← Clinical Trial Results
Migraine Phase 3 Randomized NCT02614183

EVOLVE-1: Complete Statistical Analysis of Galcanezumab in Migraine

An independent statistical analysis of the phase 3 EVOLVE-1 study evaluating galcanezumab in the prevention of episodic migraine, with emphasis on its randomized parallel-group design, longitudinal mixed-effects analysis, responder odds ratios, Fisher exact testing, and interpretation of uncertainty.

EVOLVE-1  ·  NCT02614183  ·  Completed  ·  Enrollment 862
Scope of this record

This page separates reported trial results from statistical interpretation. Numerical results and trial-design facts are restricted to the ClinicalTrials.gov record. Where the registry does not report a detail, it is not reconstructed from outside sources.

Registry note: This page provides an independent statistical analysis and educational interpretation of publicly reported results. ClinicalTrials.gov provides the official trial registry record.

1. Trial at a Glance

EVOLVE-1 was a completed phase 3, randomized, double-masked, parallel-group trial evaluating galcanezumab versus placebo in migraine. The registry reports an enrollment of 862 participants and three arms, with galcanezumab 120 mg, galcanezumab 240 mg, and placebo represented in the posted safety data.

862
Enrollment
Participants
3
Arms
Parallel design
2
Primary comparisons
120 mg and 240 mg vs placebo
<.001
Primary P-values
Both reported comparisons
FeatureEVOLVE-1
Trial nameEVOLVE-1
Brief titleEvaluation of Galcanezumab in the Prevention of Episodic Migraine- the EVOLVE-1 Study
PhasePhase 3
ConditionMigraine
AllocationRandomized
Design modelParallel
MaskingDouble
Primary purposeTreatment
Enrollment862
InterventionsGalcanezumab; placebo
Results postedYes
Outcome measures posted10
Statistical analyses posted22
ClinicalTrials.govNCT02614183

2. Clinical Question

The statistical question is whether treatment with galcanezumab is associated with a different change in migraine-related outcomes than placebo over the registered assessment periods.

Population

Participants enrolled in the phase 3 EVOLVE-1 study for the prevention of episodic migraine.

Intervention

Galcanezumab, represented in the registry results by 120 mg and 240 mg treatment groups.

Comparator

Placebo.

Primary question

How did galcanezumab compare with placebo for the overall mean change from baseline in the number of monthly migraine headache days from baseline through Month 6?

3. Trial Design

01
Randomize862 enrolled
02
3 armsPlacebo and two galcanezumab groups
03
Double maskedParallel-group design
04
Longitudinal assessmentBaseline through Month 6
05
Outcome analysisMixed-effects and categorical methods
ARM · GALCANEZUMAB 120 MG

Galcanezumab 120 mg

  • Galcanezumab treatment arm
  • Included in the primary efficacy comparison with placebo
  • Serious adverse events: 6/206 during the treatment phase
  • Serious adverse events: 4/183 during the post-treatment phase
ARM · GALCANEZUMAB 240 MG

Galcanezumab 240 mg

  • Galcanezumab treatment arm
  • Included in the primary efficacy comparison with placebo
  • Serious adverse events: 0/220 during the treatment phase
  • Serious adverse events: 4/185 during the post-treatment phase
ARM · PLACEBO

Placebo

  • Placebo comparator arm
  • Used as the reference group for the posted efficacy comparisons
  • Serious adverse events: 5/432 during the treatment phase
  • Serious adverse events: 2/372 during the post-treatment phase
Allocation caution: The ClinicalTrials.gov record reports three arms and the arm-specific denominators used in the safety results, but they do not provide a randomized allocation ratio. The denominators in the safety tables should therefore not be used to infer the original randomization ratio.

4. Trial Timing and Registry Status

November 30, 2015

Study start

The registry lists the study start date as 2015-11-30.

March 22, 2017

Primary completion

The registry lists the primary completion date as 2017-03-22.

Completed

Registry status

The ClinicalTrials.gov record identifies EVOLVE-1 as completed and reports posted results.

5. Primary Endpoint

EndpointRegistry definition / time frameAnalysis
Overall Mean Change From Baseline in the Number of Monthly Migraine Headache Days Baseline, Month 1 through Month 6 Mixed Models Analysis; mean difference

The registry defines a Migraine Headache Day (MHD) as a calendar day on which a migraine headache or probable migraine headache occurred. The posted primary analysis uses the longitudinal outcome across baseline and Month 1 through Month 6 rather than reducing the entire study to a single unadjusted endpoint measurement.

Analysis population: The primary analyses included all randomized participants who received at least one dose of study drug and had baseline and at least one post-baseline value. The registry therefore defines the analyzed population more narrowly than the enrollment count of 862.

6. Statistical Methodology

Mixed-effects model for repeated measurements

The primary endpoint was measured longitudinally, with observations spanning baseline through Month 6. The registry reports a Mixed Models Analysis, normalized here as a mixed-effects model.

A mixed-effects model is useful when the same participants contribute multiple measurements over time. Rather than treating each time point as an unrelated analysis, the model can represent the repeated-measure structure and estimate treatment differences while using the longitudinal information available from participants with qualifying observations.

Conceptual longitudinal model
Yij = fixed effects + subject-specific effects + error

Here, Yij represents an outcome measured for participant i at time j. The exact covariance structure and fixed-effect specification are not provided in the ClinicalTrials.gov record and therefore are not reconstructed here.

Mean difference

The primary treatment effect is reported as a mean difference in final values. The direction matters. Because the outcome is change in monthly migraine headache days, a negative treatment-versus-placebo difference indicates a more negative change in the galcanezumab group relative to placebo.

Interpretation
Mean difference = galcanezumab outcome − placebo outcome

For this endpoint, a negative estimate indicates a greater reduction in monthly migraine headache days for the galcanezumab group under the reported comparison.

Odds ratios for responder outcomes

The registry also reports odds ratios for binary responder endpoints defined by reductions from baseline of at least 50%, at least 75%, and 100% in monthly migraine headache days. An odds ratio greater than 1 means the estimated odds of meeting the specified responder criterion were higher in the galcanezumab group than in placebo under the reported comparison.

Fisher exact test

For anti-drug antibody outcomes, the registry reports Fisher exact testing. This is an exact method for comparing categorical proportions between two groups and is particularly useful when event counts are sparse enough that large-sample approximations may be unreliable.

Superiority framework

The statistical analyses posted on ClinicalTrials.gov are classified as superiority. The reported primary comparisons therefore ask whether the observed outcome differs between galcanezumab and placebo, rather than whether galcanezumab falls within a prespecified non-inferiority margin.

7. Primary Results: Monthly Migraine Headache Days

Galcanezumab 120 mg vs Placebo

Mean difference in change from baseline

-1.92 days

95% CI: -2.48 to -1.37   ·   P < .001

Mixed Models Analysis; baseline, Month 1 through Month 6

Clinical Biostats interpretation

The reported mean difference of -1.92 days means that the modeled overall mean change in monthly migraine headache days was 1.92 days more negative for the galcanezumab 120 mg group than for placebo, according to the registry's treatment comparison.

It does not mean that every participant experienced exactly 1.92 fewer migraine headache days. It is a group-level estimated difference from a longitudinal model.

The 95% confidence interval from -2.48 to -1.37 describes statistical uncertainty around the estimated mean difference under the analysis framework. It does not describe the range of individual treatment responses.

The P < .001 value addresses the statistical evidence against the null hypothesis specified for the comparison; it is not a measure of the magnitude or clinical importance of the effect. Effect size and statistical evidence should be considered separately.

Because the endpoint was analyzed longitudinally, interpretation depends on the mixed-effects model specification and the assumptions underlying the handling of repeated observations and missingness. Those model details are not fully reported in the ClinicalTrials.gov record.

Galcanezumab 240 mg vs Placebo

Mean difference in change from baseline

-1.76 days

95% CI: -2.31 to -1.20   ·   P < .001

Mixed Models Analysis; baseline, Month 1 through Month 6

Clinical Biostats interpretation

The reported mean difference of -1.76 days indicates a more negative modeled change in monthly migraine headache days for the galcanezumab 240 mg group than for placebo.

The estimate is a population-level treatment comparison, not an individual prediction. A participant's observed change can be substantially different from the group mean difference.

The 95% CI of -2.31 to -1.20 quantifies uncertainty around the estimated group difference. Its width provides information about precision; it does not indicate that individual effects must lie inside the interval.

The P < .001 result indicates strong statistical evidence for a difference under the reported superiority analysis. It should not be interpreted as a percentage probability that the treatment works, nor as a measure of effect size.

The two primary comparisons involve two galcanezumab doses against placebo. The ClinicalTrials.gov record does not describe an alpha-allocation or multiplicity-adjustment procedure for these two primary comparisons, so no additional multiplicity conclusion is imposed here.

Primary comparisonEstimate95% CIP-valueMethod
Placebo vs Galcanezumab 120 mg-1.92-2.48 to -1.37<.001Mixed-effects model
Placebo vs Galcanezumab 240 mg-1.76-2.31 to -1.20<.001Mixed-effects model

8. Secondary Efficacy Results: Responder Outcomes

The registry reports a secondary endpoint measuring the mean percentage of participants with reductions from baseline of at least 50%, at least 75%, and 100% in monthly migraine headache days. The analyses posted on ClinicalTrials.gov report odds ratios for individual responder thresholds.

Responder thresholdComparisonOdds ratio95% CI
Reduction ≥50%Galcanezumab 120 mg vs placebo2.632.05 to 3.37
Reduction ≥50%Galcanezumab 240 mg vs placebo2.481.94 to 3.18
Reduction ≥75%Galcanezumab 120 mg vs placebo2.652.04 to 3.45
Reduction ≥75%Galcanezumab 240 mg vs placebo2.622.01 to 3.41
Reduction =100%Galcanezumab 120 mg vs placebo2.801.96 to 4.01
Reduction =100%Galcanezumab 240 mg vs placebo2.611.81 to 3.75
How to read the odds ratios

An odds ratio of 2.63, for example, means that the estimated odds of meeting the ≥50% reduction criterion were 2.63 times the corresponding odds in the placebo group for the reported galcanezumab 120 mg comparison.

An odds ratio is not a risk ratio. If the responder outcome is common, the odds ratio can be farther from 1 than the corresponding ratio of probabilities. Without the underlying responder counts, the odds ratios should not be converted into absolute response percentages.

The confidence intervals provide information about precision. For the ≥50% endpoint, the intervals are 2.05 to 3.37 for 120 mg and 1.94 to 3.18 for 240 mg. These intervals describe uncertainty around the odds-ratio estimates rather than variability among individual participants.

9. Secondary Quality-of-Life Result: MSQ Role Function Restrictive Domain

ComparisonMean difference95% CIP-valueMethod
Galcanezumab 120 mg vs placebo7.745.20 to 10.28<.001Mixed-effects model
Galcanezumab 240 mg vs placebo7.404.83 to 9.97<.001Mixed-effects model

The endpoint was Mean Change From Baseline in the Migraine-Specific Quality of Life Questionnaire (MSQ) Version 2.1 (v2.1) Role Function Restrictive Domain, measured from baseline through Month 4 through Month 6. Both reported treatment comparisons used mixed-effects models and a mean difference in final values.

These estimates describe differences in the change in a quality-of-life scale between randomized treatment groups. They should not be interpreted as percentages or as the proportion of participants achieving a particular clinically meaningful threshold, because the registry analysis reports mean differences rather than responder proportions.

10. Secondary Migraine-Related Outcomes

EndpointComparisonEstimate95% CIP-value
Overall Mean Change From Baseline in Monthly Migraine Headache Days Requiring Medication for Acute Treatment of Migraine or Headache Galcanezumab 120 mg vs placebo -1.81 -2.28 to -1.33 <.001
Overall Mean Change From Baseline in Monthly Migraine Headache Days Requiring Medication for Acute Treatment of Migraine or Headache Galcanezumab 240 mg vs placebo -1.61 -2.09 to -1.14 <.001
Mean Change From Baseline in Patient Global Impression of Severity (PGI-S) Rating Galcanezumab 120 mg vs placebo -0.32 -0.52 to -0.12 <.001
Mean Change From Baseline in Patient Global Impression of Severity (PGI-S) Rating Galcanezumab 240 mg vs placebo -0.28 -0.48 to -0.07 <.001
Overall Mean Change From Baseline in Headache Hours Galcanezumab 120 mg vs placebo -13.98 -18.99 to -8.97 <.001
Overall Mean Change From Baseline in Headache Hours Galcanezumab 240 mg vs placebo -13.64 -18.68 to -8.60 <.001
Mean Change From Baseline on MIDAS Total Score Galcanezumab 120 mg vs placebo -6.29 -9.45 to -3.13 <.001
Mean Change From Baseline on MIDAS Total Score Galcanezumab 240 mg vs placebo -5.19 -8.39 to -1.98 <.001
Registry-method inconsistency: For the galcanezumab 240 mg comparison on monthly migraine headache days requiring medication for acute treatment, the ClinicalTrials.gov record identifies the endpoint as continuous but also label the endpoint type as binary and the effect measure as an odds ratio, while reporting an estimate of -1.61 with a confidence interval of -2.09 to -1.14. Because that numerical pattern is consistent with a mean difference rather than an odds ratio, this page preserves the registry-reported estimate and does not relabel or recompute it.

11. Why the Longitudinal Analysis Matters

The primary endpoint was not simply a single post-treatment migraine-day measurement. It was defined as an overall mean change from baseline across baseline and Month 1 through Month 6. This creates a repeated-measures statistical problem: each participant can contribute multiple observations, and those observations are naturally related because they come from the same person.

More information per participant

A longitudinal model can use repeated measurements rather than discarding all but one observation per participant.

Within-person correlation

Repeated measurements from the same participant are generally not statistically independent. Mixed-effects models are designed to accommodate that structure.

Change from baseline

The treatment effect is framed around change rather than simply comparing an isolated follow-up value.

Missing observations

Longitudinal models can be useful when follow-up measurements are incomplete, although the validity of the resulting inference depends on the missing-data assumptions and model specification.

12. Statistical Methods Explained

Why was a mixed-effects model used?

The primary outcome was collected repeatedly from baseline through Month 6. A mixed-effects model is appropriate for this type of longitudinal structure because it can represent treatment, time, and repeated observations within participants while accounting for the correlation among observations from the same individual. The exact model specification is not reported in the ClinicalTrials.gov record, so the educational explanation should not be mistaken for a reconstruction of the full statistical analysis plan.

What does a mean difference of -1.92 mean?

For the 120 mg comparison, the reported mean difference was -1.92 days. Because the endpoint is change from baseline in monthly migraine headache days, the negative sign indicates that the modeled change was more negative in the galcanezumab group than in placebo by 1.92 days. It is a difference between group-level estimated means, not a statement that every participant had exactly the same change.

What does the 95% confidence interval tell us?

The interval from -2.48 to -1.37 quantifies uncertainty around the 120 mg mean-difference estimate under the reported analysis framework. A narrower interval generally indicates greater statistical precision than a wider interval, but the interval is not a range containing 95% of individual treatment effects.

What does an odds ratio of 2.63 mean?

For the ≥50% responder threshold, an odds ratio of 2.63 means the estimated odds of meeting that responder criterion were 2.63 times the odds under placebo for the reported 120 mg comparison. Odds are not probabilities, so the estimate cannot be read as "2.63 times as many participants responded."

Why use Fisher exact testing for anti-drug antibodies?

Anti-drug antibody outcomes are binary, and the relevant positive-event counts can be relatively sparse. Fisher exact testing calculates an exact probability for a two-group categorical comparison rather than relying on a large-sample approximation. The registry specifically reports Fisher exact testing for these analyses.

Why doesn't P < .001 tell us the size of the treatment effect?

A P-value describes the statistical evidence against a null hypothesis under the specified testing framework. It does not measure clinical magnitude. In EVOLVE-1, the magnitude is better described by the mean difference, odds ratio, and corresponding confidence interval, while the P-value addresses a different question about statistical evidence.

Why should the two primary dose comparisons be considered separately?

The ClinicalTrials.gov record contains two primary analyses: galcanezumab 120 mg versus placebo and galcanezumab 240 mg versus placebo. Each has its own estimate and confidence interval. Because the ClinicalTrials.gov record does not specify an alpha-allocation or multiplicity procedure for these two comparisons, the page does not claim that the reported P-values have been adjusted for testing both doses.

13. Anti-Drug Antibody Analyses

The registry reports the percentage of participants developing anti-drug antibodies (ADA) to galcanezumab from Month 1 through Month 6. The analysis population required at least one non-missing ADA test result during both the baseline and post-baseline periods.

ADA endpointComparisonMethodP-value
TE ADA PositiveGalcanezumab 120 mg vs placeboFisher exact test<.160
TE ADA PositiveGalcanezumab 240 mg vs placeboFisher exact test.020
Neutralizing AntibodiesGalcanezumab 120 mg vs placeboFisher exact test.131
Neutralizing AntibodiesGalcanezumab 240 mg vs placeboFisher exact test.009
How to interpret these analyses

These are categorical comparisons rather than estimates of treatment effect expressed as mean differences. The registry supplies P-values but no corresponding effect estimates or confidence intervals for these ADA analyses.

A P-value such as .020 should not be converted into an estimated probability that the treatment caused an antibody response. The test evaluates the observed categorical data under its specified null framework.

Because the ClinicalTrials.gov record contains multiple ADA analyses, the individual nominal P-values should not automatically be treated as independently confirmatory without knowing the prespecified multiplicity strategy.

14. Safety Results

The ClinicalTrials.gov record reports serious adverse events separately for the treatment and post-treatment phases. These counts are presented as affected participants divided by participants at risk.

ArmTreatment phasePost-treatment phase
Placebo5/4322/372
Galcanezumab 120 mg6/2064/183
Galcanezumab 240 mg0/2204/185
Serious adverse events during the treatment phase: affected / at risk
Placebo
5/432
Galcanezumab 120 mg
6/206
Galcanezumab 240 mg
0/220

The reported safety data are counts of participants with serious adverse events and participants at risk. They should not be confused with the primary efficacy analysis population or with the total enrollment of 862. The ClinicalTrials.gov record does not provide a formal comparative statistical analysis of serious adverse events, so this page does not calculate one.

Why no new P-value is calculated: The task is to report the registry data rather than reconstruct an unreported safety analysis. Calculating a new comparison from the displayed serious-adverse-event counts would create a statistic that is not part of the registry analysis.

15. Confidence Intervals and Statistical Precision

Confidence intervals appear throughout the posted efficacy analyses. They are particularly useful here because the point estimate alone does not communicate how precisely the treatment effect has been estimated.

OutcomeComparisonEstimate95% CI
Monthly migraine headache days120 mg vs placebo-1.92-2.48 to -1.37
Monthly migraine headache days240 mg vs placebo-1.76-2.31 to -1.20
MSQ Role Function Restrictive120 mg vs placebo7.745.20 to 10.28
MSQ Role Function Restrictive240 mg vs placebo7.404.83 to 9.97
Headache hours120 mg vs placebo-13.98-18.99 to -8.97
Headache hours240 mg vs placebo-13.64-18.68 to -8.60
MIDAS total score120 mg vs placebo-6.29-9.45 to -3.13
MIDAS total score240 mg vs placebo-5.19-8.39 to -1.98

The intervals for the continuous efficacy outcomes all exclude zero in the analyses posted on ClinicalTrials.gov. That is consistent with the reported superiority P-values. However, a confidence interval crossing or not crossing zero is not itself a measure of clinical importance. Clinical interpretation requires understanding the endpoint's scale and the magnitude of the observed difference.

16. Missing Data and Analysis Population

The primary efficacy population was defined as all randomized participants who received at least one dose of study drug and had baseline and at least one post-baseline value. This criterion is important because it determines who contributes to the reported mixed-effects analysis.

Randomization

Randomization establishes the treatment groups before post-randomization outcomes are observed and is central to the causal comparison.

At least one dose

The registry analysis population excludes randomized participants who did not receive at least one dose.

Baseline measurement

A qualifying baseline value is required for the primary analysis population.

Post-baseline measurement

At least one post-baseline value is required. The ClinicalTrials.gov record does not specify the complete missing-data assumptions or imputation rules used by the model.

This distinction illustrates why "enrollment" and "analysis population" are not interchangeable. The registry reports 862 enrolled participants, while the primary analysis population has additional eligibility requirements based on treatment exposure and available outcome measurements.

17. What the Primary Effect Estimates Do — and Do Not — Mean

Mean difference

The primary estimate of -1.92 for 120 mg is a modeled difference in mean change from baseline in monthly migraine headache days between the galcanezumab and placebo groups.

It does not mean that every participant had exactly 1.92 fewer migraine headache days. It also does not describe the probability that an individual participant will respond.

Odds ratio

The responder odds ratio of 2.63 for the ≥50% threshold represents a ratio of odds, not a ratio of probabilities. It therefore should not be described as saying that 63% more participants responded.

P-value

The reported P < .001 for the two primary comparisons indicates strong statistical evidence against the relevant null hypothesis under the reported testing framework. It does not quantify effect size, probability of replication, or probability that the null hypothesis is true.

Confidence interval

A 95% confidence interval describes uncertainty in the estimated treatment effect under the statistical model and sampling framework. It is not a prediction interval for individual participants.

18. Multiplicity and Multiple Comparisons

EVOLVE-1 contains multiple statistical comparisons and multiple secondary outcomes. The ClinicalTrials.gov record classifies the efficacy analyses as superiority and report nominal P-values, but they do not provide a multiplicity-adjustment strategy or an alpha hierarchy.

Analysis familyReported informationInterpretive implication
Primary endpointTwo dose comparisons against placeboTwo primary treatment comparisons are reported.
Responder endpoint≥50%, ≥75%, and 100% thresholds for two dosesMultiple secondary comparisons are present.
Quality-of-life endpointTwo dose comparisonsAdditional secondary testing is present.
Other efficacy outcomesHeadache days requiring acute medication, PGI-S, headache hours, MIDASMultiple additional treatment comparisons are reported.
ADA outcomesTE ADA Positive and neutralizing antibodies for two dosesMultiple categorical tests are reported.

The appropriate interpretation of a P-value depends partly on whether it was part of a prespecified multiplicity-controlled testing hierarchy. Because that information is absent from the ClinicalTrials.gov record, the page reports the posted P-values without assigning them a stronger confirmatory status than the registry supports.

19. Limitations

20. Why This Trial Matters Statistically

EVOLVE-1 is a useful statistical teaching case because it combines randomized treatment allocation with a longitudinal continuous endpoint, responder definitions, quality-of-life measurements, categorical safety-related analyses, and multiple comparisons.

ConceptHow it appears in EVOLVE-1
RandomizationThe trial uses randomized allocation to parallel treatment groups.
Double maskingThe registry classifies the study as double masked.
Repeated measurementsThe primary endpoint spans baseline through Month 6.
Mixed-effects modelUsed for the primary monthly migraine headache-day endpoint and several secondary continuous outcomes.
Mean differenceUsed to quantify differences in modeled continuous outcomes.
Odds ratioUsed for reported responder thresholds of ≥50%, ≥75%, and 100% reduction.
Fisher exact testUsed for anti-drug antibody categorical comparisons.
Confidence intervalsReported for primary and several secondary effect estimates.
P-valuesReported for primary, continuous secondary, and ADA analyses.
MultiplicityMultiple doses, responder thresholds, quality-of-life outcomes, and other secondary endpoints create a multiple-testing problem.

21. Results Summary

Outcome family120 mg vs placebo240 mg vs placebo
Primary monthly migraine headache days-1.92 (95% CI -2.48 to -1.37), P < .001-1.76 (95% CI -2.31 to -1.20), P < .001
≥50% responder odds ratio2.63 (95% CI 2.05 to 3.37)2.48 (95% CI 1.94 to 3.18)
≥75% responder odds ratio2.65 (95% CI 2.04 to 3.45)2.62 (95% CI 2.01 to 3.41)
100% responder odds ratio2.80 (95% CI 1.96 to 4.01)2.61 (95% CI 1.81 to 3.75)
MSQ Role Function Restrictive7.74 (95% CI 5.20 to 10.28), P < .0017.40 (95% CI 4.83 to 9.97), P < .001
Headache hours-13.98 (95% CI -18.99 to -8.97), P < .001-13.64 (95% CI -18.68 to -8.60), P < .001
MIDAS total score-6.29 (95% CI -9.45 to -3.13), P < .001-5.19 (95% CI -8.39 to -1.98), P < .001

The statistical pattern is internally coherent across several continuous outcomes: the reported mean differences are accompanied by two-sided 95% confidence intervals and P-values below .001. The responder analyses use odds ratios instead, illustrating why the choice of effect measure should follow the scale and structure of the endpoint.

22. Related Tutorials

Learn more about the methods used in this trial:

23. Related Calculators

24. Sources

Continue through the Clinical Biostats statistical pathway

Explore the underlying concepts used to understand randomized trials, longitudinal outcomes, confidence intervals, categorical tests, odds ratios, and statistical evidence.

25. Record Summary

EVOLVE-1 provides a compact example of how a randomized clinical trial can generate several distinct statistical questions from the same underlying treatment comparison. Its primary endpoint is longitudinal and analyzed with a mixed-effects model, producing mean differences and two-sided 95% confidence intervals. Secondary responder outcomes use odds ratios, while anti-drug antibody outcomes use Fisher exact testing. The result is a useful illustration of why endpoint type determines the appropriate effect measure and statistical method.

The most important interpretive distinction is between effect size, precision, and statistical evidence. The mean differences and odds ratios describe estimated treatment effects; the confidence intervals describe uncertainty around those estimates; and the P-values address evidence under the relevant null hypotheses. None of these quantities alone describes the experience of every individual participant.

Clinical Biostats methodology: A trial-results page should not merely reproduce a collection of P-values. The objective is to explain how randomization, endpoint definition, analysis population, longitudinal structure, effect measure, confidence interval, and hypothesis testing fit together into a coherent statistical interpretation.