Quality Control

TERMS TO REMEMBER:

  • Quality: a feature/characteristic of a product which meets the expected criteria of a consumer (customer).
  • Control: a solution that resembles a human sample that is used for QC purposes only
  • Standard: a colorless solution with known concentration of substances used for calibration
  • Specificity: defined as the ability of a method to measure the analyte of interest ONLY.
  • Sensitivity: defined as the ability of a method to measure analytes even at its lowest concentration
  • Accuracy: nearness of measured value to that of the target value
  • Precision: nearness of measured values to each other
  • Diagnostic specificity: defined as the ability of a method to detect a population of individuals absent of a disease process
  • Diagnostic sensitivity: defined as the ability of a method to detect a population of individuals having the presence of disease
  • Intralab QC (internal QC): control samples are run simultaneously with a patient to ensure reliability of methods and result. Used for daily monitoring of accuracy and precision of method used. Detects random and systematic errors.
  • Interlab QC (external QC): laboratories are given samples with unknown concentrations for them to test and results are compared with other laboratories thus maintaining “long-term accuracy” to methods utilized.
    • Results difference of greater than 2SD indicates disagreement with other lab included.
  • Mean: average of a set of values (mean = Σx/n). Measures central tendency.
  • Median: midpoint of a set of values
  • Mode: the most frequent among all values/data
  • Range: Simplest expression of spread or distribution
  • Standard Deviation: it is defined as the measure of dispersion of values to that of the mean. Most frequent used measure of variation.
  • Coefficient of variation: mean expression in percentile. Index of precision
  • Variance: square of SD. V=SD2
  • T-test: this is used to assess if there is a statistical difference between the means of 2 groups of data
  • F-test: this is used to assess if there is a statistical difference between the SD of 2 groups of data
  • Shewhart Levey-Jennings Chart: most widely used QC chart
  • Trend: six or more consecutive values that either increase or decrease gradually (will cross the mean) – main cause: reagent deterioration
  • Shift: six or more consecutive values that are distributed on one side or other side of the mean (does NOT cross the mean) – main cause: improper instrument calibration

WESTGARD RULES

TYPE OF ERROR

RULES

SOURCES OF ERROR

RANDOM

–  Tests for imprecision

12s (warning rule), 13s and R4s

By chance errors: mislabeling, pipetting error, fluctuations in temperature & voltage

SYSTEMATIC

–  Tests for inaccuracy

22s, 41s and 10x

Improper calibration, reagent deterioration, contaminated solutions, instability of both samples and solutions 

  • Delta check: used to check if there are significant differences between present set of values to that of past values on the sample of same individual.
  • Six Sigma: a way of improving product processing to eliminate defects
WESTGARD RULES
TYPE OF ERRORRULESSOURCES OF ERROR

RANDOM

Test for imprecision

12s (warning rule), 13s, and R4SBy chance errors: mislabeling, pipetting error, fluctuations in temperature & voltage

SYSTEMATIC

Test for inaccuracy 

 

22S, 41S and 10xImproper calibration, reagent deterioration, contaminated solutions, instability of both samples and solutions