Internal Quality Control (IQC) – Comprehensive Study Notes

Internal Quality Control (IQC): Definition, Purpose & Scope

• WHO (1981) definition: “A set of procedures for continuously assessing laboratory work and emergent results.”
• Operates under laboratory-specific IQC protocols while meeting NATA obligations.
• Core objective: guarantee day-to-day consistency so patient results are reliable before release.
  • Patient reports are issued only when IQC fulfils pre-established acceptance criteria.
• Two complementary pillars of analytical-phase validation:
  1. Internal Quality Control (current focus)
  2. External Quality Assurance (EQA / proficiency testing) – covered in next lecture.

Control Materials (QC Materials)

• Commercial examples: Bio-Rad frozen, freeze-dried (lyophilised) or chemically-preserved vials with manufacturer-assigned target values.
• Lyophilised vials require reconstitution:
  • Strict adherence to protocol; accurate pipetting of diluent critical to restore correct analyte concentration.
• Non-routine analytes may use in-house pooled patient sera:
  • Laboratory establishes its own target values via senior-scientist verification studies.
• Success of IQC hinges on selecting appropriate material.

ISO 15189 Requirements for Selecting QC Material

• Concentration range must cover clinical decision points encountered in patient samples.
  • Minimum two levels (e.g. low vs high, normal vs abnormal, positive vs negative, reactive vs non-reactive).
• Matrix should mimic patient sample composition as closely as possible.
• QC must be analysed identically to patient samples; errors in the run must manifest in QC data.
  • Example pitfall: never run a urine QC on a serum assay.
• Minimise variability:
  • Long-term stability, minimal vial-to-vial / batch-to-batch differences.
  • Hence commercial, certified QCs preferred over in-house preparations.

Establishing Control Limits

• Laboratories usually derive limits from their own data.
• Data collection guideline:
  • At least 20 independent data points from 20 separate, stable operating days.
  • Include all results—even failed values—to capture true precision.
• Statistical groundwork:
  • Calculate mean (\overline{x}) and standard deviation (SD).
  • Assume Gaussian distribution.
  • Conventional control limits: \overline{x} \pm 2SD ⇒ 95\% confidence; 5\% probability a value randomly falls outside.

Levy-Jennings (LJ) Charts

• QC result plotted chronologically; vertical axis = result, horizontal axis = time/run number.
• Horizontal lines mark \overline{x},\; \pm1SD,\; \pm2SD,\; \pm3SD.
• Two charts usually maintained in parallel:
  • Level 1 (control 1) – typically low/normal range.
  • Level 2 (control 2) – high/abnormal range.
• Scientists interpret patterns weekly/fortnightly/monthly for long-term precision, accuracy and bias trends.

Westgard Rules (Two-Level QC System)

• Decision matrix for accepting or rejecting analytical runs.

• Numerical prefix = number of results; “S” subscript = SD distance.

  Rule	Meaning	Action	Error Type Clue
  1-2S	1 result > \pm2SD	Warning only	—
  1-3S	1 result > \pm3SD	Reject run	Random
  2-2S	2 consecutive results > \pm2SD on same side (within or across levels/runs)	Reject	Systematic
  R-4S	Range of 4SD within same run (one +2SD, one –2SD)	Reject	Random
  4-1S	4 consecutive results > \pm1SD on same side (within/across)	Reject	Systematic
  10X	10 consecutive results on same side of mean (within/across)	Reject	Systematic

• Always evaluate both levels (within-run & across-run) before releasing patient data.

Interpreting QC Data: Precision, Accuracy & Bias

• Precision = reproducibility (tight clustering).
• Accuracy = closeness to true/target value.
• Bias = consistent deviation (high or low) from target.
• Ideal IQC outcome: precise + accurate (tight cluster at target).
• Scenarios:
  • Precise but inaccurate ⇒ bias present (shift on LJ chart).
  • Inaccurate & imprecise ⇒ random scatter; invalid assay.

Error Categorisation

• Random Error (Imprecision): sporadic, typically single excursions >3SD.
  • Causes: bubbles in QC vial, pipetting variability (worn syringes), electrical interference.
  • Westgard triggers: 1-3S, R-4S.
• Systematic Error (Bias): persistent drift or shift from mean.
  • Causes: calibration bias, reagent lot changes, instrument deterioration, non-specific assay reaction.
  • Triggers: 2-2S, 4-1S, 10X.

Pattern Recognition on LJ Charts

• Trend: ≥5 consecutive points moving steadily up or down.
  • Often gradual light-source decay or ageing QC/calibrator material.
• Shift: ≥6 consecutive points abruptly above or below mean.
  • Often reagent lot change, faulty recalibration, major maintenance.

Troubleshooting Failed QC

• Do not rerun blindly; perform systematic investigation:
  1. Inspect QC material: expiry, storage, vial-to-vial variation.
  2. Check reagent pack: lot changes, contamination, expiration.
  3. Verify calibrator: value assignment, preparation errors.
  4. Examine analyser: mechanical faults, optics, probes.
  5. Review environment: temperature, humidity, electrical stability.
• After identifying cause:
  • Rectify (e.g. replace reagent, recalibrate, service analyser).
  • Rerun QC to confirm values fall within \pm2SD before releasing patient results.
• Example: A faulty analyser will consistently fail reruns until the hardware issue is fixed.

Limitations of IQC

• Compares laboratory against itself; detects changes but not pre-existing systematic errors embedded in original method validation.
• Slow drift affecting all results may remain invisible if it started before baseline was established.
• Mitigation: participate in External Quality Assurance (EQA) / Proficiency Testing to benchmark against peer laboratories.

Key Takeaways

• IQC is mandatory, continuous and immediate; EQA provides external, periodic benchmarking.
• Proper QC material selection, statistical limit setting, vigilant LJ monitoring and disciplined application of Westgard rules are essential for reliable laboratory reporting.
• Ethical imperative: Never release patient results until QC integrity is unequivocally demonstrated.