AI-rewritten: This is a summary of an article from The Conversation, rewritten by AI (Qwen, running locally) to make it easier to read. The facts come from the original article – read it for the full story.

---

Mr. V visited an emergency department with chest tightness and shallow breathing. A resident physician diagnosed heart failure based on swelling and an enlarged heart on X-ray, but Mr. V’s blood BNP level was normal. The resident dismissed this result because he learned that obesity can cause falsely normal BNP levels in heart failure patients. He treated Mr. V with a diuretic to remove fluid, but the patient felt no better. A colleague later diagnosed cardiac tamponade, a life-threatening condition where fluid compresses the heart. The diuretic could have been fatal because a squeezed heart needs more blood flow, not less. Doctors rushed Mr. V to surgery to drain the fluid and save his life.

The resident was correct that many overweight patients with heart failure have normal BNP levels, but "normal" is not simple. Doctors generally consider BNP under 100 picograms per milliliter (pg/mL) as normal. However, a value under 50 pg/mL is uncommon in heart failure even for very overweight patients. Mr. V’s BNP was 14, which argued strongly against heart failure. The label "normal" treated a level of 14 the same as 95, hiding information that could have prompted doctors to look for other explanations.

The author teaches how to interpret diagnostic tests and notes that many doctors settle for simple "normal" or "abnormal" labels despite the nuance in test results. For example, D-dimer values between 500 and 1,000 nanograms per milliliter (ng/mL) slightly decrease the odds of a blood clot, yet they are flagged as abnormal. Similarly, LDL cholesterol under 100 mg/dL is normal, but for patients with prior heart attacks or strokes, anything above 55 mg/dL is too high. The author argues that specific numbers tell doctors far more than binary labels do.

The author and colleagues conducted a study involving over 600 clinicians to test if showing interval likelihood ratios improves medical decisions. This metric shows how strongly a result points toward or away from a diagnosis rather than just a single cutoff. Clinicians seeing these ratios made the correct call 93 out of 100 times, compared to 82 out of 100 for those seeing only a standard cutoff. The study found that this method helped doctors make better decisions regardless of their experience level, potentially reducing fatal errors caused by diagnostic mistakes.