ASCLS Poster: Risk Management and Quality Control in Mass Spectrometry

ASCLS Poster: Risk Management and Quality Control in Mass Spectrometry

Risk Management and Quality Control in Mass Spectrometry

Zoe C. Brooks, ART
AWEsome Numbers Inc., Sudbury, ON, Canada

John Hopkins, BSc(Hons)
Clinical Mass Spectrometry Consulting Ltd

George Sweeney, MLS(ASCP)CM
M.O.R.E. Quality Consultant
LC/MS offers many advantages over other techniques in terms of specificity, accuracy and precision, but existing quality control processes are challenged to effectively monitor the analytical performance against healthcare objectives and risks.

“Risk [is the] combination of the probability of occurrence of harm and the severity of that harm” (ISO/IEC Guide 51).

We examined patient and QC sirolimus results from 2016 from a hospital/reference laboratory. We calculated the probability of risk from the sigma or z-value based on the monthly mean and SD of the QC sample relative to the package insert midpoint and a TEa limit of +/- 15%. We created scatter plots of the full year’s patient results below 50, plus 3 QC samples with package insert means of 3.6, 10.8 and 17 ng/L. Because of daily calibration, we saw no obvious multi-day or multi-week shifts as seen in the biochemistry data previously examined.

We examined histograms of monthly patient data. In the month of September, 16% of reported results were above 21, twice as many as the overall average of 8%. The month of December reported only 2% of results > 21, just ¼ the oval average and only 1/8 as many as September. We considered these clinically significant changes creating unacceptable patient risk.

In September, the sigma value of was below the acceptable risk criteria of 2 sigma. In December, sigma passed at 2.2. We used software to simulate a shift to the acceptable risk criteria of 2 sigma. The existing QC chart limits of +/- 15% failed to flag this failure.

Software designed a mathematically-optimiZed QC strategy based a Margin for Error (the number of SD the mean can shift before risk becomes unacceptable) that was effective to detect unacceptable risk in one day.

We believe that this process of “Mathematically-OptimiZed Risk Evaluation” will improve analytical compliance and reduce patient risk and costs.

/ Events

Share the Post

About the Author

Comments

Comments are closed.