Checking Your Cognitive Biases in the ER
David Klein, MD
Emergency clinicians are faced with a challenging task. During a busy shift, there can be significant pressure to evaluate and treat patients quickly, without giving each case the care it warrants. Thus, physicians often categorize patients early into a specific diagnostic pathway.
At some point, however, with each evaluation, we must pause to assess the individual and ensure we are providing not only efficient care, but effective care as well. Relying on diagnostic categories and past experiences can be useful. However, as I recently learned, it can also lead to careless errors.
No signs of infection?
I was working a shift recently when a 72-year-old female presented with abdominal pain. After an initial exam, which included a history and physical, I suspected a kidney stone. Subsequently, blood work, urinalysis and a CT scan were ordered. Blood work was normal; urinalysis did not reveal any signs of infection. And as suspected, the CT scan revealed a 2 ml UVJ stone. The patient was given pain meds and felt better and so, following normal protocol, I wrote up the discharge paperwork.
As the nurse began the discharge process, she noticed the patient had suddenly become tachycardic (up to 140 bpm). She informed me and I reassessed the patient. I considered common causes of tachycardia, including infection and pulmonary embolism, however, there was no evidence to support these etiologies, and her blood pressure was normal.
The tachycardia was sinus so I proceeded to treat her with a small amount of IV beta-blocker thinking that this episode was stress induced. Following this medication, her heart rate decreased to 105 and she was feeling fine, asking to go home. I discharged her to follow up with a primary care physician and cardiologist.
The next morning, I realized my mistake. It is not my usual practice to give a beta-blocker for asymptomatic tachycardia of unknown etiology. In reality, her tachycardia was a symptom of a larger issue and I failed to use it as a clue to uncover the cause. When I later called to check on her progress, I learned she had been admitted 48 hours later with an infected kidney stone. She ended up in the ICU for 24 hours with bacteremia and hypotension. She recovered and I am very lucky the story ended like this.
“Tachycardia is our friend.”
This is a phrase we have started repeating at our peer review meetings. I know that tachycardia is a classic sign of infection. But, because previous tests had shown no infection, I dismissed that as an option completely. This was uncharacteristic of me. What happened?
Check your cognitive biases
Our peer review process allows practitioners the opportunity to critique decisions and identify potential clinical errors made. However, to truly avoid future mistakes, we must understand why the errors were made in the first place. I believe the answer lies with cognitive bias. In this regard, there are a few things I should have been on guard for.
The first is anchoring, or the common human tendency to rely too heavily on the first piece of information offered. When the patient’s urine did not reveal an infection, I dismissed the notion that she could still be at risk for infection. She had a classic presentation of kidney stone and I was anchoring on that.
Another is availability, or relying on a past case or situation that was memorable, thus making it available in your mind, despite the differences it may have from the current case. As physicians, it is good practice to rely on lessons learned while making clinical decisions. However, we must also identify anomalies. Perhaps, here, I relied too heavily on past cases and failed to treat based on the facts in front of me.
When I treat patients, I like for all symptoms to line up – to make sense. In this case, tachycardia didn’t fit neatly into the mental box I had built and so – rather than address it as a symptom of a larger issue – I tried to explain it away. This is an example of confirmation bias, or ignoring contradictory data to make the pieces of the puzzle fit neatly into a presumed picture.
Finally, there is momentum. Diagnosis momentum is accepting a previous diagnosis without sufficient skepticism. This can lead to discharge momentum when a clinician is likely to continue discharging the patient – despite conflicting information becoming available. Once I had evaluated the patient and confirmed my diagnosis, I didn’t deviate from my plan. I had begun the discharge procedure and began treating new patients. In retrospect, admitting her to observation care may have been the better choice.
Maintaining Effective Care
So now that we know what some common cognitive biases are, how can we avoid falling victim to them in the future? What can we do to ensure our patients remain safe?
In our line of work, it can be challenging to find a few minutes to regroup. However, it is incredibly important to break the momentum – and consider the patient’s specific story.
We created the Discharge Time Out to give our busy practitioners the opportunity and permission to take a step back, review the chief complaint and revaluate the differential diagnosis. Other methods include one minute in the break room to physically step away from the monotonous routine, listing the biases we’ve reviewed as a reminder to combat them, and considering the observation unit as a tool to gather more information before deciding to discharge or admit.
What are some other ways we can challenge ourselves to avoid these pitfalls? Let’s hold each other accountable to not only offer efficient care but appropriate care as well.
Return to Newsletter