In clinical laboratory quality control, traditional reliance on statistical metrics like Sigma and Westgard rules often fails to capture the measurable risk of patient harm. This Scientific Short challenges professionals to shift their focus from statistical compliance to actionable risk management. By teaching and thinking of QC data in terms of patient risk and leveraging simulations and targeted evaluations, laboratory staff can identify and prevent medically incorrect results more effectively. The article advocates for a paradigm shift: from managing statistics to managing the risk those stats represent to patients.
How many patients are you willing to send over the “Cliff of Extreme Risk”? That’s not just a provocative metaphor—it’s a necessary reframe. Every time you rely solely on QC flags or chase Sigma metrics without context, you gamble with patient safety. Every QC point past the allowable error limit represents the people who were tested with that sample.
“The prevention of harm to patients is the ultimate goal of medical laboratory services.”i Incorrect results that “do not meet the requirements for their intended medical useii” expose patients to the risk of delayed or incorrect diagnosis or treatment. They waste healthcare resources on unnecessary repeat and additional testing.i
Six Sigma’s “errors per million” might work on a factory line, but most medical labs don’t test a million patients for a single analyte in a decade, let alone a year. What matters to us isn’t long-term error frequency - it’s the immediate, individual risk of harm to patients. Patient risk and appropriate QC processes vary by instrument, by setting and by patient volume.iii
ISO 15189:2022iv and CLSI EP23Aii are clear—labs must actively manage risk and demonstrate how their systems prevent patient harm.v But while they tell us what to do, they’re vague on how to do it. Here’s one process
The author recently applied this process to review QC data from 84 samples, covering 14 analytes across two chemistry analyzers. Using a 3-sigma threshold as the performance criterion, 78 (93%) of the QC samples were deemed acceptable. However, when applying a more stringent standard of maximum one medically incorrect result (MIR) per year, only 72 (86%) met the acceptable risk level. Notably, 6 of the 12 samples that exceeded that threshold had a sigma above 3.0, highlighting that a sigma >3 does not always guarantee low clinical risk.
CLSI states that "At the least, the ability of the QC procedures to detect medically allowable error should be evaluated.”ii A simulated error rate of 1 MIR per day was introduced into the 72 QC samples that had previously demonstrated <1 MIR per year. The laboratory’s existing QC procedure failed to trigger a ‘1-2s’ rule flag in a single QC run for 16 of the samples (22%). QC rules tailored to observed method accuracy and precision achieved detection of all simulated errors in a single run across all samples. Additionally, risk-based QC processes would eliminate 80% of the flags that led to repeat testing and investigations during the month.
Think of the Cliff of Extreme Risk next time you're examining your QC charts. Those dots represent lives. Let’s manage RISK—not just stats!