A number arrives. Is it the patient's number?

A potassium result of 6.7 mmol/L prints on a basic metabolic panel for an adult emergency-department patient. The reference interval on the report is 3.5 to 5.1 mmol/L, so the number is flagged high. Two weeks earlier the same patient had a potassium of 4.1 mmol/L. Nothing about the analyzer looks wrong: quality control passed, calibration is in date, and the method performed as designed. The analyzer is not broken; the question is where along the whole path from order to clinical action this specimen or this result could already have gone wrong before anyone reads the number.

A laboratory result is not one event. It is the end of a sequence: a clinician orders a test, a collector identifies the patient and draws the specimen, a courier or tube system moves it, a receiving technologist checks it in, an analyzer measures it, a reviewer checks it against rules, a system transmits it, and a clinician reads it and acts. Plebani has described this whole sequence, from test selection through interpretation and clinical action, as the total testing process, and has shown that most detectable laboratory errors occur outside the analytical step itself, not inside it (plebani-2010). An analyzer-only view of quality misses most of the places a patient can be harmed.

This case will return with its full detail in the guided example. For now, hold the shape of the problem: a flagged result, a prior value for comparison, and an interferent flag on the specimen itself. A result can fail between the order and the action taken from it; a technologist must recognize the earliest point where something already went wrong.

Before trusting or discrediting a flagged result, ask what happened to the specimen and the record before the analyzer ever saw it, not only what the analyzer reported.