Six-Day Creatinine Shift After an Unverified Reagent Lot Change
Your 250-bed community hospital chemistry section runs a single analyzer. On Monday the level-2 creatinine control shows +2.4 SD; the previous six days read +1.6, +1.8, +2.0, +1.7, +2.2 and +2.3 SD. The level-1 control has stayed within ±1 SD throughout. A new creatinine reagent lot was placed on the analyzer eight days ago; it was verified against the manufacturer's insert values rather than by parallel patient testing, and the analyzer was calibrated at the same time. The technologist on duty notes that the 1-3s rule has not been violated and released all results. Over those eight days, 156 patient creatinine results and 41 calculated eGFR values have been reported, including 12 from the oncology clinic. The chemistry supervisor is on vacation and you are the designated lead.
1. Is this QC acceptable under Westgard multirule? Which rules apply, and what should have happened?
Read the worked answer
The QC is not acceptable. The 1-3s rule is a rejection rule for a single control observation, but it detects random error and is largely irrelevant here. Systematic error is detected by the 2-2s, 4-1s and 10x rules and by trend rules such as 7T.
The readings +1.6, +1.8, +2.0 and +1.7 are four consecutive control values exceeding +1 SD on the same side of the mean - a 4-1s violation on day 4. By the sixth day, with a run of values consistently above the mean and a new reagent lot on board, the 10x rule and the trend rule are also in play. The correct action was to reject the run at the first 4-1s violation, stop reporting patient creatinine, and investigate the reagent lot and calibration before releasing further results.
Relying on 1-3s alone is a common and consequential error, because a slow, sustained shift - which is exactly what a bad reagent lot or calibration produces - never triggers that rule.
2. How do you investigate, and how do you determine whether the 156 reported results are affected?
Read the worked answer
Recalibrate with a fresh, correctly stored calibrator, re-run both control levels, and check reagent and calibrator storage, expiry and lot records. Then parallel-test at least 20 patient samples by the old and new reagent lots, or against a reference method, including several samples from the affected eight days retrieved from refrigerated storage.
Quantify the bias. If the shift is 8%, a true creatinine of 1.00 mg/dL reads as about 1.08; that moves an eGFR by roughly 8–10% and can reclassify chronic kidney disease stage, and it matters even more for the oncology patients if carboplatin dosing is based on creatinine clearance. Compare the observed bias with the allowable total error for creatinine - CLIA's acceptance limit is ±15% or 0.3 mg/dL, whichever is greater - and with your own method's total error budget.
If the bias exceeds the limit, the affected results must be repeated, corrected and re-reported, with direct clinician notification for the oncology patients. If it does not, document the evidence and retain the records.
3. Draft the corrective and preventive action plan, including communication.
Read the worked answer
Root cause: lot verification was performed against the manufacturer's insert values instead of parallel patient testing, and QC review was limited to the 1-3s rule.
Corrective: identify and retest patient samples from the affected period, re-report corrected values, notify the ordering clinicians - prioritizing the 12 oncology patients - and enter an incident report with the affected date range.
Preventive: revise the lot verification SOP to require parallel testing of at least 20 patient samples across the measuring range before a new lot is released for patient testing; add Westgard multirule training and competency assessment for all staff who release results; institute a second review of QC by a lead technologist; and add a monthly audit of QC review practice.
Communication: brief the medical staff on the corrected results, and confirm whether the event meets your institution's criteria for a reportable laboratory error under accreditation requirements. The supervisor's vacation does not defer accountability - as designee you must act on the shift today, not when the supervisor returns.