A flagged result that reached the chart anyway

At 09:40 on a day-shift core lab bench, a chemistry analyzer released a plasma potassium result of 7.6 mmol/L. The specimen was hemolyzed. The analyzer's hemolysis index (H-index) flagged that fact inside the instrument software, but the flag never reached the laboratory information system (LIS) work list the technologist actually worked from, so the result went out unreviewed against the flag.

The ordering unit called at 10:15, questioning a value that did not match a normal ECG. A repeat specimen drawn at 10:30 and reported at 10:52 came back 4.4 mmol/L, not hemolyzed, H-index within acceptable limits. The reference interval used locally for adult ambulatory plasma potassium is 3.5-5.1 mmol/L; a laboratory sets and verifies its own interval for its method, and this case declares that interval as local policy for the exercise.

The supervisor was notified at 10:20. The LIS report was amended with a corrected-report comment at 10:55, and the ordering clinician had the corrected value by 10:58. That sequence, notice, amendment, callback, is a correction: it fixes the one report in front of you. It says nothing yet about why the flag did not stop the release, or whether the same gap will let the next hemolyzed potassium result through next week.

Potassium and hemolysis interference are the vehicle here; the method for investigating and correcting a nonconforming event applies broadly across analytes and error types. CLIA regulation at 42 CFR 493.1282 and 493.1289 requires a laboratory performing nonwaived testing to have a corrective-action policy that is actually followed, to document corrective action for analytic failures such as this one, and to run an ongoing quality-assessment process that checks whether the fix worked.

CLSI QMS11 calls an occurrence like this a nonconforming event: something that does not conform to the laboratory's own procedure, does not meet an applicable requirement, or affected or could have affected safety. CAP Laboratory General Checklist requirements expect the laboratory's quality management system to include a documented process to identify, investigate, resolve, and follow up on events like this one across every section and shift.

The question the supervisor should be asking by 10:30, not 10:58, is what system condition allowed this event, what action will actually reduce the chance it happens again, and how will the laboratory know, with evidence rather than confidence, that the action worked. Correcting the report closes the patient-safety gap for one result; it does not close the investigation.