Quality Control in Clinical Biochemistry

QC Fundamentals - Laying Groundwork

• Quality Control (QC): Operational techniques ensuring test result reliability; part of QA.

• Quality Assurance (QA): Overall measures for quality at all testing stages (pre-analytical, analytical, post-analytical).

• Total Quality Management (TQM): Organization-wide philosophy for continuous quality improvement.

• Aims of QC: Monitor accuracy & precision; detect errors. Ensures reliable patient results for diagnosis & treatment.

• Errors:

  • Random (Imprecision): Unpredictable (e.g., pipetting error). Affects precision. Measured by SD/CV.
  • Systematic (Inaccuracy/Bias): Predictable, consistent (e.g., miscalibrated instrument). Affects accuracy.
  • Total Allowable Error (TEa): Maximum acceptable error for a test.

• Formula: $TE = Bias + Z \times SD$ (Z often 1.65, 1.96, or 2).

⭐ QC is product-oriented (focuses on test results), while QA is process-oriented (focuses on the entire testing lifecycle).

QC Statistics - Numbers Don't Lie

• Central Tendency:
  • Mean ($\bar{x}$): Average value.
  • Median: Middle value in sorted data.
  • Mode: Most frequent value.
• Dispersion (Spread):
  • Standard Deviation (SD): Measures data spread. $SD = \sqrt{\frac{\sum (x_i - \bar{x})^2}{n-1}}$
  • Coefficient of Variation (CV%): Relative SD; key imprecision measure. $CV% = (SD/\bar{x}) \times 100$
• Gaussian (Normal) Distribution:
  • Coverage: $\pm \textbf{1SD}$ (\textbf{68.3%}), $\pm \textbf{2SD}$ (\textbf{95.5%} - common QC limits), $\pm \textbf{3SD}$ (\textbf{99.7%}). 
• Accuracy vs. Precision:
  • Accuracy: Closeness to true value (reflects systematic error).
  • Precision: Reproducibility of measurements (reflects random error).
  • 📌 Accuracy hits the Aim (bullseye); Precision is Packed together (tight group).
• Sigma Metrics: $(TE_a - |Bias|)/SD$. Goal: \u22656 Sigma (world-class quality; <3.4 DPMO).

⭐ The Coefficient of Variation (CV%) is a standardized measure of imprecision.

Internal QC - Charting Control

• Control Materials:
  • Types: Lyophilized (freeze-dried), liquid; Assayed (known values), Unassayed (values by lab); Homemade, Commercial.
  • Desirable: Stable, matrix similar to patient samples, analyte levels at medical decision points (low/normal/high).
• Levey-Jennings (LJ) Charts:
  • Plot control values vs. time. Mean line; control limits: $\ extpm 1SD, \ extpm 2SD, \ extpm 3SD$.
  • Interpretation: Trends (gradual drift), Shifts (abrupt change).
• Westgard Multirules: (Error: R=Random, S=Systematic)
  • 1₂s: Warning (1 value > $\ extpm 2SD$).
  • 1₃s: REJECT (1 value > $\ extpm 3SD$) - R/S.
  • 2₂s: REJECT (2 consecutive values > same $\ extpm 2SD$) - S.
  • R₄s: REJECT (1 control >+2SD & another <-2SD) - R.
  • 4₁s: REJECT (4 consecutive values > same $\ extpm 1SD$) - S.
  • 10x: REJECT (10 consecutive values one side of mean) - S.
• Actions on Rule Violation:
  • Stop patient testing. Identify & correct error. Re-run controls, then patient samples.

⭐ The 1₃s rule is a rejection rule indicating significant random or systematic error.

External QC & Broader QA - Ensuring Excellence

• External Quality Assessment Schemes (EQAS) / Proficiency Testing (PT):
  • Purpose: Inter-lab comparison to identify systematic errors (bias) and assess accuracy.
  • Process: Labs analyze external blind samples; results submitted for peer comparison.
  • Interpretation: Performance uses metrics like Standard Deviation Index (SDI), Variance Index Score (VIS).
• Broader Quality Assurance (QA) considers:
  • Pre-analytical variables: Sample collection, handling, storage impact QC.
  • Post-analytical variables: Result reporting, interpretation impact QC.
  • Corrective and Preventive Actions (CAPA): For continuous quality improvement.

⭐ EQAS primarily assesses the accuracy of a laboratory's performance.

High‑Yield Points - ⚡ Biggest Takeaways

• Accuracy (true value closeness) vs. Precision (reproducibility) are key.
• Systematic errors (SE) affect accuracy; Random errors (RE) affect precision.
• Levey-Jennings (LJ) charts visually monitor QC data against mean ± SD.
• Westgard rules (e.g., 1₃₅, 2₂₅, R₄₅) applied to LJ charts detect SE & RE.
• Shifts (≥6 points one side) & Trends (≥6 points ↑/↓) indicate SE.
• Coefficient of Variation (CV%) = (SD/Mean) x 100; measures imprecision.
• IQC (daily) and EQA (inter-lab) ensure result reliability.