Linearity and Range in Method Validation: How to Define, Test, and Report

By Dr. Zubair Khalid, DVM, MS, PhD · 2026-08-21

Quick Answer

• Define linearity as the proportional relationship between analyte concentration and measured response, then verify it with residual analysis and lack-of-fit testing instead of correlation coefficients alone.
• Establish range from the lowest to highest concentration where accuracy and precision meet predefined acceptance criteria, using spiked sample recovery and replicate measurements.
• Report linearity and range with the calibration model, regression equation, residual behavior, and the specific accuracy and precision values that justify each range boundary.

At a Glance

The Problem with R² as the Sole Linearity Criterion

Many validation protocols treat the correlation coefficient as the primary evidence of linearity. A high R² value can accompany a systematically curved response, especially when the data span a wide concentration range or when variance increases at higher concentrations. The correlation coefficient measures the strength of a linear association, but it does not confirm that the relationship is actually linear across the entire interval.

Residual analysis provides a more direct check. Residuals are the differences between the measured response and the response predicted by the regression line. When plotted against concentration, residuals should scatter randomly around zero. A pattern, such as a U-shape or a funnel shape, indicates that the linear model does not adequately describe the data across the full range.

The lack-of-fit test compares the variability of replicate measurements around their group means with the variability of those means around the regression line. If the regression line deviates significantly from the group means, the test flags a lack of fit. This approach requires replicate measurements at each concentration level, which is standard practice in validation studies.

A practical guide to validation and verification of analytical methods in the clinical laboratory emphasizes that linearity studies should be part of a broader validation framework that includes precision, bias, method comparison, and analytical sensitivity and specificity. The guide notes that laboratories often handle validation inconsistently, and that incorrect implementation can lead to false conclusions about method performance. For linearity specifically, the guide calls for clear statistical tests and acceptance criteria instead of reliance on a single coefficient.

Defining Linearity in Analytical Terms

Linearity in analytical chemistry means that the measured response is directly proportional to the analyte concentration within a defined interval. The relationship is expressed as a calibration equation, typically y = a + bx, where y is the response, x is the concentration, b is the slope, and a is the intercept. A zero intercept is not always required, and forcing the line through zero can mask curvature at low concentrations.

The linear range is the concentration interval where this proportional relationship holds. Outside this interval, the response may plateau, become nonlinear, or lose precision. The range is the broader interval where accuracy and precision meet acceptance criteria, which may include nonlinear portions of the response curve.

For diagnostic assays, the reportable range is the interval of analyte values that can be reported for patient samples. This range must be verified with matrix-appropriate samples, because the calibration curve prepared in buffer may not reflect behavior in serum, plasma, or other clinical matrices.

The distinction between linear range and reportable range matters in practice. A method may show linear response from 1 to 100 units in buffer, but matrix effects may limit the reportable range to 5 to 80 units in patient samples. Reporting results outside the verified range requires dilution protocols or reflex testing, and these procedures must themselves be validated.

Statistical Methods for Linearity Assessment

Ordinary Least Squares Regression

Ordinary least squares regression fits a line that minimizes the sum of squared residuals. This method assumes that the x-values (concentrations) are known without error and that the variance of the response is constant across the concentration range. In practice, both assumptions may be violated, and the analyst should examine the data for heteroscedasticity before relying on the regression output.

When variance increases with concentration, weighted least squares regression may be more appropriate. Weights are typically inversely proportional to the variance at each concentration level. This approach gives less influence to high-concentration points with larger variance and produces more reliable slope and intercept estimates.

Residual Plots

A residual plot shows the residuals on the y-axis and the concentration or predicted response on the x-axis. Random scatter around zero supports linearity. Systematic patterns indicate problems:

• A curved pattern suggests the relationship is nonlinear and a quadratic or other model may fit better.
• A funnel shape, where residuals spread wider as concentration increases, indicates heteroscedasticity.
• Clusters of residuals at specific concentrations may indicate matrix effects or calibration errors at those levels.

Residual plots should be examined for every validation run, also at the final acceptance stage. They provide immediate visual feedback that numerical summaries can miss.

Lack-of-Fit Test

The lack-of-fit test requires replicate measurements at each concentration level. The test partitions the residual sum of squares into pure error, which comes from the variation among replicates, and lack-of-fit error, which comes from the deviation of group means from the regression line. An F-test compares the lack-of-fit mean square to the pure error mean square.

A significant lack-of-fit indicates that the linear model does not adequately describe the data. The analyst should then consider whether a nonlinear model is appropriate, whether the range should be narrowed, or whether the method itself needs modification.

The lack-of-fit test has limitations. It requires sufficient replicates to estimate pure error reliably, and it may not detect subtle curvature if the concentration levels are too few or too widely spaced. The test should be combined with residual inspection and accuracy and precision data across the range.

Mandel's Fitting Test

Mandel's fitting test compares the residual variance of a linear fit with that of a quadratic fit. If the quadratic model significantly reduces the residual variance, the response is not linear across the tested range. This test is useful when the analyst suspects curvature but does not have replicate measurements at each level.

The test calculates an F-statistic from the difference in residual sums of squares between the two models. A significant result indicates that the quadratic term improves the fit enough to reject linearity. The analyst can then narrow the range or adopt a nonlinear calibration model.

Setting the Range Based on Accuracy and Precision

The range is not determined by the calibration curve alone. It is the concentration interval where the method meets predefined accuracy and precision criteria. These criteria should be established before the validation study begins, based on the intended use of the method and any regulatory or accreditation requirements.

Accuracy is typically assessed by recovery studies, where known amounts of analyte are spiked into matrix samples and the measured concentration is compared to the expected value. Recovery is expressed as a percentage, and acceptance limits are set based on the analyte concentration and the method's intended use.

Precision is assessed by replicate measurements at each concentration level. Repeatability, or intra-assay precision, is measured by one operator on one day. Intermediate precision, or inter-assay precision, is measured across multiple days or operators. Precision is expressed as the relative standard deviation (RSD), and acceptance limits are set based on the concentration level and the method's purpose.

A validation study of free formaldehyde determination in diphtheria-tetanus vaccine and tetanus toxoid antigen illustrates this approach. The study assessed accuracy and precision at different concentration levels by spiking known amounts of formaldehyde into real samples. Recovery expressed accuracy, and RSD expressed precision. Repeatability was performed by one operator in one day at three concentration levels, and intermediate precision was conducted by one operator on three separate days at the same three levels. The study reported RSD values below 5% at the medium concentration of the linear range, with accuracy between 90% and 110% at that level.

This example shows that range and linearity are assessed together with accuracy and precision. The range extends only as far as the data support acceptable recovery and reproducibility. A method may show linear response beyond the validated range, but if accuracy or precision fails at those concentrations, the range must be restricted.

Practical Workflow for Linearity and Range Validation

Step 1: Define the Intended Use and Acceptance Criteria

Before collecting data, specify what the method will measure, what matrix will be used, and what concentration interval is clinically or analytically relevant. Set acceptance criteria for accuracy, precision, and linearity based on the method's purpose and any applicable standards.

For diagnostic assays, the intended use determines the required performance. A screening assay may tolerate wider imprecision than a quantitative assay used for treatment decisions. The acceptance criteria should reflect these differences.

Step 2: Prepare Calibration Standards and Quality Control Samples

Prepare at least five to eight concentration levels spanning the expected range. Include a blank or zero calibrator when appropriate. Use matrix-matched calibrators whenever possible, because the response in buffer may differ from the response in the clinical matrix.

Prepare quality control samples at low, medium, and high concentrations within the expected range. These samples are used to assess accuracy and precision and should be independent of the calibration standards.

Step 3: Run Replicate Measurements

Run each calibration standard in duplicate or triplicate. The replicates are necessary for the lack-of-fit test and for estimating precision at each concentration level. Run the quality control samples in multiple replicates across multiple days to assess intermediate precision.

Document the operator, instrument, reagent lot, and date for each run. This documentation supports the precision assessment and provides a record for troubleshooting if problems arise.

Step 4: Fit the Regression Model and Examine Residuals

Fit the calibration data with ordinary least squares regression. Plot the residuals against concentration and inspect the plot for patterns. If the residual plot shows curvature or heteroscedasticity, consider whether a weighted regression or a nonlinear model is more appropriate.

Calculate the correlation coefficient and the coefficient of determination, but do not use these values as the sole evidence of linearity. Report the regression equation, the standard error of the estimate, and the residual behavior.

Step 5: Perform the Lack-of-Fit Test

Use the replicate measurements to perform the lack-of-fit test. If the test is significant, examine the residual plot and the accuracy and precision data to determine whether the range should be narrowed or whether a nonlinear model is needed.

A significant lack-of-fit does not necessarily invalidate the method. It may indicate that the linear range is narrower than the tested range, and the reportable range should be adjusted accordingly.

Step 6: Assess Accuracy and Precision Across the Range

Calculate recovery at each concentration level by comparing the measured concentration to the expected value. Calculate the RSD for repeatability and intermediate precision at each level. Compare these values to the predefined acceptance criteria.

The range is the interval where both accuracy and precision meet the criteria. If accuracy fails at the lowest or highest concentration, narrow the range. If precision fails at a specific level, investigate the cause before adjusting the range.

Step 7: Document the Results

Record the calibration model, the regression equation, the residual analysis, the lack-of-fit test result, and the accuracy and precision data for each concentration level. State the validated range and the criteria that support it.

Document any deviations from the protocol and the actions taken. This documentation supports the validation report and provides a basis for troubleshooting if the method fails during routine use.

Records and Measurements

Common Failure Patterns in Linearity and Range Validation

Overreliance on Correlation Coefficient

A high R² value can mask curvature, especially when the concentration range is wide or when the data include points near the detection limit. The correlation coefficient should be reported but never used as the sole criterion for linearity.

Forcing the Intercept to Zero

Some protocols require the calibration line to pass through zero. This practice can hide curvature at low concentrations and may produce biased results when the intercept is actually nonzero. The intercept should be estimated from the data and reported.

Testing Too Few Concentration Levels

A linearity study with three or four concentration levels cannot detect curvature that occurs between the tested points. At least five to eight levels are recommended, with replicates at each level to support the lack-of-fit test.

Ignoring Heteroscedasticity

When variance increases with concentration, ordinary least squares regression gives too much weight to high-concentration points. Weighted regression should be used, and the residual plot should be examined for a funnel pattern.

Extending the Range Beyond Validated Accuracy and Precision

A calibration curve may appear linear beyond the range where accuracy and precision are acceptable. The reportable range must be limited to the interval where all performance criteria are met.

Using Buffer Calibrators for Matrix Samples

Calibration curves prepared in buffer may not reflect the response in serum, plasma, or other clinical matrices. Matrix effects can shift the response and alter the linear range. Matrix-matched calibrators should be used whenever possible.

Linearity and Range in the Context of Method Validation

Method validation is a Good Manufacturing Practice principle that demonstrates a procedure actually leads to the expected results. The validation of free formaldehyde determination in biological products illustrates this principle in practice. The study was based on pharmacopoeial monographs and did not require full validation, but the suitability of the method under actual conditions of use was verified. Performance characteristics including accuracy, repeatability, intermediate precision, linearity, range, and limit of quantitation were investigated and calculated.

This example shows that the extent of validation depends on the method's purpose and regulatory context. A method based on a pharmacopoeial monograph may require verification instead of full validation, but the verification must still include linearity and range assessment under the actual conditions of use.

The practical guide to validation and verification of analytical methods in the clinical laboratory emphasizes that validation is required before a method is first implemented and during ongoing operation. The guide notes that laboratories often handle validation inconsistently, and that incorrect implementation can lead to false conclusions about method performance. For linearity studies specifically, the guide calls for appropriate statistical tests and clear acceptance criteria.

The Role of Regression in Broader Assay Development

Regression techniques extend beyond linearity assessment. Logistic regression, for example, is used to analyze the effect of independent variables on a binary outcome. The method quantifies each independent variable's unique contribution and identifies the strongest linear combination of variables for predicting the observed outcome.

Important considerations for logistic regression include selecting independent variables, ensuring that relevant assumptions are met, and choosing an appropriate model building strategy. Assumptions include independence of errors, linearity in the logit for continuous variables, absence of multicollinearity, and lack of strongly influential outliers. The number of events per independent variable should be adequate to avoid an overfit model, with commonly recommended minimums ranging from 10 to 20 events per covariate.

These principles apply to assay development more broadly. The analyst should select variables based on accepted theory, previous empirical investigations, and clinical considerations, and should acknowledge potential confounding variables. The same rigor applies to linearity assessment, where the choice of concentration levels, the number of replicates, and the statistical model all affect the validity of the conclusions.

Quality Controls and Ongoing Verification

Linearity and range are not established once and forgotten. The practical guide to validation and verification of analytical methods in the clinical laboratory emphasizes that method performance must be evaluated before first implementation and during ongoing operation. This ongoing evaluation includes periodic verification of linearity and range, especially after changes to reagents, instruments, or procedures.

Quality control samples at low, medium, and high concentrations should be included in each analytical run. Control results that drift outside established limits may indicate a change in the calibration curve or a shift in the linear range. The analyst should investigate the cause before reporting patient results.

Levey-Jennings charts and Westgard rules are commonly used to monitor quality control results over time. These tools detect trends and shifts that may indicate a change in method performance. When a trend is detected, the analyst should verify linearity and range before continuing to report results.

Biosafety and Regulatory Context

Validation studies involve handling biological samples, reagents, and potentially infectious materials. Biosafety principles apply to the entire workflow, from sample collection to disposal. The Biosafety in Microbiological and Biomedical Laboratories manual from CDC and NIH provides guidance on risk assessment, containment, and institutional oversight. The WHO Laboratory Biosafety Manual provides additional international guidance.

Institutional biosafety committees oversee work involving recombinant or synthetic nucleic acid molecules. The NIH Guidelines for Research Involving Recombinant or Synthetic Nucleic Acid Molecules describe the oversight requirements for such work. Analysts should confirm that their validation protocols comply with institutional requirements before beginning the study.

The National Institute of Standards and Technology provides measurement science, standards, calibration, and quality context that supports analytical method validation. NIST laboratory programs contribute to the development of reference materials and measurement standards that underpin reliable analytical results.

When to Escalate to Professional Oversight

The analyst should escalate to a laboratory director, quality manager, or institutional biosafety officer when:

• The lack-of-fit test is significant and the cause cannot be identified through routine troubleshooting.
• Accuracy or precision fails at multiple concentration levels across multiple runs.
• The residual plot shows a pattern that persists after reanalysis and recalibration.
• Matrix effects are suspected but cannot be resolved with available calibrators or dilution protocols.
• The validation data suggest that the method is not suitable for its intended use.

Escalation should include a summary of the data, the acceptance criteria, the observed results, and the troubleshooting steps already taken. The laboratory director or quality manager can determine whether the method requires modification, whether the range should be restricted, or whether an alternative method should be considered.

Limitations of Linearity and Range Studies

Linearity and range studies have inherent limitations. The results apply only to the matrix, instrument, reagent lot, and operator conditions tested. Changes to any of these variables may alter the linear range and require reverification.

The lack-of-fit test requires sufficient replicates to estimate pure error reliably. With too few replicates, the test may lack power to detect curvature. With too many concentration levels, the test may detect minor curvature that has no practical impact on results.

The range is defined by the accuracy and precision criteria chosen before the study. Different criteria may produce different ranges for the same method. The criteria should be justified by the intended use of the method and any applicable regulatory requirements.

Linearity and range studies do not address all aspects of method performance. Analytical sensitivity, specificity, carry-over, and reference ranges are separate validation parameters that must be assessed independently. The practical guide to validation and verification of analytical methods in the clinical laboratory addresses these topics alongside linearity studies.

Reporting Linearity and Range in Validation Documents

The validation report should state the linear range and the reportable range separately when they differ. For each range, the report should include:

• The calibration model and equation.
• The number of concentration levels and replicates.
• The residual plot and its interpretation.
• The lack-of-fit test result.
• The accuracy data, expressed as recovery at each concentration level.
• The precision data, expressed as RSD for repeatability and intermediate precision.
• The acceptance criteria and whether they were met.

The report should also describe any deviations from the protocol and the actions taken. This documentation supports the validity of the validation and provides a basis for troubleshooting if the method fails during routine use.

The MIQE guidelines for quantitative PCR provide a model for transparent reporting of analytical methods. The guidelines emphasize the need for detailed reporting of controls, reagents, and interpretation criteria. The same principles apply to validation reports for any analytical method.

Practical Implementation Steps for the Laboratory

Establish a Validation Protocol Before Data Collection

Write a protocol that specifies the concentration levels, the number of replicates, the acceptance criteria, and the statistical tests to be used. Review the protocol with the laboratory director or quality manager before beginning the study.

Use Matrix-Matched Calibrators

Prepare calibrators in the same matrix as the samples to be tested. If matrix-matched calibrators are not available, document the potential for matrix effects and verify the range with patient samples or matrix-appropriate quality control materials.

Include Sufficient Replicates

Run each concentration level in at least duplicate, and preferably triplicate, to support the lack-of-fit test and the precision assessment. Run quality control samples across multiple days to assess intermediate precision.

Examine Residuals Before Accepting the Model

Plot the residuals against concentration and inspect the plot for patterns. Do not accept the linear model based on R² alone. If the residual plot shows curvature or heteroscedasticity, consider whether the range should be narrowed or whether a different model is needed.

Document Everything

Record the calibration data, the regression equation, the residual plot, the lack-of-fit test result, and the accuracy and precision data. Document any deviations from the protocol and the actions taken. This documentation supports the validation report and provides a basis for troubleshooting.

Verify Linearity and Range Periodically

Recheck linearity and range after any change to reagents, instruments, or procedures. Monitor quality control results for trends or shifts that may indicate a change in the calibration curve. Investigate any change before reporting patient results.

A Decision Framework for Setting Range Boundaries When Linearity and Accuracy Conflict

Validation studies frequently produce data where the statistical linearity assessment and the accuracy or precision results do not align at the same concentration boundaries. A calibration curve may remain linear across ten concentration levels, but recovery may fall outside acceptance limits at the lowest two levels, or precision may deteriorate at the highest level. Analysts need a structured method for resolving these conflicts instead of making an arbitrary choice between the statistical linear range and the accuracy-based range.

The Conflict Between Statistical Linearity and Performance-Based Range

The linear range is a property of the calibration model. It answers the question of whether the response is proportional to concentration. The range is a property of the method's total performance. It answers the question of whether results at a given concentration are trustworthy for the intended use. These two properties frequently do not share the same boundaries.

Consider a method where the lack-of-fit test is not significant across concentrations from 2 to 200 units. The residual plot shows random scatter. The correlation coefficient exceeds 0.99. By statistical criteria, the linear range extends from 2 to 200 units. However, recovery at 2 units may be 130 percent because of matrix interference near the detection limit, and the relative standard deviation at 200 units may be 12 percent because of signal saturation or dilution errors. The performance-based range would be narrower, perhaps 5 to 150 units.

The validation report must state both intervals and explain which one governs reporting decisions. The linear range describes the calibration model. The reportable range describes what can be reported for samples. These are different claims, and confusing them creates a validation failure that auditors and regulators will identify.

A Tiered Decision Framework for Range Determination

The framework below provides a structured approach for resolving conflicts between linearity and accuracy or precision data. It uses three tiers of evidence and a clear escalation path.

Tier 1: Establish the Statistical Linear Range

Fit the calibration model and perform the lack-of-fit test. Examine the residual plot for curvature or heteroscedasticity. Record the concentration interval where the linear model holds without significant lack-of-fit. This interval is the candidate linear range.

If the lack-of-fit test is significant across the entire tested interval, do not proceed to range determination. The calibration model itself is inadequate. Investigate whether a quadratic model fits better, whether the instrument response has changed, or whether the concentration levels were too widely spaced to characterize the response curve.

Tier 2: Overlay Accuracy and Precision Data

For each concentration level within the statistical linear range, compare the recovery and relative standard deviation against the predefined acceptance criteria. Mark each level as pass or fail for accuracy and for precision. A level passes only when both criteria are met.

The performance-based range is the contiguous interval where all levels pass. If the lowest tested level fails accuracy but the next level passes, the range starts at the first passing level. If the highest tested level fails precision, the range ends at the last passing level.

Tier 3: Resolve Gaps Between the Two Ranges

When the statistical linear range and the performance-based range differ, apply the following rules in order.

First, restrict the reportable range to the performance-based range. The reportable range cannot extend beyond the interval where accuracy and precision meet acceptance criteria, regardless of the statistical linearity results. Reporting results outside the performance-based range requires a validated dilution protocol or reflex testing procedure.

Second, document the reason for the discrepancy. Common causes include matrix effects at low concentrations, signal saturation at high concentrations, insufficient replication at the range boundaries, or acceptance criteria that are tighter than the method can achieve at the extremes. The documentation should state the specific cause and the evidence supporting it.

Third, determine whether the discrepancy can be resolved by method modification. If low-concentration recovery fails because of matrix interference, a matrix-matched calibrator set or a different sample preparation procedure may improve recovery. If high-concentration precision fails because of dilution error, an automated dilution step may reduce variability. These modifications require a new validation run, not an adjustment to the acceptance criteria.

Fourth, if the discrepancy persists after method modification, accept the narrower performance-based range and report it as the validated range. The statistical linear range may be reported as supplementary information, but the reportable range governs sample reporting.

Applying the Framework to a Worked Example

A laboratory validates a quantitative immunoassay for a serum biomarker. The intended use is monitoring treatment response, and the clinically relevant interval is 10 to 500 ng/mL. The acceptance criteria are recovery within 85 to 115 percent and relative standard deviation below 10 percent for repeatability and below 15 percent for intermediate precision.

The analyst prepares eight concentration levels from 5 to 600 ng/mL, with triplicate measurements at each level. The lack-of-fit test is not significant across the full tested interval. The residual plot shows random scatter. The statistical linear range extends from 5 to 600 ng/mL.

Accuracy and precision data show the following pattern. At 5 ng/mL, recovery is 124 percent and the relative standard deviation for repeatability is 14 percent. At 10 ng/mL, recovery is 108 percent and the relative standard deviation is 8 percent. At 500 ng/mL, recovery is 96 percent and the relative standard deviation is 9 percent. At 600 ng/mL, recovery is 91 percent and the relative standard deviation is 16 percent.

Applying the framework, the performance-based range is 10 to 500 ng/mL. The 5 ng/mL level fails both accuracy and precision. The 600 ng/mL level fails precision. The reportable range is set at 10 to 500 ng/mL, which matches the clinically relevant interval.

The analyst documents that the low-concentration failure likely results from matrix interference near the detection limit and that the high-concentration failure likely results from dilution error. The laboratory director reviews the data and approves the reportable range. The validation report states the statistical linear range of 5 to 600 ng/mL and the reportable range of 10 to 500 ng/mL, with the accuracy and precision data supporting the narrower range.

Records Needed to Support Range Decisions

The validation record must contain enough data for an auditor or reviewer to reconstruct the range decision. The following records are required.

The range decision log is the most frequently missing record in validation files. Without it, a reviewer cannot determine why the reportable range differs from the statistical linear range or whether the discrepancy was investigated. The log should be a dated entry signed by the analyst and the laboratory director or quality manager.

Troubleshooting Persistent Range Conflicts

When the performance-based range is substantially narrower than the statistical linear range, the analyst should investigate the cause before accepting the narrower range. The following troubleshooting sequence addresses the most common causes.

Low-Concentration Failures

If recovery exceeds the upper acceptance limit at the lowest concentration levels, check whether the blank or zero calibrator response is subtracted correctly. A nonzero intercept in the calibration equation can cause overestimation at low concentrations. Examine the intercept estimate and its confidence interval. If the intercept is significantly different from zero, the calibration model should include the intercept instead of forcing the line through zero.

Check whether the lowest calibrator concentration is above the limit of quantitation. The limit of quantitation is the lowest concentration where precision and accuracy meet acceptance criteria. If the lowest tested level falls below the limit of quantitation, the failure is expected and the range should start at the limit of quantitation.

High-Concentration Failures

If precision fails at the highest concentration levels, check whether the response is approaching signal saturation. A plateau in the response at high concentrations produces increasing relative standard deviation as the signal change per unit concentration decreases. Examine the raw response values for evidence of saturation.

Check the dilution protocol. If high-concentration samples require dilution to bring them into the calibration range, the dilution step introduces additional variability. An automated dilution system may reduce this variability compared with manual dilution. The dilution protocol itself should be validated as part of the range determination.

Mid-Range Failures

If accuracy or precision fails at a middle concentration level but passes at lower and higher levels, investigate the specific level instead of narrowing the entire range. Possible causes include a preparation error for that calibrator, an evaporation or pipetting issue, or a matrix effect specific to that concentration. Repeat the measurements at that level before deciding whether the range must be split into two intervals.

When to Escalate Range Decisions

The analyst should escalate to the laboratory director or quality manager when the performance-based range is narrower than the clinically relevant interval, when the discrepancy between linear range and performance-based range cannot be explained by a identifiable cause, or when the range decision affects patient reporting.

Escalation should include the calibration data, the lack-of-fit test result, the accuracy and precision data at each level, the acceptance criteria, and the proposed reportable range. The laboratory director determines whether the method requires modification, whether an alternative method should be considered, or whether the narrower range is acceptable for the intended use.

The practical guide to validation and verification of analytical methods in the clinical laboratory notes that incorrect implementation of validation procedures can lead to false conclusions about method performance and potentially compromise patient safety. Range decisions directly affect which sample results are reported and which are flagged for dilution or reflex testing. A structured decision framework reduces the risk of arbitrary range boundaries and supports defensible validation documentation.

Integrating the Framework into the Validation Protocol

The decision framework should be written into the validation protocol before data collection begins. The protocol should specify the acceptance criteria for accuracy and precision, the statistical tests for linearity, and the rules for resolving conflicts between the statistical linear range and the performance-based range. This advance specification prevents post hoc adjustments to the range based on the observed data.

The protocol should also specify the number of concentration levels and replicates needed to support the range decision. At least five to eight concentration levels with triplicate measurements provide sufficient data for the lack-of-fit test and for estimating precision at each level. Fewer levels may not detect curvature or may produce imprecise estimates of recovery and relative standard deviation at the range boundaries.

The validation report should state the decision framework that was applied and the specific data that supported each range boundary. This transparency allows reviewers to verify that the range was determined by predefined criteria instead of by subjective judgment after the data were collected.

Frequently Asked Questions

What is the difference between linearity and range in method validation?

Linearity describes the proportional relationship between analyte concentration and measured response within a defined interval. Range is the broader concentration interval where the method meets predefined accuracy and precision criteria. A method may show linear response beyond the validated range, but the range is limited to the interval where all performance criteria are met.

Why is R² not sufficient to prove linearity?

The correlation coefficient measures the strength of a linear association, but it does not confirm that the relationship is actually linear across the entire interval. A high R² can accompany systematic curvature, especially when the concentration range is wide or when variance increases at higher concentrations. Residual plots and lack-of-fit tests provide more direct evidence of linearity.

How many concentration levels are needed for a linearity study?

At least five to eight concentration levels are recommended, with replicates at each level. Fewer levels cannot detect curvature that occurs between the tested points. Replicates are necessary for the lack-of-fit test and for estimating precision at each concentration level.

What is a lack-of-fit test and when should it be used?

The lack-of-fit test compares the variability of replicate measurements around their group means with the variability of those means around the regression line. A significant result indicates that the linear model does not adequately describe the data. The test requires replicate measurements at each concentration level and should be used whenever linearity is assessed.

How is the range determined in method validation?

The range is the concentration interval where accuracy and precision meet predefined acceptance criteria. Accuracy is assessed by recovery studies, where known amounts of analyte are spiked into matrix samples. Precision is assessed by replicate measurements, expressed as RSD for repeatability and intermediate precision. The range extends only as far as the data support acceptable recovery and reproducibility.

What should be done if the lack-of-fit test is significant?

Examine the residual plot and the accuracy and precision data to determine whether the range should be narrowed or whether a nonlinear model is needed. A significant lack-of-fit does not necessarily invalidate the method. It may indicate that the linear range is narrower than the tested range, and the reportable range should be adjusted accordingly.

How often should linearity and range be reverified?

Linearity and range should be reverified after any change to reagents, instruments, or procedures. Quality control results should be monitored continuously for trends or shifts that may indicate a change in the calibration curve. The practical guide to validation and verification of analytical methods in the clinical laboratory emphasizes that method performance must be evaluated during ongoing operation in addition to before first implementation.

What is the difference between linear range and reportable range?

The linear range is the concentration interval where the response is directly proportional to the analyte concentration. The reportable range is the interval of analyte values that can be reported for patient samples, which must be verified with matrix-appropriate samples. Matrix effects may limit the reportable range to a narrower interval than the linear range observed in buffer.

Related Diagnostic Guides

• Method Validation in Clinical Laboratories: A Practical Protocol
• Biosafety Level 4 Laboratories: Design, Operational Protocols, and Diagnostic Applications
• Biosafety Protocols for Diagnostic Specimen Transport: Ensuring Safety and Compliance
• Automated NGS Library Preparation: Platforms, Benefits, and Selection Criteria
• Cell Viability Assays: Choosing the Right Method for Your Experiment

References and Further Reading

• Molecular Biology and Laboratory Methods. National Library of Medicine.
• Biosafety in Microbiological and Biomedical Laboratories. CDC and NIH.
• NIH Guidelines for Recombinant or Synthetic Nucleic Acid Molecules. National Institutes of Health.
• Addgene Protocols. Addgene.
• MIQE Guidelines. Clinical Chemistry.
• NIST Laboratory Programs. National Institute of Standards and Technology.
• Laboratory Biosafety Manual. World Health Organization.
• NCBI Literature Resources. National Center for Biotechnology Information.
• A practical guide to validation and verification of analytical methods in the clinical laboratory.. Advances in clinical chemistry, 2019.
• Logistic regression: a brief primer.. Academic emergency medicine : official journal of the Society for Academic Emergency Medicine, 2011.
• Validation of Free Formaldehyde Determination Method in DT Vaccine and Tetanus Toxoid Antigen.. Archives of Razi Institute, 2023.

This article is educational and does not replace validated laboratory procedures, institutional biosafety review, manufacturer instructions, or professional interpretation.