The New Lot and the Thirty Student Samples
The teaching laboratory's chemistry analyzer was switched to a new creatinine reagent lot on Friday morning. Only level 2 quality control material was run once with the new lot and gave 1.18 mg/dL against an established mean of 1.20 mg/dL and SD of 0.05; the instructor considered this acceptable. Over Friday afternoon, 30 student practice samples were run; students used the results for a graded case interpretation assignment. On Monday, level 2 QC is 1.42 mg/dL and level 1 is 1.10 mg/dL (mean 1.10, SD 0.04). The analyzer was last calibrated four months ago with the previous reagent lot; the new lot's package insert says calibration is required when the lot changes. The laboratory calibrates every six months and has no lot-to-lot verification procedure. The instructor's plan is to recalibrate, repeat QC, and move on, because the student samples were only practice. The program director asks whether the 30 results can stand and what to document.
1. Interpret the QC data using appropriate rules. Can the 30 student results be released or used for grading? Explain your reasoning.
Read the worked answer
Level 2 is 1.42 against 1.20 plus or minus 0.05, a shift of +4.4 SD. That violates the 1-3s rejection rule. Because only one level shifted, it does not meet 2-2s, but it is still a rejection. Level 1 is at its mean. A single high QC point appearing with a new reagent lot and no calibration is a classic calibration bias, not random error. Under a Westgard multirule approach, reject the analytical run.
The 30 student results were produced after the new lot was placed but before a valid calibration and before any lot-to-lot verification, so they are not traceable to a stable measurement system. 'Practice only' does not make the results usable for a graded assignment: students used them to make interpretive decisions, and the assignment rewards the numbers. The defensible position is that the results cannot stand as valid patient-equivalent data. Do not release them as reportable, and do not grade them as if accurate. Either re-run retained samples after calibration and verification, or replace the assignment with a known dataset. If re-run, compare at least 20 samples across the reportable range against the old lot to detect bias, and accept only if the difference is within the clinically allowable total error.
2. What is the most likely root cause, and what corrective action would you require before the analyzer returns to service? Include verification steps.
Read the worked answer
The most likely root cause is that the new reagent lot was placed into service without calibration or lot-to-lot verification. The package insert states calibration is required with a lot change, yet the laboratory calibrates every six months and has no lot verification procedure. Contributing causes: only one QC level was run once, so a level-specific bias could hide; no reagent lot log; and no documented acceptance criteria.
Corrective action: remove the lot from service. Calibrate with the new lot using a fresh calibrator, verifying calibrator lot and expiry. Run both QC levels, ideally in duplicate, and apply Westgard rules. Perform a method comparison with 20 samples spanning low, normal, and high creatinine, using the previous lot or a reference method. Calculate bias and compare it with the allowable total error; for creatinine, a commonly used goal is about 15% or 0.3 mg/dL, depending on the program's stated quality goal. If acceptable, document and release; if not, reject the lot and obtain another. Check reagent storage and probe function if calibration does not correct the bias. Do not simply recalibrate and move on; the verification is the evidence that the lot is fit for purpose.
3. How should the grading and the students be handled, and what quality records should be created?
Read the worked answer
Handle the students transparently and without penalty. Tell them the QC failed because of a laboratory process error, not their performance, and that their graded assignment will be based on verified results or a replacement dataset. Use the event as a live QC case: have them calculate the SD index, apply the 1-3s rule, and write the corrective action, which is more instructive than the original assignment.
Records: a QC failure report; a CAPA with root cause, actions, responsible persons, and due date; a reagent lot log; an updated procedure for lot-to-lot verification and calibration; training records for all instructors; and a review at the next quality meeting. If the program reports results outside the teaching laboratory, determine whether any actual patient results were affected in the same period and notify the responsible person. Review whether graded assignments should use anonymized validated data in future. The program director should sign off on the CAPA and verify effectiveness by auditing the next three lot changes.