Internal Quality Control (IQC): Complete Guide to Laboratory Quality Control
Learn the essential principles of internal quality control, the differences between QA, QC, IQC, and EQA, the complete laboratory QC workflow, and the major sources of pre-analytical, analytical, and post-analytical error.
1. Introduction to Internal Quality Control
Medical laboratory results influence diagnosis, treatment, medication adjustment, disease monitoring, transfusion decisions, and many other aspects of patient care. For this reason, laboratories must produce results that are accurate, precise, reliable, timely, and clinically useful.
Internal Quality Control, commonly abbreviated as IQC, is one of the most important tools used to monitor the analytical performance of laboratory testing systems. It helps laboratory professionals detect problems before incorrect patient results are reported.
Simple explanation
Internal Quality Control is like a daily safety check for laboratory testing. Before trusting patient results, the laboratory tests control materials with known or expected values to confirm that the analyzer, reagents, calibration, environment, and procedure are performing correctly.
IQC does not prevent every possible error. However, when it is designed, performed, interpreted, and documented correctly, it significantly reduces the risk of reporting analytically unreliable results.
2. What Is Internal Quality Control?
Internal Quality Control is the routine testing and evaluation of control materials within the laboratory to monitor the stability and performance of an analytical examination process.
Control materials are analyzed in a way that is similar to patient specimens. The obtained result is compared with a target value and an acceptable range. The laboratory then determines whether the analytical run is acceptable or whether investigation and corrective action are required.
Main objectives of IQC
- Detect significant analytical errors before patient results are released.
- Monitor the stability of an analyzer, method, reagent lot, and calibration.
- Identify shifts, trends, random errors, and systematic errors.
- Provide evidence that the examination process is under control.
- Support safe and consistent patient testing.
- Guide troubleshooting and corrective action.
- Provide documented evidence for quality assessment and accreditation.
Important principle
A control result is not simply “normal” or “abnormal.” It must be interpreted in relation to the established mean, standard deviation, control limits, previous performance, and the laboratory’s selected quality control rules.
3. QA vs QC vs IQC vs EQA
These terms are related, but they are not identical. Understanding the difference is essential for students, technicians, technologists, supervisors, and quality officers.
| Term | Meaning | Main Purpose | Practical Example |
|---|---|---|---|
| Quality Assurance (QA) | The complete system of planned activities used to ensure quality throughout the laboratory service. | To manage quality across the entire testing process. | Staff training, document control, internal audits, equipment maintenance, complaint management, and quality indicators. |
| Quality Control (QC) | Operational procedures used to evaluate and control the quality of laboratory testing. | To identify testing problems and maintain acceptable performance. | Running control materials and reviewing their results before reporting patient samples. |
| Internal Quality Control (IQC) | Quality control performed regularly within the laboratory. | To monitor day-to-day analytical stability and error. | Running two levels of chemistry control and plotting the results on a Levey–Jennings chart. |
| External Quality Assessment (EQA) | Evaluation of laboratory performance using specimens provided by an external organization. | To compare laboratory performance with peer laboratories, reference methods, or assigned targets. | Receiving unknown proficiency-testing samples and submitting the laboratory results for external evaluation. |
Internal Quality Control
- Performed inside the laboratory.
- Usually performed daily or according to the test schedule.
- Detects immediate or developing analytical problems.
- Supports run acceptance or rejection.
External Quality Assessment
- Organized by an external provider.
- Performed periodically.
- Compares performance with other laboratories or assigned targets.
- Evaluates longer-term comparability and competence.
IQC and EQA are complementary
A laboratory may have acceptable daily internal controls but still show poor comparability during external assessment. Similarly, good EQA performance does not replace daily internal quality control. Both are necessary components of a strong laboratory quality system.
4. Why Every Laboratory Needs IQC
Even a modern automated analyzer can produce an incorrect result. Problems may arise from reagent deterioration, calibration instability, pipetting faults, temperature variation, blocked tubing, optical failure, contamination, maintenance errors, or operator-related factors.
Without effective internal quality control, an analytical problem may remain undetected and affect a large number of patient results.
Clinical importance
Protects patients
Reliable laboratory results reduce the risk of incorrect diagnosis, inappropriate treatment, delayed care, and unnecessary follow-up procedures.
Supports medical decisions
Clinicians depend on stable laboratory performance when monitoring glucose, electrolytes, coagulation, blood counts, therapeutic drugs, cardiac biomarkers, and many other analytes.
Detects analytical instability
IQC can reveal gradual deterioration, sudden failures, changes in bias, increasing imprecision, shifts, and trends.
Supports accreditation
Documented quality control activities demonstrate that laboratory processes are monitored, evaluated, and corrected when necessary.
Patient safety principle
The purpose of quality control is not merely to produce acceptable control charts. Its real purpose is to reduce the probability that an incorrect laboratory result will negatively affect patient care.
5. Internal Quality Control Workflow
A reliable IQC process follows a structured sequence. Each step should be defined in the laboratory’s procedures and performed consistently by trained personnel.
Select appropriate control materials
Use control levels and concentrations that are appropriate for the clinical range and analytical method.
Prepare and handle controls correctly
Follow storage, reconstitution, mixing, stability, and handling instructions. Incorrect control preparation can create false QC failure.
Analyze the control material
Run the control using the same examination system used for patient samples, unless the validated procedure states otherwise.
Compare the result with established limits
Review the result against the control mean, standard deviation, control limits, chart pattern, and selected QC rules.
Accept or reject the analytical run
If performance is acceptable, patient testing may proceed. If the QC result is unacceptable, patient results should not be released until the cause has been evaluated and resolved.
Investigate failure
Review control preparation, reagent status, calibration, maintenance, environmental conditions, analyzer alarms, and previous QC patterns.
Take corrective action
Correct the identified problem, repeat quality control, and verify that performance has returned to an acceptable state.
Document the complete event
Record the failure, investigation, action taken, repeated results, responsible person, and final disposition of patient results.
Never repeat controls without investigation
Repeating the same control until it produces an acceptable result can hide an analytical problem. A failed control should trigger a logical investigation, not repeated testing without documented justification.
6. Types of Laboratory Errors
The total testing process can be divided into three major phases: pre-analytical, analytical, and post-analytical. Errors may occur in any phase, and not all of them can be detected by IQC.
| Testing Phase | When It Occurs | Examples | Can IQC Detect It? |
|---|---|---|---|
| Pre-Analytical | Before the specimen is analyzed | Incorrect patient identification, wrong tube, hemolysis, clotting, delayed transport, poor storage | Usually no. These require specimen-quality checks and process controls. |
| Analytical | During the measurement process | Reagent failure, calibration error, imprecision, contamination, analyzer malfunction | Often yes. This is the main area monitored by IQC. |
| Post-Analytical | After the result is generated | Transcription error, delayed reporting, wrong reference interval, failure to communicate a critical result | Usually no. These require reporting controls and verification procedures. |
7. Pre-Analytical Errors
Pre-analytical errors occur before the examination is performed. They may begin during test ordering, patient preparation, specimen collection, identification, transportation, processing, or storage.
Common pre-analytical errors
- Incorrect patient identification.
- Incorrect test request or missing clinical information.
- Use of the wrong collection tube or anticoagulant.
- Insufficient specimen volume.
- Incorrect blood-to-anticoagulant ratio.
- Hemolyzed, clotted, lipemic, or contaminated specimen.
- Prolonged tourniquet application.
- Collection from an intravenous line containing fluids or medication.
- Incorrect order of draw.
- Delayed transport or delayed separation of serum or plasma.
- Improper storage temperature.
- Exposure to light for light-sensitive analytes.
- Failure to follow fasting or timing requirements.
Why IQC usually cannot detect pre-analytical errors
Internal control materials do not pass through the complete patient collection process. Therefore, a control may produce a perfectly acceptable value while a patient specimen is unsuitable because of hemolysis, clotting, contamination, incorrect identification, or improper transport.
Example
A potassium control may be within range while a patient’s potassium is falsely increased because the blood specimen was severely hemolyzed. IQC confirms analytical stability, but it does not prove that every patient specimen is acceptable.
8. Analytical Errors
Analytical errors occur during the measurement process. This is the phase most directly monitored by internal quality control.
Common analytical errors
- Incorrect or unstable calibration.
- Deteriorated or improperly stored reagents.
- Incorrect reagent preparation.
- Analyzer pipetting failure.
- Blocked probes or tubing.
- Contamination or carryover.
- Incorrect incubation time or temperature.
- Optical, electrical, mechanical, or software malfunction.
- Failure to perform required maintenance.
- Control material deterioration.
- Incorrect instrument settings or test parameters.
- Operator technique error in manual methods.
Random and systematic analytical error
Random Error
Random error causes unpredictable variation in repeated measurements. It mainly affects precision.
Examples:
- Intermittent pipetting instability.
- Air bubbles in the measurement system.
- Inconsistent manual technique.
- Temporary electrical or mechanical disturbance.
Systematic Error
Systematic error shifts results consistently in one direction. It mainly affects accuracy or bias.
Examples:
- Incorrect calibration.
- Reagent lot bias.
- Gradual reagent deterioration.
- Temperature or wavelength calibration problem.
What IQC can show
Quality control charts and statistical rules may reveal isolated outliers, increased scatter, a sudden shift, a progressive trend, or repeated results on one side of the mean. These patterns help identify whether an analytical problem is likely random or systematic.
9. Post-Analytical Errors
Post-analytical errors occur after the examination result has been generated. A technically correct measurement can still harm the patient if it is reported incorrectly, delayed, assigned to the wrong patient, or interpreted using inappropriate information.
Common post-analytical errors
- Transcription or data-entry errors.
- Reporting the result for the wrong patient.
- Incorrect units or decimal placement.
- Use of an inappropriate reference interval.
- Failure to verify implausible or inconsistent results.
- Delayed result release.
- Failure to communicate critical values promptly.
- Incorrect automated comments or interpretation.
- Failure to document result correction.
- Interface problems between the analyzer and LIS.
Post-analytical quality requires result verification procedures, delta checks, critical-value communication policies, validated laboratory information systems, appropriate reference intervals, and clear documentation.
Mini Case Study
A laboratory’s sodium control results are acceptable. The patient result is correctly measured as 139 mmol/L, but it is entered manually into the system as 193 mmol/L.
This is a post-analytical transcription error. The IQC process cannot detect it because the analyzer performed correctly.
10. Key Takeaways from Part 1
- Internal Quality Control monitors the day-to-day analytical performance of laboratory examination systems.
- Quality Assurance is broader than Quality Control and includes the complete laboratory quality system.
- IQC is performed within the laboratory, while EQA evaluates performance through an external program.
- IQC is most effective for detecting analytical errors, but it does not detect every pre-analytical or post-analytical problem.
- Failed quality control requires investigation, corrective action, successful re-evaluation, and documentation.
- The ultimate goal of laboratory quality control is patient safety, not simply obtaining acceptable control values.
Coming in Part 2
Part 2 will explain control materials, control levels, target values, mean, standard deviation, coefficient of variation, accuracy, precision, bias, total error, and the basic principles of Sigma Metrics.
Internal Quality Control Statistics: Control Materials, Mean, SD, CV%, Accuracy, Precision, Bias, Total Error, and Sigma Metrics
Learn how laboratory control materials are selected and evaluated, how statistical limits are established, and how accuracy, precision, bias, total allowable error, and Sigma Metrics support reliable laboratory performance.
1. Control Materials
Control materials are specially prepared specimens with expected analyte concentrations. They are tested regularly to evaluate whether an examination system is performing within acceptable limits.
Control material is a stable material that contains one or more measurable components and is analyzed to monitor laboratory method performance.
A control material should behave as similarly as possible to a patient specimen during the analytical process. It should pass through the same instrument steps, reagent reactions, incubation conditions, detection systems, and calculations used for patient samples.
Ideal characteristics of control material
- Stable during the intended storage and use period.
- Homogeneous between bottles, vials, or aliquots.
- Compatible with the analytical method.
- Available at clinically meaningful concentrations.
- Similar to patient specimens whenever possible.
- Easy to prepare, store, mix, and analyze.
- Free from unexpected contamination or deterioration.
- Supported by documented target values or laboratory-established values.
Control material is not calibration material
A calibrator establishes the relationship between the instrument signal and analyte concentration. A control independently monitors whether that calibrated measurement system continues to perform acceptably.
2. Types of Control Materials
Control materials may be classified according to their source, preparation, target assignment, and relationship to the analytical system.
Assayed Controls
Assayed controls are supplied with manufacturer-assigned target values and expected ranges for specific methods or instruments.
- Convenient for initial implementation.
- May include method-specific target values.
- Manufacturer ranges may be wider than laboratory performance.
- Laboratories should still establish or verify their own mean.
Unassayed Controls
Unassayed controls do not provide definitive target values for the laboratory’s specific method.
- The laboratory establishes its own mean and SD.
- Useful for independent long-term monitoring.
- Requires sufficient initial measurements.
- Can provide a more realistic laboratory-specific range.
Independent Third-Party Controls
These controls are produced independently of the reagent or analyzer manufacturer.
- Reduce dependence on manufacturer-specific materials.
- May detect problems that matched-system controls fail to reveal.
- Useful across multiple reagent lots or instruments.
- Often support peer-group comparison programs.
Manufacturer-Matched Controls
These controls are designed by the same manufacturer as the analyzer or reagent system.
- Generally easy to use.
- May closely match the system’s expected performance.
- Can be optimized for a specific analytical platform.
- May be less independent from the complete reagent system.
Liquid, lyophilized, and frozen controls
| Control Format | Main Features | Advantages | Possible Limitations |
|---|---|---|---|
| Liquid Ready-to-Use | No reconstitution is usually required. | Reduces preparation error and saves time. | May have a shorter open-vial stability period. |
| Lyophilized | Freeze-dried material requiring reconstitution. | Often provides long unopened stability. | Reconstitution volume and mixing errors may affect results. |
| Frozen | Stored below freezing and thawed before use. | May preserve unstable analytes effectively. | Freeze-thaw cycles and incomplete thawing may create variation. |
3. Control Levels
Laboratories commonly use two or three control concentration levels. Each level evaluates analytical performance in a different clinical range.
| Control Level | Typical Concentration | Main Purpose | Example |
|---|---|---|---|
| Level 1 | Low or abnormal-low range | Evaluates method performance near low clinical decision limits. | Low hemoglobin, low glucose, or prolonged low platelet control. |
| Level 2 | Normal or mid-range | Monitors performance around common reference or therapeutic ranges. | Normal-range chemistry or hematology control. |
| Level 3 | High or abnormal-high range | Evaluates performance at elevated or clinically critical levels. | High glucose, high bilirubin, or high coagulation control. |
Why are multiple levels necessary?
A method may perform correctly at one concentration but poorly at another. For example, a glucose method may be stable near the normal range but show unacceptable bias at very low or very high concentrations.
The number and frequency of controls should be appropriate for the method, analytical risk, testing volume, test stability, manufacturer instructions, regulatory requirements, and the laboratory’s own quality control plan.
4. Control Preparation, Storage, and Handling
Incorrect control handling is a common cause of false QC rejection. Laboratories must treat control preparation as a standardized laboratory process.
Check identification and lot number
Confirm the control name, concentration level, lot number, expiration date, and compatibility with the intended method.
Follow storage requirements
Store unopened and opened materials at the specified temperature and protect them from light or contamination when required.
Reconstitute accurately
Use the exact diluent type and volume specified. Use calibrated volumetric equipment when required.
Allow complete stabilization
Follow the required standing time after reconstitution and mix carefully to achieve a homogeneous specimen.
Avoid contamination
Use clean pipettes and containers, close vials promptly, and avoid returning unused material to the original vial.
Record open-vial stability
Document the opening or reconstitution date, expiration after opening, initials, and any aliquot preparation details.
Common control-handling mistake
Repeated freezing and thawing can damage some analytes and increase variability. When appropriate, divide material into single-use aliquots to reduce freeze-thaw exposure.
5. Target Values and Acceptable Control Ranges
A target value is the expected central result for a control material. The acceptable control range defines how far a result may vary around that target before it requires evaluation.
Manufacturer Target
The target assigned by the control manufacturer using selected methods, instruments, or participating laboratories.
Laboratory Mean
The mean established by the individual laboratory using its own method, analyzer, reagent lot, environment, and operating conditions.
Peer-Group Mean
A comparison mean generated from laboratories using the same or similar method, instrument, reagent, and control lot.
Reference-Method Target
A value assigned using a higher-order reference measurement procedure or certified reference process.
Establishing a laboratory control mean
A laboratory may establish an initial mean by collecting repeated control results under stable operating conditions. A common educational approach is to collect at least 20 independent measurements over multiple days, although the laboratory should follow its validated protocol and applicable requirements.
Avoid using an artificially wide range
Manufacturer ranges may include variation from many laboratories and methods. Using an excessively wide range can reduce the ability of the QC system to detect clinically important analytical change.
6. Mean
The mean is the arithmetic average of a group of control results. It represents the central value around which control measurements are expected to cluster.
Mean formula:
Mean = Sum of all control results ÷ Number of results
x̄ = Σx ÷ n
Example
Five glucose control measurements are: 98, 101, 100, 99, and 102 mg/dL.
Sum = 98 + 101 + 100 + 99 + 102 = 500 mg/dL
Mean = 500 ÷ 5 = 100 mg/dL
Clinical laboratory meaning
If the established glucose control mean is 100 mg/dL, repeated results should generally fluctuate around this value. A persistent movement away from 100 mg/dL may indicate a shift, trend, or developing bias.
7. Standard Deviation
Standard deviation, abbreviated as SD, describes the dispersion of control results around the mean. It is one of the most important statistical measurements in laboratory quality control.
Standard deviation measures how closely repeated results are grouped around their mean.
Small SD
Results are closely grouped around the mean, indicating better precision.
Large SD
Results are widely dispersed, indicating greater imprecision or variability.
Control limits based on SD
For a normally distributed control process, control limits are commonly placed at one, two, and three standard deviations from the mean.
| Control Limit | Calculation | General Statistical Interpretation |
|---|---|---|
| Mean ±1 SD | Mean plus or minus one standard deviation | Approximately 68% of results are expected within this interval. |
| Mean ±2 SD | Mean plus or minus two standard deviations | Approximately 95% of results are expected within this interval. |
| Mean ±3 SD | Mean plus or minus three standard deviations | Approximately 99.7% of results are expected within this interval. |
Example
Suppose a glucose control has:
- Mean: 100 mg/dL
- SD: 2 mg/dL
| Limit | Lower Value | Upper Value |
|---|---|---|
| ±1 SD | 98 mg/dL | 102 mg/dL |
| ±2 SD | 96 mg/dL | 104 mg/dL |
| ±3 SD | 94 mg/dL | 106 mg/dL |
Important
A result located within ±2 SD is not automatically acceptable in every situation. QC interpretation must also consider previous results, multiple control levels, shifts, trends, and selected Westgard rules.
8. Coefficient of Variation (CV%)
The coefficient of variation expresses standard deviation as a percentage of the mean. It allows precision to be compared across analytes or control levels with different concentrations.
CV% = (SD ÷ Mean) × 100
Example
A glucose control has a mean of 100 mg/dL and an SD of 2 mg/dL.
CV% = (2 ÷ 100) × 100 = 2%
Lower CV%
Usually indicates less analytical variation and better precision.
Higher CV%
Usually indicates greater analytical variation and poorer precision.
CV must be interpreted appropriately
A low CV does not prove that results are accurate. A method can be highly precise but consistently biased. CV primarily evaluates imprecision, not closeness to the true value.
9. Accuracy and Precision
Accuracy and precision are related but different concepts. A high-quality analytical method should ideally demonstrate both.
Accuracy
Accuracy describes how close a measurement is to the correct, accepted, or reference value.
Poor accuracy is commonly associated with systematic error or bias.
Precision
Precision describes how closely repeated measurements agree with one another.
Poor precision is commonly associated with random error.
| Performance Pattern | Accuracy | Precision | Interpretation |
|---|---|---|---|
| Results close to target and closely grouped | Good | Good | Ideal analytical performance |
| Results away from target but closely grouped | Poor | Good | Consistent systematic error or bias |
| Results centered around target but widely scattered | May appear acceptable on average | Poor | Random error and high imprecision |
| Results away from target and widely scattered | Poor | Poor | Both systematic and random error may be present |
Practical Example
The assigned control target is 100 mg/dL. A laboratory repeatedly obtains 108, 109, 108, 109, and 108 mg/dL.
The values are closely grouped, so the method is precise. However, they are consistently higher than the target, so the method is not accurate.
This pattern suggests a positive systematic bias.
10. Bias
Bias is the consistent difference between the laboratory’s measured result and an accepted target or reference value.
Bias = Laboratory Mean − Target Value
Bias% = [(Laboratory Mean − Target Value) ÷ Target Value] × 100
Example
Laboratory mean = 104 mg/dL
Target value = 100 mg/dL
Bias = 104 − 100 = +4 mg/dL
Bias% = (4 ÷ 100) × 100 = +4%
Positive Bias
The laboratory result is consistently higher than the accepted target.
Negative Bias
The laboratory result is consistently lower than the accepted target.
Possible causes of bias
- Incorrect calibration.
- Reagent lot change.
- Calibrator lot change.
- Method-specific interference.
- Instrument temperature or wavelength problem.
- Incorrect assigned control target.
- Traceability or standardization differences.
- Matrix differences between control material and patient specimens.
11. Total Error and Total Allowable Error
Total analytical error combines the effect of systematic error and random error. It provides a broader assessment of the possible difference between a measured result and the correct value.
Total Error reflects the combined influence of bias and imprecision on analytical results.
Total Allowable Error
Total Allowable Error, abbreviated as TEa, is the maximum analytical error that may be accepted without creating an unacceptable risk for the clinical use of a laboratory result.
Observed Error
The actual analytical performance demonstrated by the laboratory, including measured bias and imprecision.
Allowable Error
The maximum analytical error permitted by the selected quality requirement or performance specification.
TEa sources
Laboratories may select analytical performance specifications from regulatory requirements, professional recommendations, biological variation models, clinical outcome models, manufacturer claims, or institutional risk assessment.
Do not mix incompatible quality specifications
The laboratory should clearly document the source and rationale for each selected TEa value. Different sources may define different allowable limits for the same analyte.
12. Sigma Metrics in Laboratory Quality Control
Sigma Metrics combine allowable error, bias, and imprecision into a single performance indicator. They help laboratories assess method capability and design an appropriate quality control strategy.
Sigma = (TEa% − |Bias%|) ÷ CV%
In this formula:
- TEa% is the selected total allowable error.
- Bias% is the method’s systematic difference from target.
- CV% represents analytical imprecision.
General interpretation of Sigma performance
| Sigma Level | General Performance | Possible QC Approach |
|---|---|---|
| ≥6 Sigma | Excellent or world-class method performance | Less intensive QC may be sufficient if risk assessment supports it. |
| 5 to <6 Sigma | Very good performance | Relatively simple multirule QC may provide effective monitoring. |
| 4 to <5 Sigma | Good performance | Increased control frequency or additional rules may be considered. |
| 3 to <4 Sigma | Marginal performance | Strong multirule QC and careful monitoring are generally required. |
| <3 Sigma | Poor method capability | Method improvement, additional controls, frequent QC, or method replacement may be necessary. |
Sigma calculation example
Assume:
- TEa = 10%
- Bias = 2%
- CV = 2%
Sigma = (10 − 2) ÷ 2
Sigma = 8 ÷ 2 = 4 Sigma
Interpretation
A 4-Sigma method may be clinically usable, but it generally requires a more careful QC strategy than a 6-Sigma method. The laboratory may need multiple control levels, appropriate Westgard rules, and sufficient QC frequency.
Sigma is not the only decision tool
Sigma Metrics should be interpreted with clinical risk, analytical frequency, stability, critical decision limits, patient population, reagent changes, instrument reliability, and regulatory requirements.
13. Complete Worked Quality Control Example
A laboratory evaluates a glucose method using the following data:
| Parameter | Value |
|---|---|
| Target value | 100 mg/dL |
| Laboratory mean | 102 mg/dL |
| Standard deviation | 2 mg/dL |
| Total allowable error | 10% |
Step 1: Calculate CV%
CV% = (SD ÷ Mean) × 100
CV% = (2 ÷ 102) × 100 = 1.96%
Step 2: Calculate Bias%
Bias% = [(102 − 100) ÷ 100] × 100
Bias% = 2%
Step 3: Calculate Sigma
Sigma = (10 − 2) ÷ 1.96
Sigma = 4.08
Final Interpretation
The glucose method demonstrates approximately 4-Sigma performance. Its precision is relatively good, but measurable positive bias is present.
The laboratory should maintain an appropriate multirule QC strategy and continue monitoring calibration, reagent lot performance, and long-term bias.
14. Practical Case Studies
Case Study 1: Good Precision but Poor Accuracy
A calcium control has a target of 9.5 mg/dL. Repeated measurements are 10.1, 10.2, 10.1, 10.2, and 10.1 mg/dL.
Interpretation: The results are closely grouped, showing good precision. However, they are consistently higher than the target, indicating positive bias and poor accuracy.
Possible action: Review calibration, calibrator lot, reagent lot, control target, and method comparison results.
Case Study 2: Poor Precision
A sodium control has a mean of 140 mmol/L. Results are 137, 143, 139, 145, 136, and 142 mmol/L.
Interpretation: The results show wide scatter around the mean. This suggests increased random error and poor precision.
Possible action: Check pipetting stability, air bubbles, probe condition, mixing, control preparation, instrument maintenance, and environmental stability.
Case Study 3: False QC Failure Due to Control Preparation
A lyophilized control suddenly produces results below the lower control limit. Patient results appear clinically consistent, analyzer checks are acceptable, and a newly prepared control is within range.
Interpretation: The original control may have been reconstituted with excessive diluent or inadequately mixed.
Lesson: Control failure does not always mean analyzer failure. Control preparation must be evaluated during troubleshooting.
Case Study 4: High Sigma but Incorrect Target
A laboratory calculates excellent Sigma performance using a very wide allowable error and an incorrect manufacturer target.
Interpretation: The calculated Sigma result may be misleading because the quality specification and target value were not appropriate.
Lesson: Statistical calculations are only reliable when the input data are valid, traceable, and clinically appropriate.
15. Key Takeaways from Part 2
- Control materials monitor analytical stability and should resemble patient specimens whenever possible.
- Two or three control levels are commonly used to evaluate performance across low, normal, and high concentration ranges.
- The mean represents the central control value, while standard deviation describes result dispersion.
- CV% expresses imprecision relative to the mean and is calculated as SD divided by mean multiplied by 100.
- Accuracy describes closeness to the accepted value, while precision describes agreement among repeated measurements.
- Bias represents systematic difference from a target value.
- Total Allowable Error defines the maximum analytical error permitted for the intended clinical use.
- Sigma Metrics combine TEa, bias, and CV to estimate method capability and guide QC design.
- Control failure may result from instrument, reagent, calibration, control preparation, storage, or operator-related problems.
Coming in Part 3
Part 3 will explain the Levey–Jennings chart, how to plot control results, how to identify shifts, trends, outliers, random error, systematic error, and how to interpret practical laboratory QC patterns step by step.
Internal Quality Control Statistics: Control Materials, Mean, SD, CV%, Accuracy, Precision, Bias, Total Error, and Sigma Metrics
Learn how laboratory control materials are selected and evaluated, how statistical limits are established, and how accuracy, precision, bias, total allowable error, and Sigma Metrics support reliable laboratory performance.
1. Control Materials
Control materials are specially prepared specimens with expected analyte concentrations. They are tested regularly to evaluate whether an examination system is performing within acceptable limits.
Control material is a stable material that contains one or more measurable components and is analyzed to monitor laboratory method performance.
A control material should behave as similarly as possible to a patient specimen during the analytical process. It should pass through the same instrument steps, reagent reactions, incubation conditions, detection systems, and calculations used for patient samples.
Ideal characteristics of control material
- Stable during the intended storage and use period.
- Homogeneous between bottles, vials, or aliquots.
- Compatible with the analytical method.
- Available at clinically meaningful concentrations.
- Similar to patient specimens whenever possible.
- Easy to prepare, store, mix, and analyze.
- Free from unexpected contamination or deterioration.
- Supported by documented target values or laboratory-established values.
Control material is not calibration material
A calibrator establishes the relationship between the instrument signal and analyte concentration. A control independently monitors whether that calibrated measurement system continues to perform acceptably.
2. Types of Control Materials
Control materials may be classified according to their source, preparation, target assignment, and relationship to the analytical system.
Assayed Controls
Assayed controls are supplied with manufacturer-assigned target values and expected ranges for specific methods or instruments.
- Convenient for initial implementation.
- May include method-specific target values.
- Manufacturer ranges may be wider than laboratory performance.
- Laboratories should still establish or verify their own mean.
Unassayed Controls
Unassayed controls do not provide definitive target values for the laboratory’s specific method.
- The laboratory establishes its own mean and SD.
- Useful for independent long-term monitoring.
- Requires sufficient initial measurements.
- Can provide a more realistic laboratory-specific range.
Independent Third-Party Controls
These controls are produced independently of the reagent or analyzer manufacturer.
- Reduce dependence on manufacturer-specific materials.
- May detect problems that matched-system controls fail to reveal.
- Useful across multiple reagent lots or instruments.
- Often support peer-group comparison programs.
Manufacturer-Matched Controls
These controls are designed by the same manufacturer as the analyzer or reagent system.
- Generally easy to use.
- May closely match the system’s expected performance.
- Can be optimized for a specific analytical platform.
- May be less independent from the complete reagent system.
Liquid, lyophilized, and frozen controls
| Control Format | Main Features | Advantages | Possible Limitations |
|---|---|---|---|
| Liquid Ready-to-Use | No reconstitution is usually required. | Reduces preparation error and saves time. | May have a shorter open-vial stability period. |
| Lyophilized | Freeze-dried material requiring reconstitution. | Often provides long unopened stability. | Reconstitution volume and mixing errors may affect results. |
| Frozen | Stored below freezing and thawed before use. | May preserve unstable analytes effectively. | Freeze-thaw cycles and incomplete thawing may create variation. |
3. Control Levels
Laboratories commonly use two or three control concentration levels. Each level evaluates analytical performance in a different clinical range.
| Control Level | Typical Concentration | Main Purpose | Example |
|---|---|---|---|
| Level 1 | Low or abnormal-low range | Evaluates method performance near low clinical decision limits. | Low hemoglobin, low glucose, or prolonged low platelet control. |
| Level 2 | Normal or mid-range | Monitors performance around common reference or therapeutic ranges. | Normal-range chemistry or hematology control. |
| Level 3 | High or abnormal-high range | Evaluates performance at elevated or clinically critical levels. | High glucose, high bilirubin, or high coagulation control. |
Why are multiple levels necessary?
A method may perform correctly at one concentration but poorly at another. For example, a glucose method may be stable near the normal range but show unacceptable bias at very low or very high concentrations.
The number and frequency of controls should be appropriate for the method, analytical risk, testing volume, test stability, manufacturer instructions, regulatory requirements, and the laboratory’s own quality control plan.
4. Control Preparation, Storage, and Handling
Incorrect control handling is a common cause of false QC rejection. Laboratories must treat control preparation as a standardized laboratory process.
Check identification and lot number
Confirm the control name, concentration level, lot number, expiration date, and compatibility with the intended method.
Follow storage requirements
Store unopened and opened materials at the specified temperature and protect them from light or contamination when required.
Reconstitute accurately
Use the exact diluent type and volume specified. Use calibrated volumetric equipment when required.
Allow complete stabilization
Follow the required standing time after reconstitution and mix carefully to achieve a homogeneous specimen.
Avoid contamination
Use clean pipettes and containers, close vials promptly, and avoid returning unused material to the original vial.
Record open-vial stability
Document the opening or reconstitution date, expiration after opening, initials, and any aliquot preparation details.
Common control-handling mistake
Repeated freezing and thawing can damage some analytes and increase variability. When appropriate, divide material into single-use aliquots to reduce freeze-thaw exposure.
5. Target Values and Acceptable Control Ranges
A target value is the expected central result for a control material. The acceptable control range defines how far a result may vary around that target before it requires evaluation.
Manufacturer Target
The target assigned by the control manufacturer using selected methods, instruments, or participating laboratories.
Laboratory Mean
The mean established by the individual laboratory using its own method, analyzer, reagent lot, environment, and operating conditions.
Peer-Group Mean
A comparison mean generated from laboratories using the same or similar method, instrument, reagent, and control lot.
Reference-Method Target
A value assigned using a higher-order reference measurement procedure or certified reference process.
Establishing a laboratory control mean
A laboratory may establish an initial mean by collecting repeated control results under stable operating conditions. A common educational approach is to collect at least 20 independent measurements over multiple days, although the laboratory should follow its validated protocol and applicable requirements.
Avoid using an artificially wide range
Manufacturer ranges may include variation from many laboratories and methods. Using an excessively wide range can reduce the ability of the QC system to detect clinically important analytical change.
6. Mean
The mean is the arithmetic average of a group of control results. It represents the central value around which control measurements are expected to cluster.
Mean formula:
Mean = Sum of all control results ÷ Number of results
x̄ = Σx ÷ n
Example
Five glucose control measurements are: 98, 101, 100, 99, and 102 mg/dL.
Sum = 98 + 101 + 100 + 99 + 102 = 500 mg/dL
Mean = 500 ÷ 5 = 100 mg/dL
Clinical laboratory meaning
If the established glucose control mean is 100 mg/dL, repeated results should generally fluctuate around this value. A persistent movement away from 100 mg/dL may indicate a shift, trend, or developing bias.
7. Standard Deviation
Standard deviation, abbreviated as SD, describes the dispersion of control results around the mean. It is one of the most important statistical measurements in laboratory quality control.
Standard deviation measures how closely repeated results are grouped around their mean.
Small SD
Results are closely grouped around the mean, indicating better precision.
Large SD
Results are widely dispersed, indicating greater imprecision or variability.
Control limits based on SD
For a normally distributed control process, control limits are commonly placed at one, two, and three standard deviations from the mean.
| Control Limit | Calculation | General Statistical Interpretation |
|---|---|---|
| Mean ±1 SD | Mean plus or minus one standard deviation | Approximately 68% of results are expected within this interval. |
| Mean ±2 SD | Mean plus or minus two standard deviations | Approximately 95% of results are expected within this interval. |
| Mean ±3 SD | Mean plus or minus three standard deviations | Approximately 99.7% of results are expected within this interval. |
Example
Suppose a glucose control has:
- Mean: 100 mg/dL
- SD: 2 mg/dL
| Limit | Lower Value | Upper Value |
|---|---|---|
| ±1 SD | 98 mg/dL | 102 mg/dL |
| ±2 SD | 96 mg/dL | 104 mg/dL |
| ±3 SD | 94 mg/dL | 106 mg/dL |
Important
A result located within ±2 SD is not automatically acceptable in every situation. QC interpretation must also consider previous results, multiple control levels, shifts, trends, and selected Westgard rules.
8. Coefficient of Variation (CV%)
The coefficient of variation expresses standard deviation as a percentage of the mean. It allows precision to be compared across analytes or control levels with different concentrations.
CV% = (SD ÷ Mean) × 100
Example
A glucose control has a mean of 100 mg/dL and an SD of 2 mg/dL.
CV% = (2 ÷ 100) × 100 = 2%
Lower CV%
Usually indicates less analytical variation and better precision.
Higher CV%
Usually indicates greater analytical variation and poorer precision.
CV must be interpreted appropriately
A low CV does not prove that results are accurate. A method can be highly precise but consistently biased. CV primarily evaluates imprecision, not closeness to the true value.
9. Accuracy and Precision
Accuracy and precision are related but different concepts. A high-quality analytical method should ideally demonstrate both.
Accuracy
Accuracy describes how close a measurement is to the correct, accepted, or reference value.
Poor accuracy is commonly associated with systematic error or bias.
Precision
Precision describes how closely repeated measurements agree with one another.
Poor precision is commonly associated with random error.
| Performance Pattern | Accuracy | Precision | Interpretation |
|---|---|---|---|
| Results close to target and closely grouped | Good | Good | Ideal analytical performance |
| Results away from target but closely grouped | Poor | Good | Consistent systematic error or bias |
| Results centered around target but widely scattered | May appear acceptable on average | Poor | Random error and high imprecision |
| Results away from target and widely scattered | Poor | Poor | Both systematic and random error may be present |
Practical Example
The assigned control target is 100 mg/dL. A laboratory repeatedly obtains 108, 109, 108, 109, and 108 mg/dL.
The values are closely grouped, so the method is precise. However, they are consistently higher than the target, so the method is not accurate.
This pattern suggests a positive systematic bias.
10. Bias
Bias is the consistent difference between the laboratory’s measured result and an accepted target or reference value.
Bias = Laboratory Mean − Target Value
Bias% = [(Laboratory Mean − Target Value) ÷ Target Value] × 100
Example
Laboratory mean = 104 mg/dL
Target value = 100 mg/dL
Bias = 104 − 100 = +4 mg/dL
Bias% = (4 ÷ 100) × 100 = +4%
Positive Bias
The laboratory result is consistently higher than the accepted target.
Negative Bias
The laboratory result is consistently lower than the accepted target.
Possible causes of bias
- Incorrect calibration.
- Reagent lot change.
- Calibrator lot change.
- Method-specific interference.
- Instrument temperature or wavelength problem.
- Incorrect assigned control target.
- Traceability or standardization differences.
- Matrix differences between control material and patient specimens.
11. Total Error and Total Allowable Error
Total analytical error combines the effect of systematic error and random error. It provides a broader assessment of the possible difference between a measured result and the correct value.
Total Error reflects the combined influence of bias and imprecision on analytical results.
Total Allowable Error
Total Allowable Error, abbreviated as TEa, is the maximum analytical error that may be accepted without creating an unacceptable risk for the clinical use of a laboratory result.
Observed Error
The actual analytical performance demonstrated by the laboratory, including measured bias and imprecision.
Allowable Error
The maximum analytical error permitted by the selected quality requirement or performance specification.
TEa sources
Laboratories may select analytical performance specifications from regulatory requirements, professional recommendations, biological variation models, clinical outcome models, manufacturer claims, or institutional risk assessment.
Do not mix incompatible quality specifications
The laboratory should clearly document the source and rationale for each selected TEa value. Different sources may define different allowable limits for the same analyte.
12. Sigma Metrics in Laboratory Quality Control
Sigma Metrics combine allowable error, bias, and imprecision into a single performance indicator. They help laboratories assess method capability and design an appropriate quality control strategy.
Sigma = (TEa% − |Bias%|) ÷ CV%
In this formula:
- TEa% is the selected total allowable error.
- Bias% is the method’s systematic difference from target.
- CV% represents analytical imprecision.
General interpretation of Sigma performance
| Sigma Level | General Performance | Possible QC Approach |
|---|---|---|
| ≥6 Sigma | Excellent or world-class method performance | Less intensive QC may be sufficient if risk assessment supports it. |
| 5 to <6 Sigma | Very good performance | Relatively simple multirule QC may provide effective monitoring. |
| 4 to <5 Sigma | Good performance | Increased control frequency or additional rules may be considered. |
| 3 to <4 Sigma | Marginal performance | Strong multirule QC and careful monitoring are generally required. |
| <3 Sigma | Poor method capability | Method improvement, additional controls, frequent QC, or method replacement may be necessary. |
Sigma calculation example
Assume:
- TEa = 10%
- Bias = 2%
- CV = 2%
Sigma = (10 − 2) ÷ 2
Sigma = 8 ÷ 2 = 4 Sigma
Interpretation
A 4-Sigma method may be clinically usable, but it generally requires a more careful QC strategy than a 6-Sigma method. The laboratory may need multiple control levels, appropriate Westgard rules, and sufficient QC frequency.
Sigma is not the only decision tool
Sigma Metrics should be interpreted with clinical risk, analytical frequency, stability, critical decision limits, patient population, reagent changes, instrument reliability, and regulatory requirements.
13. Complete Worked Quality Control Example
A laboratory evaluates a glucose method using the following data:
| Parameter | Value |
|---|---|
| Target value | 100 mg/dL |
| Laboratory mean | 102 mg/dL |
| Standard deviation | 2 mg/dL |
| Total allowable error | 10% |
Step 1: Calculate CV%
CV% = (SD ÷ Mean) × 100
CV% = (2 ÷ 102) × 100 = 1.96%
Step 2: Calculate Bias%
Bias% = [(102 − 100) ÷ 100] × 100
Bias% = 2%
Step 3: Calculate Sigma
Sigma = (10 − 2) ÷ 1.96
Sigma = 4.08
Final Interpretation
The glucose method demonstrates approximately 4-Sigma performance. Its precision is relatively good, but measurable positive bias is present.
The laboratory should maintain an appropriate multirule QC strategy and continue monitoring calibration, reagent lot performance, and long-term bias.
14. Practical Case Studies
Case Study 1: Good Precision but Poor Accuracy
A calcium control has a target of 9.5 mg/dL. Repeated measurements are 10.1, 10.2, 10.1, 10.2, and 10.1 mg/dL.
Interpretation: The results are closely grouped, showing good precision. However, they are consistently higher than the target, indicating positive bias and poor accuracy.
Possible action: Review calibration, calibrator lot, reagent lot, control target, and method comparison results.
Case Study 2: Poor Precision
A sodium control has a mean of 140 mmol/L. Results are 137, 143, 139, 145, 136, and 142 mmol/L.
Interpretation: The results show wide scatter around the mean. This suggests increased random error and poor precision.
Possible action: Check pipetting stability, air bubbles, probe condition, mixing, control preparation, instrument maintenance, and environmental stability.
Case Study 3: False QC Failure Due to Control Preparation
A lyophilized control suddenly produces results below the lower control limit. Patient results appear clinically consistent, analyzer checks are acceptable, and a newly prepared control is within range.
Interpretation: The original control may have been reconstituted with excessive diluent or inadequately mixed.
Lesson: Control failure does not always mean analyzer failure. Control preparation must be evaluated during troubleshooting.
Case Study 4: High Sigma but Incorrect Target
A laboratory calculates excellent Sigma performance using a very wide allowable error and an incorrect manufacturer target.
Interpretation: The calculated Sigma result may be misleading because the quality specification and target value were not appropriate.
Lesson: Statistical calculations are only reliable when the input data are valid, traceable, and clinically appropriate.
15. Key Takeaways from Part 2
- Control materials monitor analytical stability and should resemble patient specimens whenever possible.
- Two or three control levels are commonly used to evaluate performance across low, normal, and high concentration ranges.
- The mean represents the central control value, while standard deviation describes result dispersion.
- CV% expresses imprecision relative to the mean and is calculated as SD divided by mean multiplied by 100.
- Accuracy describes closeness to the accepted value, while precision describes agreement among repeated measurements.
- Bias represents systematic difference from a target value.
- Total Allowable Error defines the maximum analytical error permitted for the intended clinical use.
- Sigma Metrics combine TEa, bias, and CV to estimate method capability and guide QC design.
- Control failure may result from instrument, reagent, calibration, control preparation, storage, or operator-related problems.
Coming in Part 3
Part 3 will explain the Levey–Jennings chart, how to plot control results, how to identify shifts, trends, outliers, random error, systematic error, and how to interpret practical laboratory QC patterns step by step.
Westgard Rules: Complete Guide to QC Rule Interpretation, Error Detection, and Corrective Action
Learn how to interpret the 12s, 13s, 22s, R4s, 41s, 8x, 10x, and trend rules, distinguish random from systematic error, and respond correctly when laboratory quality control fails.
1. Introduction to Westgard Rules
Westgard rules are statistical decision rules used to evaluate internal quality control results. They help laboratory professionals decide whether an analytical run is acceptable or whether it should be rejected and investigated.
A single control result outside a statistical limit does not always mean that the analytical system has failed. Some results occur by chance even when the system is stable. Westgard multirules improve error detection by evaluating the position, sequence, and relationship of several control results.
Westgard rules are statistical quality-control rules designed to detect significant random and systematic analytical errors while reducing unnecessary rejection of acceptable analytical runs.
Why use several rules?
Different rules are sensitive to different types of error. For example, the R4s rule is particularly useful for detecting random error, while the 22s, 41s, 8x, and 10x rules are more sensitive to systematic error.
2. Warning Rules vs Rejection Rules
Quality-control rules can function as warning rules or rejection rules. Understanding this distinction prevents unnecessary repetition and inappropriate reporting.
Warning Rule
A warning rule alerts the laboratory professional that the control pattern requires closer evaluation.
- Does not automatically reject the run.
- Triggers review of additional control rules.
- Requires examination of recent QC history.
- May indicate an early developing problem.
Rejection Rule
A rejection rule indicates that the analytical run may be unreliable and should not be accepted without investigation.
- Patient results should generally be withheld.
- The source of error should be investigated.
- Corrective action should be documented.
- QC should be acceptable before reporting resumes.
Important principle
The 12s rule is commonly used as a warning rule, while rules such as 13s, 22s, R4s, and 41s are generally used as rejection rules.
3. The 12s Warning Rule
The 12s rule occurs when one control result exceeds the mean by more than 2 standard deviations, either above or below the mean.
Pattern
One control result is located outside the +2 SD or −2 SD limit, but it remains within ±3 SD.
Example
Mean = 100 mg/dL
SD = 2 mg/dL
+2 SD = 104 mg/dL
Control result = 105 mg/dL
Interpretation
The result violates the 12s warning rule. The run should not be rejected based on this result alone. Additional Westgard rules and recent QC results should be reviewed.
Recommended action
- Review the other control level or levels.
- Check whether another rejection rule is also violated.
- Review the previous QC sequence for shifts or trends.
- Check for recent reagent, calibration, or maintenance changes.
- Do not reject the run automatically based only on 12s.
Common mistake
Rejecting every control result beyond ±2 SD creates excessive false rejection because approximately 5% of normally distributed results may fall outside ±2 SD by chance.
4. The 13s Rejection Rule
The 13s rule occurs when one control result exceeds the mean by more than 3 standard deviations.
Pattern
A single control measurement lies beyond the +3 SD or −3 SD control limit.
Example
A sodium control has a mean of 140 mmol/L and an SD of 1 mmol/L.
The ±3 SD limits are 137–143 mmol/L.
A control result of 144 mmol/L violates the 13s rule.
Likely error type
The 13s rule may detect a large random error, although a major systematic error may also produce this pattern.
Recommended action
- Reject the analytical run.
- Do not release affected patient results.
- Check control preparation, mixing, expiration, and storage.
- Inspect analyzer alarms and mechanical performance.
- Review reagent, calibration, and maintenance status.
- Correct the cause and rerun all required control levels.
5. The 22s Rejection Rule
The 22s rule occurs when two consecutive control results exceed the same +2 SD or −2 SD limit.
Two common forms
Within-Run Violation
Two control levels in the same analytical run are both located beyond +2 SD or both beyond −2 SD.
Across-Run Violation
The same control level exceeds the same 2 SD limit in two consecutive analytical runs.
Likely error type
The 22s rule usually indicates systematic error.
Possible causes
- Calibration shift.
- Reagent lot bias.
- Control target or assigned-value problem.
- Temperature instability.
- Gradual reagent deterioration.
- Instrument wavelength or sensitivity change.
Example
Day 1 glucose control: +2.4 SD
Day 2 glucose control: +2.3 SD
Both results are above +2 SD. This violates the 22s rule and suggests a positive systematic shift.
6. The R4s Rejection Rule
The R4s rule occurs when the difference between two control results in the same run exceeds 4 standard deviations.
Typical pattern
One control result is above +2 SD, while another control result is below −2 SD in the same analytical run.
Example
Normal-level control = +2.2 SD
High-level control = −2.1 SD
The difference is 4.3 SD, violating the R4s rule.
Likely error type
The R4s rule is mainly associated with random error or excessive imprecision.
Possible causes
- Intermittent pipetting failure.
- Air bubbles in reagent or sample tubing.
- Incomplete control mixing.
- Probe obstruction.
- Electrical instability.
- Inconsistent manual technique.
- Temperature fluctuation.
Important interpretation point
The R4s rule should generally be evaluated within the same analytical run. Combining results from different runs may lead to an incorrect interpretation.
7. The 41s Rejection Rule
The 41s rule occurs when four consecutive control results exceed the same +1 SD or −1 SD limit.
Pattern
Four control results are all located more than 1 SD above the mean or all located more than 1 SD below the mean.
Likely error type
This pattern suggests a systematic shift.
Positive 41s
Four consecutive control results are above +1 SD.
Negative 41s
Four consecutive control results are below −1 SD.
Possible causes
- Small calibration change.
- New reagent lot with mild bias.
- Control lot change.
- Instrument adjustment or maintenance.
- Gradual environmental effect.
8. The 8x Rule
The 8x rule occurs when eight consecutive control results fall on the same side of the mean.
None of the individual results must exceed a specific SD limit. The significance comes from the repeated pattern on one side of the mean.
Interpretation
The 8x rule indicates a possible systematic shift or persistent bias.
Example
Eight consecutive potassium control results are located slightly above the mean, ranging from +0.3 SD to +1.4 SD.
Although every individual result remains within ±2 SD, the sequence may violate the 8x rule and suggests a developing positive shift.
9. The 10x Rule
The 10x rule occurs when ten consecutive control results are located on the same side of the mean.
Likely error type
The 10x rule is primarily used to identify a persistent systematic shift.
Possible causes
- Recalibration that changed the analytical mean.
- Reagent lot change.
- Control lot or target-value problem.
- Analyzer adjustment.
- Long-term change in method bias.
8x vs 10x
Some laboratories use an 8x rule, while others use a 10x rule. The selected rule should be defined in the laboratory’s QC plan and based on method performance, risk, and validated procedures.
10. Trend Rules
A trend occurs when consecutive control results progressively increase or progressively decrease over time.
A trend is a continuous directional movement in control values, even when the individual measurements remain inside control limits.
Common educational trend patterns
- Six consecutive results increasing.
- Seven consecutive results increasing.
- Six or seven consecutive results decreasing.
The exact number used depends on the laboratory’s selected QC procedure.
Possible causes of an upward or downward trend
- Gradual reagent deterioration.
- Progressive calibration drift.
- Temperature instability.
- Electrode aging.
- Lamp or optical-system deterioration.
- Progressive probe contamination.
- Control material deterioration.
Example
Control results over seven days are:
−0.8 SD, −0.4 SD, 0 SD, +0.3 SD, +0.7 SD, +1.1 SD, +1.6 SD
This progressive increase represents an upward trend and may indicate gradual analytical drift.
11. Random Error vs Systematic Error
| Characteristic | Random Error | Systematic Error |
|---|---|---|
| Effect | Causes unpredictable variation | Causes a consistent directional change |
| Main quality affected | Precision | Accuracy and bias |
| Typical chart appearance | Wide scatter or isolated extreme result | Shift, trend, or results on one side of mean |
| Common rules | 13s, R4s | 22s, 41s, 8x, 10x, trend |
| Examples | Air bubble, intermittent pipetting problem, unstable manual technique | Calibration shift, reagent bias, temperature drift |
Remember
Random error primarily reduces precision, while systematic error primarily reduces accuracy. However, real laboratory problems may produce mixed patterns, so investigation should not rely on one rule alone.
12. Westgard Multirule Strategy
A multirule QC procedure combines several rules to increase error detection while limiting false rejection.
Review the 12s warning rule
If no result exceeds ±2 SD, continue evaluating the overall pattern and accept the run when all selected rules are satisfied.
Check the 13s rule
Reject if one control result exceeds ±3 SD.
Check the 22s rule
Reject if two consecutive control results exceed the same ±2 SD limit.
Check the R4s rule
Reject if two controls in the same run differ by more than 4 SD.
Check the 41s rule
Reject if four consecutive control values exceed the same ±1 SD limit.
Check shift and trend rules
Review 8x, 10x, and directional trend patterns according to the laboratory’s approved procedure.
Not every laboratory needs the same rules
The ideal QC strategy depends on method performance, Sigma capability, clinical risk, analytical frequency, number of control measurements, test volume, and regulatory requirements.
13. Practical QC Decision Tree
Run all required control levels
Confirm correct control identity, lot, preparation, and instrument status.
Are all results within expected limits?
If yes, review patterns and previous QC history. If no, continue with rule evaluation.
Is only the 12s warning rule present?
Review the remaining rules before deciding whether to accept or reject the run.
Is a rejection rule violated?
If yes, stop patient-result release and investigate the analytical process.
Identify the probable error type
Determine whether the pattern suggests random error, systematic error, or control-material error.
Perform targeted troubleshooting
Avoid randomly changing multiple components at the same time.
Repeat QC after corrective action
Confirm that all required levels are acceptable before resuming patient reporting.
Evaluate affected patient results
Determine whether previously tested specimens require repeat analysis, amended reports, or clinician notification.
14. QC Troubleshooting Guide
| QC Pattern | Probable Problem | Recommended Checks |
|---|---|---|
| One result beyond ±3 SD | Large random error or major analytical failure | Control preparation, pipetting, analyzer alarms, reagent status, probe condition |
| Two results beyond the same ±2 SD limit | Systematic shift | Calibration, reagent lot, control lot, temperature, target value |
| One result above +2 SD and another below −2 SD | Random error or excessive imprecision | Pipetting, bubbles, mixing, blockage, electrical or mechanical instability |
| Several results on one side of mean | Persistent bias or shift | Recalibration history, reagent change, maintenance, target mean |
| Progressive increase or decrease | Trend or drift | Reagent deterioration, electrode aging, optical system, temperature |
| All analytes suddenly fail | General system or control-material problem | Control preparation, storage temperature, instrument condition, common reagent or system issue |
| Only one analyte fails | Analyte-specific reagent or calibration problem | Reagent lot, calibration factor, wavelength, assay parameters |
| Only one control level fails | Control-level deterioration or concentration-specific problem | Control vial, stability, reconstitution, linearity, reaction range |
Recommended troubleshooting sequence
- Confirm that the correct control level and lot were used.
- Check expiration date and open-vial stability.
- Verify storage temperature and preparation procedure.
- Inspect for bubbles, clots, contamination, or inadequate mixing.
- Review analyzer flags, alarms, maintenance, and environmental conditions.
- Check reagent lot, expiration, preparation, and onboard stability.
- Review calibration status and recent calibration changes.
- Repeat control only after a reasoned investigation or corrective action.
- Use fresh control material when control deterioration is suspected.
- Escalate unresolved problems according to laboratory procedure.
Avoid “repeat until acceptable”
Repeatedly testing the same control without investigating the cause may eventually produce an acceptable result by chance. This practice can hide a real analytical problem and create unsafe patient reporting.
15. Corrective Action and Documentation
Every rejected analytical run should be investigated and documented in a traceable manner.
Documentation should include
- Date and time of the QC failure.
- Analyzer, method, analyte, and control level involved.
- Control lot and reagent lot numbers.
- Exact QC result and violated rule.
- Probable error type.
- Investigation steps performed.
- Corrective action taken.
- Repeat QC results.
- Evaluation of affected patient results.
- Name or initials of responsible personnel.
- Supervisor review when required.
Immediate Correction
An action taken to restore acceptable analytical performance, such as preparing fresh control, recalibrating, replacing reagent, or cleaning a probe.
Corrective Action
A structured action intended to remove the root cause and reduce the chance that the problem will recur.
Evaluation of patient results
When a QC failure may have affected patient testing, the laboratory should determine:
- When the analytical problem likely began.
- Which patient specimens were tested during the affected period.
- Whether retained samples can be repeated.
- Whether differences are clinically significant.
- Whether corrected reports are required.
- Whether clinicians must be notified.
Quality improvement principle
A QC failure should not be viewed only as an isolated technical problem. Repeated failures should be analyzed for root causes, recurring patterns, staff-training needs, equipment reliability, and process improvement.
16. Practical Case Studies
Case Study 1: 12s Warning Only
Level 1 control is at +2.2 SD. Level 2 control is near the mean. No other Westgard rule is violated.
Interpretation: This is a 12s warning. Review previous results and other rules. The run may remain acceptable if the approved QC procedure allows it.
Case Study 2: R4s Violation
Low-level control is +2.4 SD and high-level control is −2.0 SD in the same run.
Interpretation: The difference exceeds 4 SD. This violates the R4s rule and suggests random error or excessive imprecision.
Action: Reject the run and inspect pipetting, bubbles, control mixing, probe condition, and analyzer stability.
Case Study 3: 22s Violation After Calibration
Two consecutive calcium-control results are above +2 SD following a new calibration.
Interpretation: This is a 22s violation and suggests a positive systematic shift.
Action: Review calibration preparation, calibrator lot, assigned values, reagent lot, and instrument settings.
Case Study 4: Progressive Trend
Potassium-control values increase gradually over seven runs but remain within ±2 SD.
Interpretation: A developing upward trend may be present. The system may be drifting even though no single result is outside the standard limits.
Action: Check electrode condition, calibration drift, reagent status, temperature, and instrument maintenance.
Case Study 5: Ten Results Above the Mean
Ten consecutive glucose-control measurements remain between +0.2 SD and +1.5 SD.
Interpretation: The 10x rule is violated. This suggests a persistent positive shift even though each result is individually inside ±2 SD.
Action: Investigate calibration, reagent lot, control target, and recent analyzer adjustments.
17. Key Takeaways from Part 4
- Westgard rules evaluate control results using statistical limits, sequences, and relationships between control levels.
- The 12s rule is usually a warning rule, not an automatic rejection rule.
- The 13s rule indicates one result beyond ±3 SD and generally requires rejection.
- The 22s rule suggests systematic error when two results exceed the same ±2 SD limit.
- The R4s rule mainly detects random error or excessive imprecision.
- The 41s, 8x, 10x, and trend rules help identify smaller systematic shifts and developing analytical drift.
- Failed QC requires investigation, corrective action, acceptable repeat controls, and evaluation of affected patient results.
- Repeating controls without investigating the cause is not an appropriate corrective action.
Coming in Part 5
Part 5 will cover the complete daily QC procedure, laboratory-specific applications in hematology, clinical chemistry, coagulation, immunology, blood bank, and molecular diagnostics, as well as ISO 15189:2022 principles, documentation, case studies, FAQs, references, author box, and structured data.