What This Run Does and Does Not Prove

What is supported: calibration and calibration verification are two distinct, required activities under CLIA, with calibration verification specifically designed to check low, mid, and high across the reportable range on a defined schedule, at least every 6 months and sooner after a reagent lot change, major maintenance, an unresolved QC trend, or the laboratory's own schedule. A residual pattern, small and random versus trended, is a reliable way to judge whether a fitted curve is trustworthy, and a failed verification calls for a specific investigation, not just a repeat recalibration.

No universal acceptance limit, recalibration interval shorter than the CLIA floor, or patient-result look-back window applies across methods, because those numbers depend on the manufacturer's instructions and this laboratory's own procedure, and vary by method and policy. The numeric values above are method-specific; use the manufacturer's instructions and the laboratory's procedure for other methods.

What depends on method and local policy: the number of calibration levels, the curve-fit model, the acceptance limits for both calibration and calibration verification, the reagent-lot-change exception criteria, and the exact investigation and patient-result look-back workflow. Look these up in the manufacturer's instructions for use and this laboratory's own procedure before acting on a real result.

When a calibration signal looks wrong, separate what the data show, a residual pattern, a recovery outside a stated band, a repeated failure, from what the laboratory's policy requires you to do about it, and document both.