Method Validation Edition

Method validation software used in regulatory submissions worldwide From method comparison and precision through to linearity, reference intervals, and detection limits — 11 CLSI protocols in a single validated package, for IVD manufacturers, clinical laboratories, and regulatory professionals.

Trusted by 75,000 scientists at thousands of laboratories and 8 of the top-10 leading IVD manufacturers for over 30 years.

We use Analyse-it for the analysis of data necessary to file 510k. We chose Analyse-it because it works in Excel, includes CLSI protocols, and, unlike EP-Evaluator, lets us analyze data directly from equipment without typing.

Thomas D Harrigan, Ph.D.
Technical Product Manager
Alfa Wassermann Diagnostic Technologies

Analyse-it is developed for and is in use at thousands of ISO/IEC 17025 accredited testing and calibration laboratories, ISO 15189 accredited medical laboratories, CLIA ’88 regulated medical laboratories, and IVD manufacturers for development, support, product labelling and FDA 510(k) submissions.

Establish precision, linearity, bias, detection capability, and interferences during product development. Compare your method against a predicate or reference procedure. Establish reference intervals and diagnostic accuracy for product labelling. Verify manufacturer claims when bringing a new analyser or reagent system into clinical use. Produce the statistical evidence for a 510(k) submission, CE-IVD technical file, CAP inspection, or ISO 15189 audit.

Characterise measurement system performance

Establish the precision, trueness, linearity, interferences, and detection capability of a measurement procedure — whether you’re an IVD manufacturer characterising performance during development for a 510(k) submission, or a clinical laboratory verifying manufacturer claims:

Precision with flexible nested variance component designs — repeatability, between-run, between-day, between-operator, between-site, within-laboratory, and reproducibility (EP05-A3)
Precision profiles with six variance function models
Linearity with polynomial regression, Hsieh-Liu nonlinearity testing, and adjustable measuring interval for reportable range determination (EP06-A)
Detection capability — LoB, LoD, and LoQ using parametric, non-parametric, probit regression, and variance function approaches (EP17-A2)
Verification of precision and trueness against manufacturer claims — such as testing observed CV against a claimed 3.5% — with χ² and equivalence tests (EP15-A3)
Interference testing against allowable limits at clinically relevant concentrations (EP10-A3-AMD)

Measurement system analysis details →

Compare methods and estimate bias

When introducing a new measurement procedure, replacing an existing one, or comparing against a reference method, you need to know the bias and whether it affects clinical decisions:

Passing-Bablok regression (1983 and 1988) — non-parametric, robust to outliers, with CUSUM linearity test and bootstrap CIs
Deming and Weighted Deming regression — accounts for measurement error in both methods, systematic error decomposition into constant and proportional bias, Syx independent precision estimate
Bland-Altman limits of agreement — mean and median bias, constant and non-constant precision, linear fit for concentration-dependent bias, allowable difference band
Bias at specified clinical decision points with confidence intervals, equality and equivalence tests. Allowable difference as absolute, percentage, or combination — such as “10%, with a minimum of 5 mg/dL”
Total analytical error per EP21-A — combine bias + imprecision against allowable total error
Partition the measuring range into separate intervals, each with its own fit and allowable difference specification
Commutability per EP14-A3 — compare commutability of processed samples with prediction intervals

Method comparison details →

Bland-Altman, Deming, Passing-Bablok method comparison

Establish and verify reference intervals

Reference intervals are essential for clinicians to interpret results and make a diagnosis. Analyse-it covers the full EP28-A3C workflow — from choosing the right quantile method for your data and sample size, through outlier screening and partitioning, to transfer and verification against a new procedure or population.

Five quantile methods — parametric, non-parametric (three computation approaches), robust bi-weight, bootstrap, and Harrell-Davis
Partition by factors that affect the intervals — such as sex and age group for alkaline phosphatase
Full range of transformations — log, square root, cube root, Box-Cox, Manly exponential, two-stage exponential/modulus
Outliers with Tukey box plots, and normality assessment with Shapiro-Wilk, Anderson-Darling and frequency histogram
Transfer and verify intervals — typically 20 samples per partition per EP28-A3C

Reference interval details →

Evaluate diagnostic test accuracy

Establish and compare the ability of a diagnostic test to discriminate between outcomes. Explore how the test differentiates between positive and negative cases and determine optimum decision thresholds factoring in clinical costs of misdiagnosis.

Empirical and binormal ROC curves with DeLong AUC and confidence intervals (EP24-A2)
AUC comparison (DeLong) for up to 10 paired or independent tests — with equality, equivalence, and non-inferiority options
Sensitivity, specificity, likelihood ratios, predictive values, diagnostic odds ratio
Optimal threshold — Youden, closest-to-(0,1), cost-based
Decision plots of sensitivity vs specificity, likelihood ratios, predictive values, or cost across all possible thresholds
Qualitative test evaluation — proportion in positive/negative agreement, kappa and weighted kappa (EP12-A2)

Diagnostic performance details →

I started working with Analyse-it about 15 years ago. I was then working for a medical laboratory in method evaluation and ISO certification. The certification expert was delighted that I use the same statistical tool that he uses. Later in the pharma industry I applied it again to method evaluation and statistical work to analyse our projects. Now I have a company for clinical trials. Even tasks with ten thousands of numbers are no problem with Analyse-it.

Peter J.
Clinical Immunology FAMH

Includes the full Standard edition

All features from the Standard edition are included — ANOVA, PCA, regression, correlation, distribution analysis, hypothesis tests, and more.

So if, or when, a validation study raises questions — an unexpected distribution, a suspect outlier, a relationship that needs further examination — the full general statistics toolkit is right there in the same workbook.

Software you can trust

Validated calculations you can defend at inspection Every calculation is performed by Analyse-it — no Excel formulas, no third-party functions. Results are validated against CLSI reference datasets, published datasets, and thousands of internal test cases before every release. See how we develop and validate Analyse-it →

Data stays in your facility Analyse-it runs entirely within Microsoft Excel on your PC. No cloud processing, no data transmission. Pre-submission data, patient-adjacent data, and in-process results stay within your facility under your own data governance controls.

Standard Excel workbooks anyone can open Every analysis is an ordinary .xlsx workbook. Share with colleagues, submit to regulatory affairs, archive for audit, open on any PC with Excel. No proprietary format, no licence required to view results. Colleagues and auditors see exactly what you see.

Results that can’t be accidentally broken Analysis output contains computed values, not formulas. Nothing to accidentally overwrite, no cell references to break, no formula errors to introduce. The results you reported are exactly what you’ll find when you reopen the workbook months or years later for an audit.

Support for 11 CLSI protocols

Analyse-it has led the field in method validation statistics software for over 30 years. Many of the procedures now codified in CLSI guidelines — including Passing-Bablok regression, Weighted Deming regression, and non-parametric reference interval methods — were first made widely accessible through Analyse-it, and their adoption into formal guidelines followed.

EP05-A3	Evaluation of Precision of Quantitative Measurement Procedures	Precision
EP06-A	Evaluation of the Linearity of Quantitative Measurement Procedures	Linearity
EP09-A3	Measurement Procedure Comparison and Bias Estimation Using Patient Samples	Method comparison
EP10-A3-AMD	Preliminary Evaluation of Quantitative Clinical Laboratory Measurement Procedures	MSA
EP12-A2	User Protocol for Evaluation of Qualitative Test Performance	Diagnostic performance
EP14-A3	Evaluation of Commutability of Processed Samples	Method comparison
EP15-A3	User Verification of Precision and Estimation of Bias	MSA
EP17-A2	Evaluation of Detection Capability for Clinical Laboratory Measurement Procedures	Detection capability
EP21-A	Estimation of Total Analytical Error for Clinical Laboratory Methods	Method comparison
EP24-A2	Assessment of the Diagnostic Accuracy of Laboratory Tests Using Receiver Operating Characteristic Curves	Diagnostic performance
EP28-A3C	Defining, Establishing, and Verifying Reference Intervals in the Clinical Laboratory	Reference intervals

Example analyses

Download CLSI example datasets, open them in the trial, and see exactly what the output looks like.

EP09-A3 — Appendix I
Bland-Altman with partitioned measuring range. 0–1.8 µg/L (allowable difference ±0.06 µg/L) and 1.8–100 µg/L (allowable difference ±6%). Mean difference with 95% CI and equality hypothesis test for each interval.

EP09-A3 — Appendix I
All five method comparison fits. OLS, WLS, Deming, Weighted Deming, and Passing-Bablok on 79 observations. Bias at medical decision point 5 µg/L with CI and hypothesis test for each fit.

EP05-A3 — Appendix B
Glucose precision study, single site. 20 days × 2 runs × 2 observations. Repeatability and between-run/between-day variance components, ANOVA table, and variability of measurements plot.

EP05-A3 — Appendix B
CA19-9 precision study, multi-site. 6 samples × 3 laboratories × 5 runs × 5 observations. Repeatability, within-laboratory, and reproducibility components. Precision profile with 3-parameter variance function fit.

EP15-A3 — Table 8
Ferritin precision and trueness verification. 3 samples × 5 runs × 5 observations. Within-run and total precision with χ² tests against manufacturer’s claimed imprecision. Generalised ESD outlier detection.

EP06-A — Appendix C
IgM linearity evaluation. 5 dilutions × 2 observations. Linear and 2nd order polynomial fits. Difference plot with ±5% allowable nonlinearity band. Dilution 4 meets the requirement; dilutions 1, 2, 3, and 5 do not.

EP17-A2 — Appendix A
Estradiol detection capability. Parametric LoB from 5 blank samples and LoD from pooled SD of 5 low-level samples. Frequency density histogram with LoB and LoD lines.

EP28-A3C — Table 4
Calcium reference intervals. Partitioned by sex. Non-parametric (N+1)p quantile method, 120 observations per group. Histogram with reference limits and 90% CIs.

EP24-A2 — Appendix D
OxLDL and LDL diagnostic accuracy. Two paired tests. ROC curves with DeLong AUC and CIs. AUC comparison between tests. Bi-histogram and decision plot.

EP12-A2 — Example 10.3.1
H. pylori qualitative test evaluation. Two paired tests against true state. Sensitivity, specificity, predictive values with Wilson CIs. Score Z test.

Technical details

Method comparison

CLSI protocols

EP09-A3: Measurement Procedure Comparison and Bias Estimation Using Patient Samples
EP14-A3: Evaluation of Commutability of Processed Samples
EP21-A: Estimation of Total Analytical Error for Clinical Laboratory Methods

Quantitative methods (EP09-A3)

Singlicate, duplicate, and replicate measurements
Reduce measuring interval to linear range, or partition into multiple intervals with different relationships new in v4.00
Ordinary and Weighted linear regression average bias with confidence intervals
Deming and Weighted Deming regression average bias with jackknife confidence intervals
Passing-Bablok regression (1983 and 1988) average bias with Passing-Bablok or bootstrap confidence intervals
Bland-Altman limits of agreement with mean, median, and linear fit bias new in v3.75
Predict bias with confidence intervals at important decision levels
Test equality (no difference) or equivalence (difference within allowable difference) at decision levels
Compare commutability of samples with prediction intervals new in v4.90
Precision (SD or CV) and precision plots for each method
Pearson r correlation coefficient

Total analytical error (EP21-A)

Combine bias from method comparison with imprecision of test method
Assess against allowable total error
Allowable error as absolute, percentage, or combination

Qualitative methods

Proportion in positive/negative agreement (Clopper-Pearson exact, Wilson score CIs)
Kappa and Weighted Kappa (Wald Z CI)
Kappa test for agreement

Plots

Scatter plot with average bias, confidence bands, identity line, and allowable difference band
Scatter plot with allowable error bands
Vary colour of points by a factor
Difference / relative difference / ratio plot against X or mean of methods with allowable difference band and histogram
Mountain plot with allowable difference band new in v3.71
Residual plot (raw and standardised) with histogram
CUSUM linearity plot and Kolmogorov-Smirnov linearity test

Measurement systems analysis

CLSI protocols

EP05-A3: Evaluation of Precision of Quantitative Measurement Procedures
EP06-A: Evaluation of the Linearity of Quantitative Measurement Procedures
EP10-A3-AMD: Preliminary Evaluation of Quantitative Clinical Laboratory Measurement Procedures
EP15-A3: User Verification of Precision and Estimation of Bias
EP17-A2: Evaluation of Detection Capability for Clinical Laboratory Measurement Procedures

Experiment design

Flexible balanced and unbalanced experiment design: up to 3 random nested factors new in v4.60 and 1 fixed factor
Generalised ESD outlier identification new in v4.95

Precision (EP05-A3)

Precision as variance, SD, or CV% with exact, Satterthwaite, and MLS confidence intervals
Abbreviated reproducibility/repeatability, and detailed intermediate precision components
ANOVA table
χ² test against precision claim
Precision profile of SD or CV
Variance function fit: constant variance, constant CV, mixed constant/proportional, Sadler 3-parameter power, and 3-parameter alternative power new in v4.00
4-parameter variance function with turning point new in v4.91

Bias / trueness (EP15-A3, EP10-A3-AMD)

Bias with confidence interval new in v4.00
Test equality (no bias) or equivalence (bias within allowable bias)

Linearity (EP06-A)

Linear, polynomial (up to 5th order), forward stepwise polynomial, and best polynomial regression
Weighted models for non-constant SD across the measuring interval new in v5.00
Adjust measuring interval to find linear range new in v5.00
Difference between linear and nonlinear fit with Hsieh-Liu confidence intervals
Test equality or equivalence against allowable nonlinearity
Emancipator-Kroll linearity

Detection capability (EP17-A2) new in v4.00

LoB: parametric (SD of blank material), non-parametric (quantile of blank material), or from precision profile variance function
LoD: pooled SD of non-blank materials, from precision profile variance function, or probit regression
LoQ: from precision profile variance function

Plots

Scatter and difference plots
Variability of measurements plot
Precision profile with variance function
Difference plot of bias against assigned values with allowable bias bands
Linearity plot with polynomial fits
Difference plot with allowable nonlinearity band
Frequency density histogram with LoB and LoD
Probit regression curve

Reference intervals

CLSI protocols

EP28-A3C: Defining, Establishing, and Verifying Reference Intervals in the Clinical Laboratory

Establish reference limits

Normal (parametric) quantile
Non-parametric quantile: (N+1)p, Np+½, (N+⅓)p+⅓
Harrell-Davis quantile
Bootstrap quantile
Robust bi-weight quantile
Confidence intervals on all reference limits

Transfer / verify

Transfer using method comparison regression function
Binomial test for proportion inside reference interval

Partitioning & transformations

Partition by factor(s) new in v4.00
Reciprocal transformation
Log transformation
Square and cube root
Box-Cox new in v3.52
Manly exponential new in v4.00
2-stage exponential / modulus new in v4.00

Normality testing

Shapiro-Wilk test
Anderson-Darling test
Normal Q-Q plot

Plots

Frequency distribution histogram with normal overlay and reference limits
Tukey outlier box plot
Normal Q-Q plot with Lilliefors confidence band

Diagnostic performance

CLSI protocols

EP24-A2: Assessment of the Diagnostic Accuracy of Laboratory Tests Using Receiver Operating Characteristic Curves
EP12-A2: User Protocol for Evaluation of Qualitative Test Performance

ROC analysis (EP24-A2)

1 test, up to 10 paired tests, or up to 10 independent tests/groups
Empirical and binormal ROC curves
AUC with DeLong-DeLong-Clarke-Pearson CIs
AUC comparison (DeLong) for paired or independent tests
Partial AUC over specified FPR range
Sensitivity, specificity, PPV, NPV, likelihood ratios, odds ratio, Youden index
Optimal threshold (Youden, closest-to-(0,1), cost-based)

Qualitative tests (EP12-A2)

Sensitivity, specificity, prevalence
Positive/negative agreement
Kappa and weighted kappa

Plots

ROC curve with confidence band
Overlaid ROC curves for test comparison
Decision plot (sensitivity/specificity vs threshold)
Dot plot

System requirements

Microsoft Excel 2013, 2016, 2019, 2021, 2024 and Microsoft 365 for Microsoft Windows (32- and 64-bit)
Microsoft Windows 8, 10, 11, Server 2016, 2019, 2022
2 GB RAM minimum recommended
80 MB disk space