Cognitive Errors in Clinical Diagnosis
Framing and Diagnosis Momentum
Margaret A. Fitzgerald, DNP, FNP-BC, NP-C, FAANP, CSP, FAAN, DCC, FNAP
Clinicians process a steady flow of information, frequently under conditions of stress, fatigue, and time constraints. Given these challenges, communication between clinicians and healthcare workers can be a source of cognitive errors in clinical diagnosis, especially communication that takes place during patient handoffs.
Case
A 21-year-old woman is brought to the emergency department (ED) by emergency medical services (EMS) after experiencing anxiety, dyspnea, chest discomfort, and a brief loss of consciousness at work; EMS reports the patient’s main complaints as “anxiety and dyspnea in a patient with a history of panic attacks.” Over the past two years, the patient has presented three times to the ED with similar symptoms, which have been diagnosed as panic attack episodes. The triage nurse remembers the patient from her previous presentations and records the patient’s main complaint “as anxiety and dyspnea with chest pain” and notes her history of panic attacks.
In the ED, the patient reports she was sitting at her desk at work when she became anxious, felt tightness in her chest, and had trouble breathing. Her coworkers noted her condition and called 9-1-1 after she briefly lost consciousness. In the ED, she reports dyspnea and left chest discomfort, which is not reproduced with palpation but is made worse with deep breaths. Physical exam reveals the following: temperature, 98.4°F (36.9°C); heart rate, 112 beats/min; blood pressure, 108/70 mm Hg; respirations, 30 breaths/min; and oxygen saturation, 92% on room air. Body mass index is 31.1. Pulmonary exam is unremarkable and cardiovascular exam reveals tachycardia.
The patient denies any recent trauma. She has a history of anxiety and depression, for which she takes fluoxetine, and is also taking a combined oral contraceptive. She does not use recreational drugs or take supplements but has smoked a half pack of cigarettes daily for the past 5 years. She works in an office setting as a sales support specialist.
Chest x-ray, electrocardiogram, and routine blood testing are reported as unremarkable. The patient is given clonazepam for her anxiety and is told to return to the ED if her symptoms don’t improve.
Later that evening, the patient collapses at home and is brought back to the ED, where further testing is initiated, including D-dimer, which is elevated, and computed tomography angiography of the chest, which demonstrates pulmonary embolism involving the right posterior basal segmental pulmonary artery.
Discussion
Several cognitive biases contributed to the delayed diagnosis of pulmonary embolism in this patient. She had signs and symptoms of pulmonary embolism, including low oxygen saturation and tachycardia, but her presentation was framed in the context of her history of panic attacks, both by the EMS provider and then by the triage nurse, a bias known as framing effect. The framing effect is the tendency for the listener to be affected by how information is framed or presented or how a question is posed (eg, posed in terms of whether a patient might “die” or “live”).1,2 Framing can have considerable influence on a clinician’s diagnostic thinking when forming or adjusting diagnostic impressions. Both context and source of information (eg, trusted source, senior clinician) and context can influence framing.2 In this case the healthcare providers viewed the patient’s initial presentation through the lens of her mental health issues.
Framing the patient’s case with a diagnostic label (ie, panic attack) can lead to another cognitive bias, diagnosis momentum (or bandwagon effect), if the diagnostic label is transmitted repeatedly by all those providing care for the patient. With this bias, once a label (diagnosis) has been assigned, it sticks to a patient and momentum builds and interferes with clinicians’ ability to consider other alternatives. Patients receive their diagnostic “label” and their diagnosis is carried on from person to person unchallenged. Diagnosis momentum can affect both the work-up of the patient and how handoffs are framed.3
In this case, the providers who cared for the patient did not consider pulmonary embolism a diagnostic possibility on initial presentation, despite objective information that suggested this diagnosis. The patient’s history of smoking and combined oral contraceptive use along with the presence of low oxygen saturation and tachypnea should have led the clinicians to include pulmonary embolism in the differential diagnosis. Framing effect along with diagnostic momentum led them to focus on certain details while ignoring or failing to account for others. The clinicians should have paused and asked themselves whether there is something different in this case from the patient’s previous presentations.
Providers can avoid framing-related errors by avoiding diagnostic labels when framing challenging cases and by maintaining appropriate uncertainty about the cause(s) of a patient’s signs or symptoms.1 With regard to diagnosis momentum, providers must treat each encounter with a patient on its own; the presence of a prior diagnosis does not affect the possibility of the patient having another diagnosis.4
References
1. Etchells E. Anchoring bias with critical implications. PSNet Patient Safety Network. June 1, 2015. Accessed April 26, 2022. https://psnet.ahrq.gov/web-mm/anchoring-bias-critical-implications
2. The Joint Commission. Cognitive biases in health care. Quick Safety. Issue 28. October 2016. https://www.jointcommission.org/-/media/tjc/documents/newsletters/quick_safety_issue_28_oct_2016pdf.pdf
3. Croskerry P. 50 cognitive and affective biases in medicine. May 2013. CriticaThinking-Listof50-biases.pdf
4. Frye KL, Adewale A, Martinez CJ, Mora Montero C. Cognitive errors and risks associated with provider handoffs. Cureus. 2018;10(10):e3442. doi:10.7759/cureus.3442