Cognitive Errors in Clinical Diagnosis: Representativeness

Pattern recognition is one of the many cognitive heuristics, or short-cuts, clinicians use to make the complex process of clinical decision-making more manageable. Faced with a swell of information and needing to make multiple decisions daily, often under stressful conditions, clinicians use heuristics to save time and effort. In this way, pattern recognition is essential to clinical reasoning and practice. However, like all heuristics, pattern recognition, when not accompanied by an analytic approach that overrides intuitive thinking, can misdirect clinical reasoning and lead to diagnostic error.

CASE

Office Visit

A 52-year-old woman presents to her clinician’s office with burning pain in the abdomen-epigastric region that began approximately 24 hours ago. The pain has grown progressively worse, with no associated, aggravating, or relieving factors noted. The patient reports some dizziness and nausea, and notes that she has felt fatigued for the past several weeks but denies shortness of breath.

Past history includes gastroesophageal reflux disease (GERD) diagnosed at age 47 and hypertension diagnosed at age 45. GERD was treated with lifestyle changes and short-term proton pump inhibitor (PPI) therapy, and is now managed with over-the-counter antacid tablets as needed. Assuming the pain was related to her GERD, the patient took antacid tablets and an over-the-counter PPI, which did not provide significant relief. She is taking chlorthalidone 25 mg once daily for hypertension, although her blood pressure is not under optimal control. Her last lipid panel 2 years ago showed borderline high low-density lipoprotein cholesterol (155 mg/dL) and triglycerides (187 mg/dL). Her clinician discussed prescribing a statin as an option for lowering her cholesterol at that time, but the patient wanted to try lifestyle changes before starting drug therapy.

The patient is a former smoker (25 pack-years), having quit 3 years ago as part of her GERD management lifestyle changes. Since quitting smoking, she has gained approximately 15 lb. Physical examination shows a blood pressure of 137/80 mmHg, heart rate of 64 bpm, BMI of 29.1 kg/m2, and waist circumference of 33.9 inches.

Given the patient’s history of GERD, the clinician diagnoses a GERD flare, prescribes PPI therapy, and advises the patient to call the office if her symptoms do not improve on longer-term PPIs or if they worsen.

Emergency Department Visit

The following morning, the patient presents to the emergency department with continued burning pain in the abdomen-epigastric region and new-onset shortness of breath. Electrocardiography shows nonspecific ST segment changes and she is admitted to the cardiac observation unit, where myocardial infarction (MI) is confirmed with elevated troponin levels.

DISCUSSION

The patient in this case presented with typical signs of acute coronary syndrome in women and common risk factors for coronary heart disease. The clinician did not identify the signs of MI, as they did not fit the prototypical or textbook combination of features of chest pain, radiating pain down the arm, shortness of breath, and sweating associated with MI, which is seen most often in men. The application of the cognitive heuristic representativeness, which leads the clinician to look for prototypical manifestations of disease, resulted in the delayed diagnosis. Specifically, with the representativeness heuristic, the clinician assumes that something that seems similar (or dissimilar) to other things in a certain category is itself a member (or not a member) of that category.1,2 The clinician in this case deemphasized evidence because it did not strongly align with a category they’ve encountered or learned about—the typical manifestations of MI described in the literature based on studies involving mostly men—and adhered to a past diagnosis.

It is well documented that the early diagnosis of acute MI is missed more frequently in younger women (<55 years) than in men because of sex differences in clinical presentation among patients with acute coronary syndromes, with the prototypical presentation reflecting what has been observed in men.3,4 A recent multicenter, observational cohort study of 3,501 young adults hospitalized with acute MI concluded that women showed significantly more variation in unique MI symptom phenotypes than men.5

In many situations, the representativeness heuristic produces accurate results because representativeness often correlates with the likelihood of a disease or condition, but it also leads clinicians to give too much weight to highly representative evidence and to undervalue other relevant options, resulting in missed diagnoses with atypical presentations.2 The patient in this case presented with typical symptoms of MI in women, which differ from the typical and more commonly described symptoms of MI in men. Because the patient’s presentation did not resemble the latter, the clinician reflexively discounted acute coronary syndrome as a possibility and focused instead on the patient’s prior diagnosis of GERD.

Clinicians can avoid being misdirected by the representativeness heuristic by knowing base rates of the occurrence of a particular condition in a population—in this case, acute MI in younger women (prevalence of 1% in women ages 35-54 years, according to NHANES data6); not giving too much weight to a single piece of information; and keeping in mind the possibility of atypical presentations.1

References
1. Klein JG. Five pitfalls in decisions about diagnosis and prescribing. BMJ. 2005;330:781-784.
2. Bornstein BH, Emler AC. Rationality in medical decision making: a review of the literature on doctors’ decision-making biases. J Eval Clin Pract. 2001;7(2):97-107.
3. Khan NA, Daskalopoulou SS, Karp I, et al. Sex differences in acute coronary syndrome symptom presentation in young patients. JAMA Intern Med. 2013;173:1863–1871.
4. Rubini Gimenez M, Reiter M, Twerenbold R, et al. Sex-specific chest pain characteristics in the early diagnosis of acute myocardial infarction. JAMA Intern Med. 2014;174:241–249.
5. Brush JE Jr, Krumholz HM, Greene EJ, Dreyer RP. Sex differences in symptom phenotypes among patients with acute myocardial infarction. Circulation. 2020;13(2):e005948.
6. Towfighi A, Zheng L, Ovbiagele B. Sex-specific trends in midlife coronary heart disease risk and prevalence. Arch Intern Med. 2009;169(19):1762-1766.

Join our email list

Get new course alerts, newsletters and more delivered directly to your inbox

By providing my personal information, including phone number, I consent to (1) receive email messages with information and offers, autodialed calls, texts, and prerecorded messages from FHEA, including current and possible future services, customer service and billing; and (2) FHEA’s Privacy Policy and Terms and Conditions. I understand that my consent is not required to purchase, and that cancellation of purchase does not automatically revoke this consent.