, the patient presented to an ED with the chief complaints of vomiting and possible aneurysm. The patient also complained of headache, neck pain, and unsteady gait.
The EP was aware of the abnormal CT, and documented that the patient had a stiff neck. "The EP diagnosed vomiting and gastroenteritis and discharged the patient to home, where he deteriorated and subsequently died from a ruptured arteriovenous malformation (AVM)," says Zafren. "The case was ultimately settled."
Zafren says the cognitive errors in this case include anchoring bias, premature closure, and failure to connect the patient’s symptoms to the recently diagnosed AVM.
• An 11-year-old girl was seen in the ED for abdominal pain, fever, and vomiting. "She was tender to palpation on abdominal examination. The EP diagnosed her with gastritis,’" says Zafren. The EP also diagnosed otitis media and treated the patient with amoxicillin, although the patient had no complaint of ear pain.
The patient’s abdominal pain worsened at home. The patient’s mother called the ED and was told to give the amoxicillin two days to work. The patient subsequently returned to the ED, where she was diagnosed with ruptured appendicitis.
"Her peritonitis was so severe that the right ovary had to be removed. The case was settled," says Zafren, adding that cognitive errors in this case include anchoring bias and premature closure.
"The EP testified at deposition that otitis media can cause abdominal pain," he says. Otitis media can be associated with fever and vomiting, Zafren notes, but is unlikely to cause abdominal pain with tenderness on exam. "Abdominal pain in a patient with an appendix should always prompt consideration of appendicitis."
• A 43-year-old man who was a T12 paraplegic with neurogenic bladder as the result of a motor vehicle crash many years previously, with a history of poorly controlled diabetes, presented to the ED with a two-day history of right flank pain and nausea, and was seen by a physician’s assistant (PA).
The patient’s urinalysis was consistent with infection and positive for glucose, and a fingerstick blood sugar was 358.
"The PA allegedly had a discussion’ with the EP. The PA diagnosed muscle spasm of the back and diabetes and did not prescribe antibiotics," says Zafren.
The patient was subsequently seen by a nurse practitioner (NP) at the office of his primary care physician, and was prescribed cotrimoxazole based on the positive urine culture that showed > 100,000 mixed flora. The patient had a second follow-up visit with his primary care physician.
"At this visit, the patient was febrile. The primary care physician diagnosed infected decubitus ulcers and prescribed cephalexin," says Zafren.
Three days later, the patient had a cardiac arrest at home. At autopsy, the patient had bilateral renocortical abscesses that grew the same organisms as the urine. "The cause of death was listed as bacterial sepsis due to chronic pyelonephritis," says Zafren. "The case against the PA was settled."
Zafren says cognitive errors in this case include anchoring bias and failure to assure proper follow-up. "The PA thought the back pain was musculoskeletal and communicated this to the primary care office," says Zafren. "This patient had a high likelihood of pyelonephritis, and should have been admitted considering his underlying comorbidities."
The NP had correctly diagnosed urinary tract infection, but, like the PA, ignored red flags for pyelonephritis.
"The primary physician compounded the problem by ascribing the fever to decubitus ulcers, by not repeating the urinalysis, and by not recognizing the likelihood of pyelonephritis as the source for the fever," says Zafren.
Anchoring bias is "one of the most common, powerful, and problematic" errors made by EPs, says Bruce Wapen, MD, a Foster City, CA-based emergency physician.
In one recent case, a 15-year-old girl reported the sudden onset of severe headache, vomiting, slurred speech, facial droop, and paralysis on the side of her body opposite from the side of the headache. There was no history of head trauma or loss of consciousness. She was airlifted to a stroke receiving center, where the flight nurse gave a verbal report to the triage nurse, which was entered into the ED’s electronic medical record. "By the time she arrived in the ED, her symptoms had pretty much resolved, except for waxing and waning facial asymmetry," says Wapen.
After the EP evaluated the patient, he documented that the patient had struck her head and was rendered unconscious, after which she experienced brief, one-sided weakness that was now resolved.
"Because the physician got, from somewhere, a history of head trauma and loss of consciousness, he thought the patient had a concussion and he anchored on that. He didn’t have [transient ischemic attack] or [cerebrovascular accident] in his differential," says Wapen.
The work-up consisted of a head and neck CT scan, which did not show an intracranial bleed or neck fracture, and the patient was discharged with a diagnostic impression of "concussion." The patient returned the next day to a different ED, and a CT angiogram showed evidence of an internal carotid dissection with brain infarction.
The diagnosis was missed because the first EP "anchored" on concussion and did a work-up appropriate for that problem, Wapen says.
Anchoring bias kept the first EP from making the correct diagnosis, says Wapen, "or at least having the correct diagnosis in the differential diagnosis, which should have led to additional imaging, admission to the hospital, and consultation with a pediatric neurologist."
For more information, contact: