11.3 Quality Control Charts, Westgard Rules, and Laboratory Quality Assurance

11.3 Quality Control Charts, Westgard Rules, and Laboratory Quality Assurance

Pulmonary function testing laboratories generate quantitative physiological data critical for diagnosing pulmonary diseases, determining surgical risk, assessing disability, and guiding therapy. To ensure diagnostic accuracy, repeatability, and legal defensibility, laboratories must implement a comprehensive Quality Assurance (QA) and Statistical Quality Control (SQC) program. Equipment calibration alone is insufficient; daily quality control testing, statistical control chart monitoring, and adherence to objective multi-rule evaluation systems (such as Westgard Multirules) are mandated by accreditation bodies like the College of American Pathologists (CAP), The Joint Commission (TJC), and ATS/ERS guidelines.

---

Statistical Control Principles: Baseline, Mean, & Standard Deviation

Statistical process control requires establishing baseline statistical parameters for every testing system, gas analyzer, and volume transducer using stable reference controls (such as 3-Liter precision calibration syringes, automated blood gas controls, or biological controls).

1. The Mean (Target Value, $\bar{x}$)

The arithmetic average of a series of baseline control measurements collected under stable operating conditions over a minimum of 20 consecutive testing days:

$\bar{x} = \frac{\sum x_i}{n}$

2. Standard Deviation (SD, $\sigma$)

A statistic measuring the dispersion or spread of control values around the mean:

$SD = \sqrt{\frac{\sum (x_i - \bar{x})^2}{n - 1}}$

3. Coefficient of Variation (CV%)

Expresses standard deviation as a percentage of the mean, allowing comparison of precision across different physiological variables:

$CV\% = \left( \frac{SD}{\bar{x}} \right) \times 100$

---

Levey-Jennings Control Charts

A Levey-Jennings chart is a graphical quality control tool where daily control measurement results are plotted chronologically on the y-axis against time (days or runs) on the x-axis. The chart includes a central line representing the target Mean ($\bar{x}$) and parallel horizontal upper and lower control limits established at $\pm 1SD$, $\pm 2SD$, and $\pm 3SD$.

+3 SD ----------------------------------------------------------- (Reject: 1-3s)
+2 SD ----------------------------------------------------------- (Warning: 1-2s)
+1 SD -----------------------------------------------------------
 MEAN =========================================================== (Target)
-1 SD -----------------------------------------------------------
-2 SD ----------------------------------------------------------- (Warning: 1-2s)
-3 SD ----------------------------------------------------------- (Reject: 1-3s)

Error Categorization on Levey-Jennings Charts

1. Random Error: Unpredictable variations occurring without pattern, caused by electrical noise, temporary ambient temperature spikes, operator technique variations, or mechanical friction. Reflected on control charts as sudden, isolated spikes exceeding control limits.
2. Systematic Error (Bias): Predictable, directional shifts or gradual trends away from the mean, caused by gas analyzer sensor depletion, dirty pneumotachometer meshes, improper calibration gas concentration entry, or software updates. Reflected as persistent displacement above or below the mean.

---

The Westgard Multirule System in Pulmonary Diagnostics

To eliminate subjective interpretation of control charts, laboratories apply Westgard Multirules—a decision tree of statistical rules that evaluate control run acceptability. When a control value falls outside specified limits, the technologist evaluates the rule combination to determine whether to accept the run, flag a warning, or halt diagnostic testing.

                            +-----------------------+
                            |   Daily QC Run Measured|
                            +-----------------------+
                                        |
                               Is Value > +/- 2 SD?
                                /               \
                              NO                 YES
                              /                   \
                   +---------------+      +-------------------------+
                   | ACCEPT RUN    |      |  1-2s WARNING FLAG      |Cumulative
                   | (In Control)  |      +-------------------------+Evaluation
                   +---------------+         /   |      |      |   \
                                            /    |      |      |    \
                                         1-3s   2-2s   R-4s   4-1s  10-x
                                          |      |      |      |     |
                                        REJECT REJECT REJECT REJECT REJECT
                                        (Random)(System)(Random)(System)(Shift)

Detailed Analysis of Westgard Rules

1. $1_{2s}$ Rule (Warning Rule)

• Definition: A single control observation exceeds either the $+2SD$ or $-2SD$ limit.
• Classification: WARNING ONLY.
• Action: Do NOT reject the testing run based on a $1_{2s}$ violation alone. Inspect previous control data and evaluate subsequent rules ($1_{3s}, 2_{2s}, R_{4s}, 4_{1s}, 10_x$). If no other rules are violated, patient testing may proceed.

2. $1_{3s}$ Rule (Rejection Rule - Random Error)

• Definition: A single control observation exceeds either $+3SD$ or $-3SD$.
• Classification: REJECTION (Severe Random Error or Major Outlier).
• Action: Immediately halt patient testing. Reject the control run. Check for acute hardware failure, loose connections, or severe electrical interference.

3. $2_{2s}$ Rule (Rejection Rule - Systematic Error)

• Definition: Two consecutive control observations exceed the SAME $+2SD$ or $-2SD$ limit.
• Classification: REJECTION (Systematic Bias).
• Action: Halt testing. Indicates a systematic shift in baseline (e.g., miscalibrated syringe volume or altered gas mixture). Perform recalibration.

4. $R_{4s}$ Rule (Rejection Rule - Random Error)

• Definition: One control observation in a run exceeds $+2SD$ and another observation in the same run (or consecutive run) exceeds $-2SD$, creating a total difference of $\ge 4SD$ between them.
• Classification: REJECTION (High-Magnitude Random Error).
• Action: Halt testing. Indicates extreme analytical instability or operator variation.

5. $4_{1s}$ Rule (Rejection Rule - Systematic Error / Shift)

• Definition: Four consecutive control observations exceed the SAME $+1SD$ or $-1SD$ limit.
• Classification: REJECTION (Minor Systematic Bias / Shift).
• Action: Halt testing. Indicates a persistent minor shift in instrument baseline, such as electronic zero drift or subtle sensor degradation.

6. $10_x$ Rule (Rejection Rule - Systematic Trend / Bias)

• Definition: Ten consecutive control observations fall on the SAME side of the mean (all above or all below), regardless of whether they exceed $1SD$.
• Classification: REJECTION (Systematic Trend / Machine Drift).
• Action: Halt testing. Indicates long-term calibration drift, aging gas sensors, or altered ambient correction factors.

---

Biological Quality Control (BioQC)

While mechanical controls (3L syringes) verify hardware volume accuracy, they cannot assess gas conditioning, breathing circuit dynamics, software algorithms, or operator coaching technique. Biological Quality Control (BioQC) fills this critical gap.

BioQC Protocol

1. Subject Selection: Healthy, non-smoking laboratory personnel (at least 2 to 3 individuals) free from acute or chronic respiratory disease.
2. Baseline Establishment: Each biological control subject performs spirometry, lung volume, and $DL_{CO}$ testing repeatedly over 10 to 20 separate days to establish personal mean and $SD$ baselines.
3. Testing Frequency: BioQC subjects are tested weekly or whenever hardware controls pass but patient data appear clinically suspect.

BioQC Acceptability Limits

If a BioQC test falls outside acceptable limits while 3L syringe calibrations pass, suspect gas analyzer drift, software calculation changes, BTPS temperature sensor errors, or valve timing malfunctions.

---

Syringe Calibration Verification vs. System Quality Control

3-Liter Calibration Syringe Verification

ATS/ERS standards require daily volumetric calibration verification of all spirometers using a standardized 3.000 L calibration syringe.

• Flow Range Requirement: Syringe discharges must be performed at three distinct flow rates:
  1. Low Flow: 0.5 to 1.5 L/s (takes 2 to 6 seconds to inject).
  2. Medium Flow: 2.5 to 3.5 L/s (takes 1 to 2 seconds to inject).
  3. High Flow: 6.0 to 12.0 L/s (explosive injection in < 0.5 seconds).
• Acceptability Limit: Measured volume at each flow rate must be within $\pm 3%$ of 3.000 L (acceptable range: 2.910 L to 3.090 L, or within 0.090 L).

Corrective Action Protocols & Logbooks

Whenever a QC test or Westgard rule triggers a rejection:

1. Halt Patient Testing immediately.
2. Document Out-of-Control Event: Record date, time, instrument ID, specific rule violated, and control values in the official QC logbook.
3. Troubleshoot Root Cause: Inspect hardware, verify gas cylinder pressures, clean flow sensors, and check ambient temperature/barometric pressure sensors.
4. Perform Recalibration: Re-calibrate instrument and repeat QC testing.
5. Verify In-Control Status: Confirm that two consecutive QC runs meet all acceptance criteria before resuming patient testing.