Unpacking the Cognitive Baggage: an Exercise in Clinical Reasoning

In this series, a clinician extemporaneously discusses the diagnostic approach (regular text) to sequentially presented clinical information (bold). Additional commentary on the diagnostic reasoning process (italics) is integrated throughout the discussion.

CASE DISCUSSION

An 82-year-old female with history of heart failure with preserved ejection fraction, pulmonary embolism and atrial fibrillation on apixaban, sick sinus syndrome status post pacemaker placement, and coronary artery disease presents to the emergency department with 4 months of progressive dyspnea at rest and on exertion.

Dyspnea, the sensation of difficulty breathing, arises from an imbalance between the respiratory drive from the brainstem and the perception of how much effort it takes to breathe. The lungs, which are responsible for gas exchange, and the heart, which is responsible for gas delivery, are the most common culprit organs for causing dyspnea. A variety of non-cardiopulmonary conditions can cause dyspnea as well, e.g., anemia, acidosis, neuromuscular disease, and deconditioning. The patient’s past medical history puts her at risk for many of the common causes of dyspnea including cardiomyopathy, pulmonary embolism, dysrhythmia, and cardiac ischemia. Her description of the dyspnea, the time course, associated symptoms, and her physical exam will help to guide this very broad differential diagnosis.

Clinicians begin formulating a differential diagnosis from the moment they hear the patient’s chief complaint.¹ Often, the etiologies with the highest prevalence come to mind first. Other times, diseases which pose a great risk to the patient’s health drive the differential diagnosis. The discussant utilizes these base-rate and severity heuristics, respectively, to list cardiopulmonary causes at the top of her differential. Complications can arise when common diseases present atypically or new data contradicts the initial hypotheses.² Clinicians can combat the resulting cognitive biases through cognitive debiasing strategies.^2,3 In this case, the discussant debiases themselves two ways. First, she specifies which non-cardiopulmonary causes of dyspnea she would consider in her differential diagnosis, rather than leaving them as a bucket of unnamed possibilities. Second, she utilizes the past medical history to name several diagnoses of interest and the hypothesis-driven inquiry she would follow to rule them in or out.⁴ Unpacking a bucket of diagnostic options into a list of specific choices increases their salience in the clinician’s mind and increases the chances a clinician will uncover clues to their presence.^4,5

She was admitted to the hospital 4 months ago when her symptoms started. An electrocardiogram showed atrial flutter without signs concerning for acute ischemia. Serial troponins were negative. Computed tomography angiography of the chest demonstrated moderate cardiomegaly, bibasilar atelectasis, and biapical pleural scarring without evidence of pulmonary embolism. A transthoracic echocardiogram showed normal left ventricular size and systolic function with an ejection fraction of 55–60%, moderate mitral regurgitation, and a dilated inferior vena cava consistent with increased right atrial pressure similar to an echocardiogram 1 year ago. Regadenoson SPECT myocardial perfusion imaging showed low probability for ischemia. Her symptoms improved with diuresis and she was discharged home.

One month prior to the current presentation, the patient again presented with dyspnea associated with dull, retrosternal chest pain, and generalized weakness. A right heart catheterization demonstrated a mean right atrial pressure of 5 mmHg, wedge pressure of 7 mmHg, pulmonary artery saturation of 63%, and mean pulmonary artery pressure of 27 mmHg. Coronary angiography showed no significant coronary artery stenosis. Telemetry monitoring showed atrial flutter with a normal rate. Her symptoms resolved without intervention and she was discharged home.

The patient has had an extensive workup done in the past several months, none of which explains her current symptoms. She has been in atrial flutter, which by itself can cause dyspnea due to loss of the atrial kick. However, this is not a new diagnosis, and attributing her current symptoms to this alone would likely constitute premature closure. The apical pleural scarring seen on her CT scan is a common finding, often idiopathic, and unlikely to be contributing significantly to her symptoms. Her unremarkable stress test and coronary angiography make coronary ischemia as an etiology of her current dyspnea very unlikely. The result of her right heart catheterization is suggestive of very mild pulmonary hypertension. Pulmonary hypertension can certainly cause dyspnea, but this degree of pulmonary hypertension usually does not cause significant dyspnea. The normal wedge pressure makes it less likely that either volume overload or mitral regurgitation has been contributing significantly to her dyspnea.

The results of extensive prior testing show that she has potentially contributing cardiopulmonary conditions, but heuristics alone have not found a diagnosis. The differential diagnosis includes a broad list of non-cardiopulmonary causes of dyspnea such as chest wall disease, neuromuscular disease, acid-base disorders, hematologic conditions, and psychiatric conditions.

Given this broad list, clinicians can be tempted to seize upon any positive finding as the sole explanation for this patient’s symptoms, for example, blaming her dyspnea on mild volume overload. This type of cognitive bias is called search satisficing. Search satisficing bias is defined as the tendency to call off a diagnostic search when sufficient evidence is found, so that clinicians satisfy themselves that they have been looking in the right place.⁶ Clinicians can sometimes immunize themselves against these biases through meta-cognition, or thinking about their thinking. Diagnostic checklists and reflection-based exercises have been tested as vehicles for meta-cognitive debiasing, with mixed results.^2,7 The discussant reflects on the results of each diagnostic study within the patient’s clinical background and concludes that she is still searching for a culprit diagnosis.

Regarding her current presentation, the dyspnea is present both at rest and on exertion. It is not associated with chest pain, diaphoresis, or coughing but is associated with worsening generalized weakness over the past year. She started using a walker 1 year ago and now can barely move from the couch to the bathroom (about 10 feet) without stopping because of fatigue and weakness. She also reports several years of progressive, intermittent difficulty swallowing liquids and solids. She describes occasional regurgitation of undigested food but denies pain with swallowing. Both the weakness and the dysphagia had progressed somewhat since her initial presentation 4 months ago but were present then as well.

Her past medical history includes atrial fibrillation, sick sinus syndrome with pacemaker implantation, prior pulmonary embolism, coronary artery disease, and childhood polio without residual deficits. Medications include apixaban, metoprolol succinate, furosemide, clopidogrel, pravastatin, and magnesium oxide. There have been no recent changes to her medication list, and she is adherent to her medications. She denies tobacco, alcohol, or other illicit drug use. Her review of systems is otherwise negative.

The patient’s problem list now has three main issues: dyspnea, generalized weakness, and dysphagia. A neuromuscular disorder with involvement of the respiratory, skeletal, and oropharyngeal or esophageal muscles would be a possible unifying diagnosis. Myopathies such as inclusion body myositis, dermatomyositis/polymyositis, and drug-induced myositis related to statin use must be considered. Inclusion body myositis is often insidious in onset, and the associated dysphagia can precede the skeletal weakness by many years, which would fit with her presentation. Neuromuscular junction disorders like myasthenia gravis can cause this constellation of problems as well, although myasthenia gravis usually presents with ocular symptoms as well. Other etiologies to consider include motor neuron diseases like amyotrophic lateral sclerosis or post-polio syndrome which can cause progressive muscle weakness in survivors of poliomyelitis decades after the initial infection.

Of course, her three problems could certainly be caused by more than one diagnosis, so we will continue to consider the very broad differential diagnoses for each of these problems separately. The differential diagnosis for generalized weakness is vast, including causes of objective weakness like the neuromuscular disorders mentioned above and causes of subjective weakness from more generalized conditions like deconditioning and hypoxia. A detailed physical exam will differentiate between objective and subjective weakness. If she has objective weakness, the exam will help to localize the level of the lesion.

The unpacking principle states that people tend to underestimate the probability of unspecified possibilities in a problem space.⁵ For example, if a clinician lists “unstable angina, pulmonary embolism, and other cardiopulmonary causes” as their differential for a patient’s dyspnea, they will naturally favor the specified diagnoses over the bucket of unnamed causes. This unconscious favoritism for named options over nameless ones is less relevant if clinicians name the true diagnosis in their initial hypotheses. However, if the initial hypotheses prove false, clinicians can still remain biased toward their initial ideas, causing them to stop gathering data prematurely and discard important diagnostic considerations.³ Clinicians can combat this cognitive bias by unpacking the differential diagnosis for a complaint to increase the availability of other possibilities in their mind. For example, the discussant previously unpacked dyspnea into a framework of system-based categories. Each of these categories can be further unpacked as new data is obtained. If the initial workup proves unrevealing, the unpacking principle would suggest clinicians return to the bedside, using their unpacked differential, to obtain a hypothesis-driven history and examination.

In this case, the discussant reformulates her problem representation to include dyspnea, weakness, and dysphagia. This problem list triggers recognition of neuromuscular disorders as a potential unifying diagnosis. Rather than considering only the commonest etiology from this category, the discussant then unpacks the differential diagnosis for neuromuscular disorders to guide further investigation.

On examination, the patient’s temperature was 36.4 °C, blood pressure 122 over 82 mmHg, heart rate 65 beats per minute, respiratory rate 18 breaths per minute, and oxygen saturation 97% on room air. The patient’s cardiovascular exam showed a regular rate and rhythm; absence of S3, S4, rubs, murmurs, and gallops; and symmetric, normal pulses with warm and dry extremities. Jugular venous pressure was normal. She had 1+ bilateral lower-extremity edema at her ankles. Lungs were clear to auscultation. A comprehensive neurologic exam was performed which revealed reflexes 1+ throughout. Babinski’s sign was not present. Motor exam revealed 4+/5 strength in the right upper and lower extremities and 5−/5 strength in the left upper and lower extremities. Sensation was normal to light touch. Generalized decreased bulk and tone were noted without fasciculations. Gait was not assessed due to the patient’s profound weakness and inability to ambulate without assistance.

An EKG showed atrial flutter with no signs of ischemia or significant changes from previous exams. Chest x-ray showed stable cardiomegaly without an acute process. Her complete blood count was notable for hemoglobin of 10.9 g/dL and her chemistry for a blood urea nitrogen of 45 mg/dL with creatinine 1.0 mg/dL and albumin 3.3 g/dL, which are at her historical baselines.

The main takeaway from her physical exam is that she does in fact have objective weakness as well as generalized decreased bulk and tone. These findings make a neuromuscular disorder very likely. In addition, there are no major cardiopulmonary abnormalities on her exam—her electrocardiogram is unchanged from prior, her chest imaging has no new abnormalities, and her basic labs are near her historic baselines—strongly suggesting that her dyspnea is due to respiratory muscle weakness. The next steps would be to assess her respiratory muscle strength by measuring her negative inspiratory force (NIF) and forced vital capacity. Diagnostic testing should be obtained to help confirm the presence of a neuromuscular disorder and help localize the level, including muscle enzymes such as creatinine kinase and aldolase, and electromyography.

By asking, “Is the new data concordant or discordant with the working hypothesis?” clinicians can encourage analytical thinking and increase their chances of avoiding cognitive bias.

Given her progressive dysphagia and dyspnea without clear cardiopulmonary etiology, a neurologic process was considered. Cyanocobalamin, folate, and thyroid-stimulating hormone levels were normal. Testing for syphilis, Lyme antibodies, and acetylcholine receptor modulating/binding antibodies had negative results. A swallowing study was not completed and deferred to the outpatient setting. CT scan of the neck 1 month prior showed no obstructive mass or laryngeal process. The patient’s pacemaker was incompatible with magnetic resonance imaging, so no imaging of the brain or spinal canal was obtained.

An important step in diagnostic reasoning is the renaming and reordering of items on a problem list in light of newly obtained data. An updated problem list would look as follows: first, chronic progressive dyspnea on exertion and at rest without clear cardiopulmonary etiology after extensive workup; second, subacute progressive generalized weakness equal in distal and proximal extremities with atrophy but without upper motor neuron signs or myalgias; and third, insidiously progressive dysphagia to solids and liquids over several years. Neuromuscular disorders, at the level of the muscle, neuromuscular junction, or motor neuron, remain the most likely explanation for the above constellation of problems. Although this laboratory workup rules out some of the common causes of neuromuscular disorders, many others remain.

Of the myopathies being considered, inclusion body myositis seems most likely as it is usually insidious in onset and slowly progressive, is more common in older adults, and involves both proximal and distal muscle groups. Muscle enzymes can be mildly elevated, and electromyography could point strongly to the diagnosis, although a muscle biopsy is usually necessary to confirm the diagnosis. Neuromuscular junction disorders are less likely: myasthenia gravis is the most common neuromuscular junction disorder, but the patient does not present the classic clinical features of ocular involvement and fluctuating fatigue, and her binding receptor antibody was negative. Seronegative variants of myasthenia gravis exist nonetheless.

Motor neuron diseases are also on the differential diagnosis. Amyotrophic lateral sclerosis, which is the most common of the degenerative motor neuron disorders, is less likely. The clinical hallmark is a combination of upper and lower motor neuron findings which is missing in this patient. Progressive muscular atrophy can present with only lower motor neuron findings and is a clinical diagnosis after excluding alternative diagnoses. Electromyography findings can be supportive. Finally, given her history of poliomyelitis in childhood, we should continue to consider post-polio syndrome which can cause insidious onset of progressive weakness many decades after initial infection and would fit her presentation well.

Muscle enzymes and electromyography would be the next diagnostic step, with consideration of a muscle biopsy and nerve conduction study. A consultation from a neuromuscular disorder specialist may also be helpful.

Clinicians increase their likelihood of recognizing the correct illness script when they unpack a patient’s complaints with as much specificity and detail as possible. In this instance, the discussant performs a diagnostic pause,⁸ a cognitive forcing strategy, to rename the patient’s problems considering all collected information. Based upon these updated problems, the discussant affirms their working diagnosis of a neuromuscular disorder. The subsequent unpacking of the neuromuscular disorder differential diagnosis allows the discussant to contextualize the results of the newest diagnostic results in preparation for obtaining a potentially confirmatory diagnostic test: electromyography.

Electromyography (EMG) was obtained which showed both active and chronic axonal processes leading to intraspinal canal lower motor axon loss affecting the cervical and lumbosacral areas. Findings were suggestive of a primary motor neuron disease. Given the patient’s history of childhood polio and the isolated lower motor neuron findings, her symptoms were thought to be from post-polio syndrome with mild respiratory involvement. Amyotropic lateral sclerosis was also considered but was thought to be less likely given the slow progression of the patient’s symptoms and widespread involvement of cranial, cervical, and lumbar spinal nerves.

Her strength improved throughout hospitalization with physical therapy. She was discharged home in stable condition with plans to follow a low-intensity physical therapy regimen and to repeat EMG if any progression of symptoms occurred. On outpatient follow-up, her symptoms had improved, although she remained tired throughout the day.

DISCUSSION

This case demonstrates several important diagnostic principles. First, the patient’s initial investigation demonstrates the appropriate usage of diagnostic heuristics and their pitfalls. For example, cardiopulmonary causes of dyspnea were ruled out first given their commonness (base-rate heuristic), their potential severity if not found early (severity heuristic), and the patient’s history of those diseases (availability heuristic). Although the initial workup was thorough, it did not investigate her dysphagia or weakness, which was just as important as her dyspnea to achieving a final diagnosis. This diagnostic delay reinforces the importance of a thorough history and evidence-based physical exam maneuvers for adjusting a differential when one’s initial heuristics have not found a diagnosis. One study has found that up to 80% of diagnoses may rely on careful history-taking alone, with 10% relying on the addition of a detailed physical examination.⁹

Second, this case demonstrates the diagnostic power of the unpacking principle and its applications. In short, clinicians naturally favor named possibilities in a problem space over unnamed possibilities. Additionally, the more detailed and specific the description of a hypothesis becomes, the more probable the hypothesis seems.^2,4–7,10 By unpacking their initial differential diagnosis and using a careful history and evidence-based exam to fully describe a patient’s problems, clinicians increase their chances of recognizing an appropriate illness script and the correct diagnosis. When unpacking does not occur, relevant diagnostic possibilities may not be considered and biases such as search satisficing can hinder timely diagnosis of a patient’s clinical problems.

This case highlights the importance of thorough history-taking to overcome cognitive biases and the crucial role of re-examining the available data when the original hypotheses does not contain a convincing diagnosis.

CLINICAL TEACHING POINTS

1. Post-polio syndrome (PPS) is characterized by new weakness, fatigue, and pain occurring decades (average 35 years) after acute poliomyelitis.¹¹ Although rare in the general population, PPS occurs in 25% of patients with a history of poliomyelitis.¹² The most important risk factors of PPS are female sex and respiratory symptoms during acute polio infection.¹¹

2. The pathophysiology of the disease is not fully understood, but axonal loss is often identified in areas previously affected during the primary illness.¹³ EMG/NCS can guide the diagnosis but does not reliably distinguish between PPS and other lower motor neuron diseases like ALS.

3. Low-intensity physical therapy, weight loss, and defatiguing techniques^14,15 are usually recommended. Respiratory support is also recommended if respiratory dysfunction is present, and orthoses can be used in cases with progressive loss of ambulation ability.¹⁵