Last week, we waged war against over-testing. This week, we’ll battle profligate differential diagnoses.
Do you remember Tom Smith, the man with fever, productive cough, and chest pain? He had a right upper lobe infiltrate and obvious community acquired pneumonia, right?
But wait, how do we know the chest pain was caused by pneumonia? What else could it be? Compleat physicians that we are, we’d hate to miss something. What about…
- Organizing pneumonia?
- Aortic dissection?
- Esophageal spasm?
- Boerhaave Syndrome?
- Peptic ulcer?
- Sickle cell?
- Something else?
You get my point. There’s merit to expansive thinking so we don’t miss anything, but our obsession with remote possibilities can also lead to unnecessary or misleading tests- CTs, troponins, echocardiograms, serologies, invasive procedures, and more. Worse, they distract us. The goal isn’t to make a long list; our mission is to home in on the one diagnosis making the patient sick.
So, think before you test. Which diagnoses merit attention? Are they likely? Are they even possible? If possible, are they “can’t miss?” Is the testing invasive? Risky? Complex? Expensive?
Remember, the history, physical, and routine diagnostic studies—the CBC, chemistries, urinalysis, and chest x-ray—are safe, simple, cheap, and effective. Usually, they tell you everything you need to know.
Let’s return to Tom. His pain began yesterday, so don’t worry about cancer. He denies hurting himself and his chest isn’t tender, so trauma is out. The pain is right-sided and it only hurts when he inhales, so stop worrying about heart disease and aortic catastrophes. He has no GI symptoms and no belly tenderness, so forget about peptic ulcer and cholecystitis. There’s no effusion, so it’s not pleuritis. There’s also no rash, so it’s not shingles.* Finally, let’s not fret about PE- emboli don’t cause productive cough and lobar infiltrates, and the Wells’ Score plummets when another diagnosis is more likely.
So, let’s end the profligacy and aim for parsimony. Just tell poor Tom he has pneumonia, start the antibiotics, and watch him get better. He needs rest, not tests.
Enjoy your Sunday, everyone.
*Someone will inevitably remind me that pain from Herpes zoster can precede the rash. That’s true, strictly speaking, but in this case, worrying about shingles would be obsessive.
On another note:
- Another essential COVID article by Zeynep Tufekci: 5 Pandemic Mistakes We Keep Repeating
- On vaccination and contagion: The False Dilemma of Post-Vaccination Risk
- And for those of you who share my love for MKSAP, prepare for MKSAP 19, due this summer!