Filed & Forgotten

Loading archive index…

    A filed archive of broken systems

    Negative Control Contamination — GMP Batch Recursion

    Incident Summary

    Batch 7714-C was released following standard QC protocol. Post-release review identified:

    • The negative control sample registered a positive result at 0.3 ng/mL
    • The negative control had been pipetted from the same intermediate stock as the test sample
    • The batch record accurately reflected all steps as performed
    • The batch record was therefore simultaneously correct and invalid

    Kindy’s Note

    “I have logged this. I am not sure what I have logged. The batch was real. The control was real. The positive was real. The negative was also real.
    I have flagged this as an Existence Check failure and submitted Form 51-E (Emotional Verification of Sample Identity).
    Results may vary. So may I.”

    QA Finding

    The deviation was classified as a documentation observation, not a batch failure, because all steps were performed as written. The SOP did not specify that the negative control must be negative. This has been noted for the next SOP revision.

    The next SOP revision has been pending since Revision 3.

    Negative Control Contamination

    Simultaneously correct and invalid. The form remained intact. The situation did not.

    The SOP did not specify that the negative control must be negative. Meaning adjusted around the winning selector.

    Release-theater: if all steps were performed as written, the result is technically irrelevant. Relief remained outside scope.

    Good manufacturing practices are only good if nothing is ever manufactured.

    We follow the guidelines to ensure the error is properly credentialed.

    A perfectly compliant system is one that refuses to start.

    The human seeks quality; the flat file provides only stasis.

    The best practice is to unplug the machine and walk away.

    Negative Control Contamination

    Haikus

    The negative test
    Glows as bright as the live one
    Something here is wrong

    Labeled from the same
    We tested the exact same thing
    Twice upon the bench

    Manufacturing
    A perfectly clean mistake
    Paperwork is right

    Batch recursion loops
    If the control is the drug
    What did we just make

    Accurate report
    Of a terrible error
    Signed and sent away

    Negative Control Contamination

    The negative tested quite hot
    From the same intermediate pot.
    It was real, it was true,
    All the positive, too,
    But the SOP said negative’s not.

    Batch seven-seven-one-four
    Was released, and then released some more.
    The control was a lie,
    But it passed—here’s why:
    The procedure said nothing to pour.

    The deviation was filed, then refiled,
    The second report was “revised original” styled.
    Both exist in the database now,
    QA accepts this somehow,
    Though logic itself was defiled.

    Said Kindy, “I’m not sure what’s true.
    The batch was real. The control too.
    The positive was real,
    But the negative deal,
    And I filed what I filed, me and you.”

    Form 51-E asked the key question:
    “Does this sample know what it’s guessing?”
    Kindy checked the box,
    The audits don’t squawk,
    And “results may vary” — our confession.

    The QA said this is just fine,
    A documentation observation line.
    The SOP’s not clear,
    That’s what got us here,
    And the next revision waits for a sign.

    Revision Three’s been on the shelf
    Since Revision Four failed all itself.
    The next SOP draft
    Will address the craft,
    But released batches don’t wait to repent themselves.

    The procedure that filed batch seven
    Never said the control can’t be leaven.
    It only described
    What steps are prescribed,
    Not what makes a control go to heaven.

    First form said the batch was complete,
    Second form was “revised,” a treat.
    QA accepted the fiction,
    That both had conviction,
    And released was the now double-sheet.

    One pot fed both control and the test,
    A mistake, but the audit passed best.
    The SOP said, “File.”
    The QA said, “Smile.”
    And pharma got dose to the chest.

    Two deviations, both filed true,
    Two forms that shouldn’t both be through.
    The system records both,
    Like they’re both a troth,
    And the batch stays released, just as new.

    The intermediates shared the same slot,
    In a crowded and steaming-hot pot.
    They contaminated deep,
    While the SOPs sleep,
    And QA called it “release what you’ve got.”