General Laboratory: Maintaining QC Targets and Using Calibration Effectively Online Training

Welcome to the Maintaining QC Targets and Using Calibration Effectively Online Training course.  This is the second of four courses in a Quality Control series. This course will cover how to maintain the targets to make your QC process as effective as possible. Select Next to continue. Course created by: Nils B. Person, Ph.D., FACB Senior Scientist Global Product Education   Upon successful completion of this course, you will be able to: Identify when QC targets need to be updated List key reasons for excluding QC results from statistics Identify impact of calibration on QC troubleshooting Select Next to continue. “Calibration is an effective tool to address out of range QC results”   “To assure results remain accurate, properly set target values for mean and SD should lot be changed for this lot of QC material.” Statistical QC detects change Accuracy is established by calibration verification, analytical measurement range verification, and external quality assessment (proficiency testing) Verify targets periodically (e.g. monthly) Compare mean and SD for current period to cumulative values Cumulative values should be based on all data over an extended time Include multiple reagent lots and calibrations Want mean and SD to reflect overall performance of method If a change is noted in current month   Investigate If valid and permanent; update targets Result > 4 SD Mislabeled QC samples Data to Exclude Confirmed system problem Problem with QC material   SD limit Good results exceeding SD limit % per run: 1 test x 1 level     2 SD 5%     2.58 SD 1%     3 SD 0.3%     4 SD 0.006%       SD limit Good results exceeding SD limit % per run: 1 test x 1 level     2 SD 5%     2.58 SD 1%     3 SD 0.3%     4 SD 0.006%       SD limit Good results exceeding SD limit % per run: 1 test x 1 level     2 SD 5%     2.58 SD 1%     3 SD 0.3%     4 SD 0.006%       SD limit Good results exceeding SD limit % per run: 1 test x 1 level     2 SD 5%     2.58 SD 1%     3 SD 0.3%     4 SD 0.006%       SD limit Good results exceeding SD limit % per run: 1 test x 1 level     2 SD 5%     2.58 SD 1%     3 SD 0.3%     4 SD 0.006%       SD limit Good results exceeding SD limit % per run: 1 test x 1 level # per month: 1 test x 1 level   2 SD 5% 1.5   2.58 SD 1% 0.3   3 SD 0.3% 0.09   4 SD 0.006% 0.002     SD limit Good results exceeding SD limit % per run: 1 test x 1 level # per month: 1 test x 1 level   2 SD 5% 1.5   2.58 SD 1% 0.3   3 SD 0.3% 0.09   4 SD 0.006% 0.002     SD limit Good results exceeding SD limit % per run: 1 test x 1 level # per month: 1 test x 1 level Total # per month: 25 tests x 2 levels 2 SD 5% 1.5 75 2.58 SD 1% 0.3 15 3 SD 0.3% 0.09 4.5 4 SD 0.006% 0.002 0.09   SD limit Good results exceeding SD limit % per run: 1 test x 1 level # per month: 1 test x 1 level Total # per month: 25 tests x 2 levels 2 SD 5% 1.5 75 2.58 SD 1% 0.3 15 3 SD 0.3% 0.09 4.5 4 SD 0.006% 0.002 0.09   SD limit Good results exceeding SD limit % per run: 1 test x 1 level # per month: 1 test x 1 level Total # per month: 25 tests x 2 levels 2 SD 5% 1.5 75 2.58 SD 1% 0.3 15 3 SD 0.3% 0.09 4.5 4 SD 0.006% 0.002 0.09   SD limit Good results exceeding SD limit % per run: 1 test x 1 level # per month: 1 test x 1 level Total # per month: 25 tests x 2 levels 2 SD 5% 1.5 75 2.58 SD 1% 0.3 15 3 SD 0.3% 0.09 4.5 4 SD 0.006% 0.002 0.09   SD limit Good results exceeding SD limit % per run: 1 test x 1 level # per month: 1 test x 1 level Total # per month: 25 tests x 2 levels 2 SD 5% 1.5 75 2.58 SD 1% 0.3 15 3 SD 0.3% 0.09 4.5 4 SD 0.006% 0.002 0.09         Bad Practices Excluding “Out of range” QC results that have no confirmed problem: Not recorded Excluded from statistics during review Incorrect QC targets SD too small Range too tight Increased false positives Mean:  85.3 SD:  6.84 CV:  8.0%   Mean:  84.8 SD:  6.00 CV:  7.1%   Actual range: +/- 1.75 SD   Mean:  83.8 SD:  5.38 CV:  6.4%   Actual range: +/- 1.57 SD   Actual assay performance Unrealistic performance expectation Remember fundamental assumption: normal analytical variation will be symmetrical around the target mean Sustained shift in QC results may invalidate that assumption QC protocols less effective If shift in mean does not impact patient results, QC mean should be updated Using Common QC targets for multiple systems Key points: Common SD usually works Reagent lot changes can impact Monitor individual means Myth of Calibration:  “Calibrating more frequently will improve method accuracy”   Calibrator Target:  70 Method CV: 5% 95% Range: +/- 10% Calibration (x5) variation: +/- 4.46% Calibrating more often than absolutely necessary only increases variation in results Calibration corrects for normal drift, which is gradual Calibration to address QC rule failure usually … Covers up the problem Delays problem resolution Wastes time and materials Increases TAT for patient results Does not improve result quality Calibration? It's time Major service event New calibration is bad TSC request Long term drift Usually not necessary to calibrate ALL methods Calibrate those methods where QC results indicate a significant change QC results are used to detect change in performance If targets are updated for lot changes, matrix effects, etc. Calibration should only be performed when necessary If targets are set based on actual performance…future change will be detected Any new change will be detected Time is not wasted investigating problems that don’t exist Calibrating in response to QC rule failures delays resolution