Unsupervised machine learning identifies symptoms of indigestion as a predictor of acute decompensation and adverse cardiac events in patients with heart failure presenting to the emergency department.

The analysis was secondary and included only heart failure patients from a parent study of people arriving with chest pain or equivalent symptoms, so it does not describe all heart failure patients presenting to an emergency department. Symptoms were coded only as present or absent from electronic health records, without severity, so the clusters cannot show how mild or severe symptoms relate to risk. The study was done at three hospitals in a small geographic area, so the clusters may not generalise to other emergency departments or populations. The findings are associations, not proof that indigestion caused decompensation or adverse events, and several outcomes were composite. The reported odds ratios had lower confidence bounds at 1.0, so the excess risk is not strong certainty. Because the parent study recruited patients with chest pain or equivalent symptoms, symptoms such as palpitations may have been captured differently than in routine emergency care.

Do not conclude that indigestion alone causes heart failure decompensation or that every patient with indigestion needs urgent cardiac treatment. The association was found only among heart failure patients presenting with chest pain or equivalent symptoms, and the reported odds ratios had lower bounds at 1.0.