University:
Eastern Kentucky University
Course:
STA 215 | Introduction to Statistical Reasoning
Academic year:

2024
Views:

110

Pages:

12
Author:

Kaushik Sen

How can we determine if a change is acceptable or not? Alerts us to changes in accuracy and precision Levey-Jennings Graph 2010- 2016 301,0 2015- 300 100.5 Understanding Bias!!! Monitors and evaluates method performance FULLY RECAP- Definitions of Terms Accuracy The closeness of measurements to the true value Bais: The difference between the expectation of a test result and an accepted reference value Precision The amount of variation in the measurements Error Inaccuracy leads to Bias (Systematic Errors) Imprecision (Random Errors) Actions Easier to detect and correct. Many of the causes are correctable.. More difficult to detect and eliminate.. RECAP - Accuracy and Precision Inaccurate but Precise Inaccurate and Imprecise BIAS SD Accurate and Precise How do you detect and Quantify? Mechanism Bias Detection Bias Quantification Imprecision Detection Imprecision Quantification EQA/PT Peer Group Computing Bias = x -True Value X True Value Bias CALCULATION Bias = x -True Value Consistent change in the analytical system Error in a given direction (negative or positive number) Bias has direction Positive (+) or Negative (-) Observed mean > True value: Bias is + Observed mean < True value: Bias is - CALCULATIONS Absolute Bias and percent Bias Since Total Error (TE) calculations involve the sum of Bias and Imprecision, Bias has to be converted to an absolute number Absolute Bias x - Positive number Used in calculations for TE - Absolute Bias % = (Abs Bias /True Value) *100 But where do you get the True Value from???? What makes it "true"? Inter-laboratory comparison programs of the QC manufacturer Good QC providers give monthly as well as cumulative means The cumulative means are robust values and will give very good anchoring of the true value Manufacturer assigned mean Long term lab mean Provided the QC lot has been running for a considerable duration Recap: Causes of Systematic errors (Trends and Shifts) Trend: Gradual loss of reliability in the test system Cause Deterioration of the instrument light source Gradual accumulation of debris in sample /reagent tubing Aging of reagents Gradual deterioration of control materials Gradual deterioration of incubation chamber temperature (enzymes only) Gradual deterioration of calibration Bias = x -True Value True Value True Value of the HWD QC material is 110 umol/L 103 umol/L 110 Bias = x -True Value X True Value True Value of the HWD QC material is 110 umol/L 109 umol/L umol.0 110 umol/L Before Calibration After Calibration