What Is Laboratory Quality Control (QC) and How Do Westgard Rules Keep Test Results Reliable?

Laboratory quality control (QC) is the systematic process of verifying that analytical instruments and methods produce accurate, precise, and reliable patient results. It is a cornerstone of ISO/IEC 17025 laboratory quality management and is detailed in CLSI guideline C24 for statistical quality control. Without robust QC, a laboratory cannot confidently report results for diagnosis, treatment, or regulatory compliance.

What It Is

Laboratory QC involves running control materials—samples with known or assigned values—alongside patient samples. The measured control values are plotted on control charts (e.g., Levey-Jennings charts) and compared against pre-established acceptance limits, typically the mean ± 2 or ± 3 standard deviations (SD). The goal is to detect both random errors (imprecision) and systematic errors (bias) before they affect patient results.

The most widely adopted interpretive framework is the Westgard multirule QC, a set of decision criteria that flags runs based on multiple statistical rules. These rules are published, non-proprietary methods used globally in clinical and analytical laboratories.

How It Works / Formula or Steps

### Step 1: Establish the Control Mean and SD
For each control level (low, normal, high), run the control material repeatedly over 20–30 days. Calculate the mean (x̄) and standard deviation (s) from these data. These become the "target" values.

### Step 2: Plot Control Data on a Chart
Each day, run the control and plot the result on a Levey-Jennings chart. Draw horizontal lines at the mean, mean ± 1s, ± 2s, and ± 3s.

### Step 3: Apply Westgard Multirules
When a control value exceeds ± 2s, do not immediately reject the run. Instead, evaluate the following rules (using the common notation where "s" = standard deviation):

• 1₃ₛ rule: Reject the run if one control value exceeds the mean ± 3s. This signals a large random error or gross systematic error.
• 2₂ₛ rule: Reject the run if two consecutive control values (same level or different levels) both exceed the mean ± 2s on the same side of the mean. This signals a systematic error.
• R₄ₛ rule: Reject the run if one control value exceeds the mean + 2s and another (within the same run) exceeds the mean − 2s. This signals increased random error.
• 4₁ₛ rule: Reject the run if four consecutive control values exceed the mean ± 1s on the same side. This signals a trend or persistent bias.
• 10ₓ rule: Reject the run if ten consecutive control values fall on the same side of the mean (regardless of magnitude). This signals a slow drift or shift.

A run is accepted only if none of the applied rules are violated. Many laboratories use a simplified combination: reject on 1₃ₛ, 2₂ₛ, and R₄ₛ.

### Step 4: Document and Act
If a rule is violated, stop reporting patient results, investigate the cause (e.g., reagent change, calibration drift, temperature issue), correct it, and repeat the control before resuming testing.

A Worked Illustrative Example

Example data (illustrative only):
A glucose control has an established mean of 100 mg/dL and s = 4 mg/dL. Control limits are therefore: ± 2s = 100 ± 8 (92–108), ± 3s = 100 ± 12 (88–112).

• Day 1: Control value = 109 mg/dL. This exceeds + 2s (108) but is below + 3s (112). Apply the rules: 1₃ₛ No (109 < 112). 2₂ₛ Need two consecutive values > 108 on the same side—only one so far. R₄ₛ No value below − 2s. Result: Accept the run, but watch the next day.
• Day 2: Control value = 110 mg/dL. Now check 2₂ₛ: Day 1 (109) and Day 2 (110) both exceed + 2s on the same side. Rule violated → Reject the run. Investigate: recalibrate the instrument, rerun controls, and document the corrective action.

This example shows why single-point ± 2s rejection is too strict (causing false rejections ~5% of the time) and why Westgard rules balance error detection with efficiency.

Common Pitfalls

• Using only ± 2s as a single rule leads to excessive false rejections; always apply multirule logic.
• Not recalculating mean and SD after major maintenance or reagent lot changes—control targets drift.
• Ignoring Westgard rule violations because "the value is close" defeats the purpose of statistical QC.
• Running too few control levels—use at least two levels (low and high) to detect errors across the reportable range.

Closing

Laboratory QC is not optional paperwork—it is a scientific safeguard. By combining control charts with Westgard multirules, laboratories meet ISO/IEC 17025 and CLSI C24 expectations while protecting patient safety. To simplify daily QC tracking and rule evaluation, try the free Laboratory Quality Control (Lab QC) tool at https://www.6sq.com/tools/labqc/ and automate your Westgard rule checks.