← Clinical Trials
COVID-19 Prevention Phase 3 Completed NCT04939428

MOVe-AHEAD: Complete Statistical Analysis of Molnupiravir in Coronavirus Disease (COVID-19)

An independent statistical analysis of the randomized phase 3 MOVe-AHEAD trial evaluating molnupiravir versus placebo for prevention of Coronavirus Disease (COVID-19) in adults.

MOVe-AHEAD  ·  MK-4482-013  ·  ClinicalTrials.gov NCT04939428
Scope of this record

This page separates reported trial results from statistical interpretation. Numerical results and trial characteristics are taken from the ClinicalTrials.gov record. The registry provides the official trial record.

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

MOVe-AHEAD was a randomized, parallel, triple-masked phase 3 prevention trial comparing molnupiravir with placebo in adults in the context of Coronavirus Disease (COVID-19). The trial enrolled 2441 participants and reported seven statistical analyses, including analyses of three registered primary endpoints.

2441
Enrollment
Randomized trial
2
Arms
Molnupiravir vs placebo
3
Primary endpoints
All binary
7
Analyses posted
3 primary · 4 secondary
FeatureMOVe-AHEAD
PhasePhase 3
ConditionCoronavirus Disease (COVID-19)
Primary purposePrevention
DesignRandomized, parallel, triple-masked
AllocationRandomized
Enrollment2441
InterventionsMolnupiravir and placebo
StatusCompleted
Start2021-08-11
Primary completion2022-11-16
Lead sponsorMerck Sharp & Dohme LLC
Sponsor typeIndustry
ClinicalTrials.govNCT04939428

2. Clinical Question

The statistical question is whether participants assigned to molnupiravir differed from participants assigned to placebo with respect to the prespecified binary outcomes, particularly the development of laboratory-confirmed symptomatic COVID-19 through Day 14 among participants with undetectable SARS-CoV-2 in baseline nasopharyngeal swabs.

Population

Adults participating in a phase 3 prevention trial for Coronavirus Disease (COVID-19).

Intervention

Molnupiravir (MK-4482).

Comparator

Placebo.

Primary question

How does the percentage of participants experiencing each registered binary endpoint differ between molnupiravir and placebo?

3. Trial Design

01
Randomize2441 enrolled
02
MaskTriple-masked design
03
InterveneMolnupiravir or placebo
04
AssessBinary endpoints
05
CompareStratified score-based analysis
ARM A

Molnupiravir

  • Intervention: molnupiravir.
  • Statistical comparisons were made against placebo.
  • Safety analyses used the APaT population.
ARM B

Placebo

  • Comparator: placebo.
  • Statistical comparisons were made against molnupiravir.
  • Safety analyses used the APaT population.

The registry describes the allocation as randomized, the design model as parallel, and masking as triple. These features establish the basic structure of the comparison: participants were assigned to one of two parallel intervention groups and the trial incorporated triple masking.

4. Endpoints

The registry lists three primary endpoints, all binary. Their statistical interpretation depends heavily on the precise event definition and time frame.

Primary endpointTime frameTypeAnalysis
Percentage of Participants Who Had Undetectable SARS-CoV-2 in Baseline Nasopharyngeal (NP) Swabs and Developed COVID-19 (Laboratory-Confirmed SARS-CoV-2 Infection With Symptoms) Through Day 14 Day 14 Binary Miettinen & Nurminen method
Percentage of Participants With ≥1 Adverse Event 29 days Binary Miettinen & Nurminen method
Percentage of Participants Discontinuing From Study Therapy Due to AE Up to 5 days Binary Miettinen & Nurminen method

Primary endpoint definitions

The first endpoint measures the percentage of participants who had undetectable SARS-CoV-2 in baseline NP swabs and subsequently developed COVID-19, defined as laboratory-confirmed SARS-CoV-2 infection with symptoms, through Day 14.

The second endpoint measures the percentage of participants with at least one adverse event during the 29-day time frame. The registry defines an adverse event as any untoward medical occurrence in a clinical study participant temporally associated with study intervention, whether or not considered related to the intervention.

The third endpoint measures the percentage of participants discontinuing from study therapy due to an adverse event through up to 5 days.

5. Statistical Analysis Populations

The registry identifies different analysis populations for efficacy and safety. That distinction is important because the population being analyzed determines exactly what the reported percentage difference represents.

PopulationRole described in registry
mITT The efficacy analysis was conducted on the modified intent-to-treat population consisting of all randomized participants who received at least 1 dose of study intervention.
APaT Safety analyses were conducted in the APaT population, consisting of all randomized participants who received at least 1 dose of study intervention.
Per-protocol analysis context The secondary endpoint analysis text states that index participants were not included in outcome-measure analysis per protocol.
Why the population label matters: “Randomized” and “received at least 1 dose” are not interchangeable definitions. The efficacy results are explicitly described using mITT terminology, while the safety results use APaT. A reported percentage difference should therefore always be interpreted together with its analysis population.

6. Statistical Methodology

Miettinen & Nurminen method

The principal statistical method reported for the trial's primary analyses is the Miettinen & Nurminen method. It belongs to the group of score-based confidence-interval methods for proportions, within the broader family of categorical-data methods.

Core quantity
Difference = pMolnupiravir − pPlacebo

For a binary endpoint, the treatment comparison can be expressed as the difference between the percentage experiencing the event in the two groups. The registry reports adjusted differences together with two-sided 95% confidence intervals.

The registry specifically states that adjusted differences and corresponding confidence intervals for the principal efficacy analysis were based on the Miettinen & Nurminen method and stratified by age and household size. The same stratification description is provided for the secondary analyses where adjusted differences were reported.

Why a score-based interval is appropriate here

For a binary outcome, the basic data structure is an event count and a number at risk in each treatment group. A confidence interval for a difference in proportions must account for uncertainty in both proportions. Score-based methods such as Miettinen–Nurminen are designed specifically for this setting and avoid simply applying a large-sample normal approximation to the difference.

The important statistical object is therefore not merely a percentage in one arm. It is the estimated between-group difference together with an interval describing uncertainty around that difference.

Stratification

The reported adjusted differences and confidence intervals were stratified by age and household size. Stratification allows the comparison to account for prespecified grouping factors rather than treating every participant as if the trial had been conducted without those factors.

Interpretive principle
Adjusted difference ≠ unadjusted difference

The registry's reported estimates are adjusted differences from the specified score-based stratified analysis. They should not be reconstructed by simply subtracting two independently rounded percentages unless the underlying event counts and analysis specification are available.

Two-sided confidence intervals

All seven posted analyses use two-sided 95% confidence intervals in the ClinicalTrials.gov record. A two-sided interval allows the treatment difference to plausibly lie on either side of zero, subject to the statistical assumptions underlying the method.

7. Results: Primary Efficacy Endpoint

COVID-19 through Day 14 among participants with undetectable baseline SARS-CoV-2

The primary efficacy analysis compared molnupiravir with placebo in the mITT population. The registry reports an adjusted difference of -2 percentage points, with a two-sided 95% confidence interval from -5 to 0.9 and a reported P-value of 0.0848.

Adjusted difference in percentages

-2

95% CI: -5 to 0.9   ·   P = 0.0848

Molnupiravir vs placebo · Day 14 · mITT

Primary efficacy resultReported value
ComparisonMolnupiravir vs Placebo
Analysis populationmITT
EndpointCOVID-19 through Day 14 among participants with undetectable SARS-CoV-2 in baseline NP swabs
MethodMiettinen & Nurminen method
Adjustment / stratificationAge and household size
Effect estimate-2
95% CI-5 to 0.9
P-value= 0.0848
Hypothesis typeSuperiority
Clinical Biostats interpretation

The estimate of -2 is an adjusted difference in percentages for the molnupiravir group relative to placebo. Its negative sign means the estimated percentage experiencing the endpoint was lower in the molnupiravir group under the reported analysis.

The estimate does not mean that every participant's individual probability was reduced by exactly 2 percentage points, nor does it establish the size of an effect in a different population or time frame. It is specific to this endpoint, its Day 14 assessment, its analysis population, and the reported stratified method.

The 95% CI of -5 to 0.9 describes statistical uncertainty around the estimated between-group difference. Because the interval extends from a negative value through zero to a positive value, the reported interval includes zero.

The P-value of 0.0848 is not a measure of effect size. It describes the compatibility of the observed data with the null hypothesis under the specified statistical framework. It should not be read as an “8.48% chance” that the treatment effect is real, nor as the probability that the treatment is ineffective.

The superiority hypothesis is also distinct from the numerical magnitude of the estimate. The registry reports a superiority hypothesis type, while the confidence interval and P-value provide the statistical evidence associated with that comparison.

8. Results: Primary Safety Endpoint — ≥1 Adverse Event

The first primary safety endpoint was the percentage of participants with at least one adverse event through 29 days. The safety analysis used the APaT population. The registry reports an adjusted difference of -1.4 percentage points, with a two-sided 95% CI of -4.8 to 2.0.

Adjusted difference in percentages

-1.4

95% CI: -4.8 to 2.0

Molnupiravir vs placebo · 29 days · APaT

Primary safety resultReported value
ComparisonMolnupiravir vs Placebo
Analysis populationAPaT
EndpointPercentage of Participants With ≥1 Adverse Event
Time frame29 days
MethodMiettinen & Nurminen method
Effect estimate-1.4
95% CI-4.8 to 2.0
Hypothesis typeOther / not stated
Clinical Biostats interpretation

The estimated difference of -1.4 indicates a lower percentage with at least one adverse event in the molnupiravir group relative to placebo under the reported analysis.

The estimate is not a statement that molnupiravir caused a 1.4-percentage-point reduction in adverse events for every patient. It summarizes a group-level difference in a binary endpoint.

The 95% CI extends from -4.8 to 2.0. Thus, the statistical uncertainty includes both a potentially larger negative difference and a positive difference. The interval is therefore important for understanding the precision of the estimate rather than focusing only on its point estimate.

The registry states that estimated differences and confidence intervals were provided for this endpoint and that the analysis used the Miettinen and Nurminen method. No P-value is in the ClinicalTrials.gov record, so none is reported here.

9. Results: Primary Safety Endpoint — Discontinuation Due to AE

The second primary safety endpoint was the percentage of participants discontinuing from study therapy due to an adverse event through up to 5 days. The registry reports an adjusted difference of 0.3 percentage points and a two-sided 95% CI of -0.4 to 1.0.

Adjusted difference in percentages

0.3

95% CI: -0.4 to 1.0

Molnupiravir vs placebo · Up to 5 days · APaT

Primary safety resultReported value
ComparisonMolnupiravir vs Placebo
Analysis populationAPaT
EndpointPercentage of Participants Discontinuing From Study Therapy Due to AE
Time frameUp to 5 days
MethodMiettinen & Nurminen method
Effect estimate0.3
95% CI-0.4 to 1.0
Hypothesis typeOther / not stated
Clinical Biostats interpretation

The estimate of 0.3 represents an adjusted percentage-point difference for discontinuation due to an adverse event, molnupiravir relative to placebo. The positive sign indicates that the estimated percentage was slightly higher in the molnupiravir group under the reported analysis.

The 95% CI of -0.4 to 1.0 includes zero. This means the uncertainty around the estimate is compatible with a small difference in either direction under the specified statistical model and sampling framework.

The confidence interval is more informative than the point estimate alone because the point estimate of 0.3 is small relative to the range of plausible values represented by the interval. No P-value was in the ClinicalTrials.gov record.

10. Secondary Endpoint Results

The registry also reports four secondary statistical analyses. These analyses use the same general score-based approach, with adjusted differences and two-sided 95% confidence intervals. Several analyses explicitly state that they were stratified by age and household size.

Secondary endpointTime frameEstimate95% CIP-value
Percentage of Participants (Regardless of SARS-CoV-2 in Baseline NP Swabs) Who Developed COVID-19 (Laboratory-Confirmed SARS-CoV-2 Infection With Symptoms) Through Day 14 Up to Day 14 -3.2 -6.3 to -0.1 = 0.0205
Percentage of Participants Who Had Undetectable SARS-CoV-2 in Baseline NP Swabs and Developed COVID-19 Through Day 29 Up to Day 29 -2.2 -5.4 to 1.0 Not reported
Percentage of Participants Who Had Undetectable SARS-CoV-2 in Baseline NP Swabs and Developed Detectable SARS-CoV-2 in NP Swabs on or Before Day 14 Up to Day 14 -3 -6.9 to 0.8 Not reported
Percentage of Participants Who Had Detectable SARS-CoV-2 in Baseline NP Swabs and Developed COVID-19 Through Day 14 Up to Day 14 -10.5 -22.7 to 2.0 Not reported

Secondary endpoint: COVID-19 through Day 14 regardless of baseline SARS-CoV-2 status

For the secondary endpoint assessing participants regardless of SARS-CoV-2 in baseline NP swabs, the reported adjusted difference was -3.2, with a two-sided 95% CI of -6.3 to -0.1 and P = 0.0205. The registry identifies the hypothesis type as superiority.

Adjusted difference

-3.2

95% CI: -6.3 to -0.1   ·   P = 0.0205

Up to Day 14 · Molnupiravir vs placebo

This estimate is distinct from the primary efficacy endpoint because the secondary endpoint does not restrict the analysis to participants with undetectable SARS-CoV-2 in baseline NP swabs.

Other secondary endpoints

For the Day 29 COVID-19 endpoint among participants with undetectable baseline SARS-CoV-2, the reported adjusted difference was -2.2 with a 95% CI of -5.4 to 1.0.

For development of detectable SARS-CoV-2 in NP swabs on or before Day 14 among participants with undetectable baseline SARS-CoV-2, the reported adjusted difference was -3 with a 95% CI of -6.9 to 0.8.

For participants with detectable SARS-CoV-2 in baseline NP swabs who developed COVID-19 through Day 14, the reported mean difference in final values was -10.5 with a 95% CI of -22.7 to 2.0. The registry describes the adjusted differences and confidence intervals as based on the Miettinen & Nurminen method stratified by age and household size.

Multiplicity caution: The ClinicalTrials.gov record identifies the primary efficacy analysis as a superiority hypothesis and reports P = 0.0848, while a secondary endpoint has P = 0.0205. A nominal P-value from a secondary endpoint should not automatically be interpreted as establishing a separately confirmatory finding unless the prespecified multiplicity strategy supports that interpretation. The ClinicalTrials.gov record does not provide a detailed multiplicity-adjustment scheme, so no such adjustment is inferred here.

11. Safety Results

The registry supplies serious adverse-event counts by arm. These data provide a direct event-count perspective that complements the primary safety endpoint analyses.

Safety measureMolnupiravirPlacebo
Serious adverse events 3 / 763 2 / 765
Serious adverse events: affected / at risk
Molnupiravir
3 / 763
Placebo
2 / 765

These counts should be kept separate from the primary endpoint “Percentage of Participants With ≥1 Adverse Event,” because a serious adverse event and an adverse event of any severity are not the same endpoint. The ClinicalTrials.gov record does not provide a formal between-arm statistical analysis for the serious adverse-event counts, so this page does not construct one.

12. Why the Difference in Proportions Matters

For this trial, the main reported effect measure is a difference in percentages rather than a hazard ratio or an odds ratio. That changes how the result should be read.

Absolute scale

A difference of -2 means the estimated percentage in the molnupiravir group was 2 percentage points lower than the placebo group under the reported analysis.

Not a relative risk

A percentage-point difference does not tell us the ratio of risks between groups. A risk ratio would answer a different question.

Not a probability of benefit

The confidence interval and P-value describe statistical uncertainty; neither gives the probability that an individual participant will benefit.

Endpoint-specific

The estimate applies only to the defined binary endpoint and its specified time frame and analysis population.

Simple interpretation
−2 percentage points = 2 percentage-point lower estimated event percentage

This is an absolute difference. It should not be translated into a relative percentage reduction without the underlying arm-specific event percentages.

13. Confidence Intervals: Reading the Reported Results

The confidence intervals in MOVe-AHEAD are central to interpretation because the trial reports adjusted differences rather than only point estimates.

EndpointEstimate95% CIDoes CI include 0?
Primary Day 14 efficacy endpoint-2-5 to 0.9Yes
Primary ≥1 AE endpoint-1.4-4.8 to 2.0Yes
Primary discontinuation due to AE0.3-0.4 to 1.0Yes
Secondary Day 14 COVID-19 endpoint-3.2-6.3 to -0.1No
Secondary Day 29 COVID-19 endpoint-2.2-5.4 to 1.0Yes
Secondary detectable SARS-CoV-2 endpoint-3-6.9 to 0.8Yes
Secondary endpoint in participants with detectable baseline SARS-CoV-2-10.5-22.7 to 2.0Yes

For a difference in proportions, zero is the natural null value. An interval that contains zero therefore includes no difference as one statistically compatible value under the stated framework. An interval entirely below zero indicates that the corresponding range of estimated differences is negative.

A confidence interval is not an individual-patient prediction

The 95% confidence interval describes uncertainty around the estimated treatment difference. It does not mean that 95% of patients experienced an effect within the interval, and it does not provide a prediction interval for individual outcomes.

14. Statistical Methods Explained

Why was the Miettinen & Nurminen method used?

The registered endpoints are binary percentages. The Miettinen & Nurminen method is a score-based approach for confidence intervals for differences in proportions. It is designed for exactly this type of categorical endpoint and provides an interval for the between-group difference rather than relying solely on an unadjusted normal approximation.

What does an estimate of -2 mean?

It means the adjusted estimated percentage in the molnupiravir group was 2 percentage points lower than the corresponding percentage in the placebo group for the specified endpoint. It does not mean a 2% relative reduction, because percentage points and relative percentages are different quantities.

Why does the confidence interval include zero for the primary efficacy endpoint?

The reported 95% CI is -5 to 0.9. Because zero lies inside that interval, the reported uncertainty includes the possibility of no between-group difference. The P-value of 0.0848 is consistent with that feature of the interval under the reported two-sided analysis.

Why was stratification by age and household size used?

The registry states that adjusted differences and their corresponding confidence intervals were based on the Miettinen & Nurminen method stratified by age and household size. Stratification allows those factors to be incorporated into the treatment comparison rather than ignoring them in the statistical analysis.

Why are mITT and APaT different?

Both populations are described as including randomized participants who received at least 1 dose, but the registry uses the term mITT for the efficacy analysis and APaT for safety. The distinction reflects the different analytic purposes: efficacy is framed around the treatment comparison for the clinical endpoint, whereas safety is framed around participants exposed to study intervention.

Why is the P-value not an effect size?

A P-value such as 0.0848 describes evidence against a null hypothesis under the specified statistical model. It does not quantify how large the treatment difference is. The effect estimate and its confidence interval are needed to describe magnitude and precision.

Why should secondary P-values be interpreted cautiously?

The registry reports seven analyses and multiple endpoints. When several hypotheses are examined, the chance of obtaining a small nominal P-value by chance can increase unless a prespecified multiplicity strategy controls the relevant error rate. The ClinicalTrials.gov record does not provide enough detail to reconstruct such a strategy, so secondary P-values should be reported as registry results rather than automatically treated as independent confirmatory findings.

15. Randomization and Blinding

MOVe-AHEAD is described as randomized, with a parallel design and triple masking. These design features are statistically important because they support a comparison in which treatment assignment is determined by the trial design rather than by participant or investigator choice.

Allocation
Randomized
Participants were assigned within a randomized two-arm trial.
Design model
Parallel
The two intervention groups form parallel treatment arms.
Masking
Triple
The registry identifies the trial as triple-masked.
Purpose
Prevention
The registry identifies prevention as the primary purpose.

Randomization is particularly valuable for causal interpretation because, in expectation, treatment assignment is independent of baseline characteristics under the randomization mechanism. Triple masking can additionally reduce opportunities for treatment assignment or knowledge of treatment to influence participant, investigator, or other trial decisions, although the ClinicalTrials.gov record does not specify the individual roles included in the masking designation.

16. Missing Data, Censoring, and Analysis Assumptions

The endpoints in the registry-reported MOVe-AHEAD data are binary rather than time-to-event outcomes. Consequently, the central statistical issue is the classification of participants into event and non-event categories within the specified assessment window and analysis population.

No hazard-ratio assumption

The reported analyses use differences in percentages, not Cox proportional-hazards models. A proportional-hazards assumption is therefore not part of the reported primary method.

No Kaplan-Meier analysis

The statistical analyses posted on ClinicalTrials.gov do not report Kaplan-Meier estimation. The endpoints are analyzed as binary outcomes at defined time frames.

Missing-data details

The ClinicalTrials.gov record does not specify an imputation method for missing binary endpoint observations.

Event definition matters

The analysis depends on the registry's precise laboratory-confirmed infection, symptom, baseline-swab, adverse-event, and discontinuation definitions.

Because the ClinicalTrials.gov record does not provide a detailed missing-data or imputation strategy, this page does not infer one. That limitation is important: the numerical estimate can only be interpreted within the population and endpoint definitions reported by the registry.

17. Multiplicity and Hypothesis Testing

The trial has three registered primary endpoints and four secondary statistical analyses in the ClinicalTrials.gov record. The primary efficacy endpoint is explicitly associated with a superiority hypothesis. The two primary safety endpoints are listed with “Other / not stated” hypothesis types.

Analysis familyNumber reportedInterpretive role
Primary endpoints3Registered primary outcomes
Secondary endpoints4Additional reported outcomes
Total statistical analyses7All analyses reported in the ClinicalTrials.gov record
Primary analyses with estimate + CI3All three primary analyses have reported estimates and 95% CIs

The presence of multiple endpoints makes it important to distinguish an endpoint's numerical result from the broader question of confirmatory error control. A P-value should not be interpreted in isolation from the prespecified testing hierarchy or multiplicity procedure. The ClinicalTrials.gov record does not provide a detailed multiplicity adjustment scheme, so this page does not assume one.

18. What the Primary Efficacy Result Does — and Does Not — Mean

Effect estimate

The adjusted difference of -2 means that the estimated percentage experiencing the specified Day 14 COVID-19 endpoint was 2 percentage points lower with molnupiravir than with placebo under the reported mITT analysis.

It does not mean that each participant had exactly a 2-percentage-point change in probability, nor does it describe an individual's treatment response.

Precision

The 95% CI of -5 to 0.9 shows that the estimate has uncertainty extending from a larger negative difference to a small positive difference. Zero is included in the interval.

P-value

The reported P-value of 0.0848 quantifies evidence relative to the specified null hypothesis under the reported statistical framework. It is not the probability that the null hypothesis is true and it is not a measure of the clinical importance of the treatment difference.

Absolute versus relative effect

The registry's reported effect measure is an adjusted difference in percentages. A relative risk, odds ratio, or relative percentage reduction cannot be recovered from the registry-reported adjusted difference alone.

19. Primary Results in Context

EndpointEstimate95% CIKey statistical observation
Day 14 primary efficacy endpoint -2 -5 to 0.9 CI includes zero; P = 0.0848
≥1 adverse event through 29 days -1.4 -4.8 to 2.0 CI includes zero
Discontinuation due to AE through up to 5 days 0.3 -0.4 to 1.0 CI includes zero

The three primary analyses should be interpreted together rather than collapsing them into a single numerical conclusion. They address different clinical and safety questions and use different endpoint definitions and time frames.

The primary efficacy endpoint is specifically about laboratory-confirmed symptomatic COVID-19 through Day 14 among participants with undetectable SARS-CoV-2 at baseline. The safety endpoints concern adverse events over 29 days and discontinuation due to adverse events through up to 5 days. These outcomes are not interchangeable.

20. Limitations and Interpretation Issues

21. Why This Trial Matters Statistically

MOVe-AHEAD is a useful teaching case because it demonstrates how a randomized prevention trial can rely on a relatively simple binary endpoint while still requiring careful statistical interpretation. The key analytical issue is not merely calculating two percentages; it is defining the population, endpoint, time frame, stratification variables, effect measure, and uncertainty interval consistently.

ConceptHow it appears in MOVe-AHEAD
RandomizationThe trial is described as randomized.
Parallel designTwo intervention arms are compared in a parallel design.
Triple maskingThe registry identifies the masking as triple.
Binary endpointsAll three registered primary endpoints are binary.
Difference in proportionsThe reported effect estimates are adjusted differences in percentages.
Score-based CIThe Miettinen & Nurminen method is the reported primary analytical method.
Stratified analysisAdjusted differences and CIs were based on stratification by age and household size.
mITT analysisThe primary efficacy analysis used the registry-defined mITT population.
APaT safety analysisPrimary safety analyses used the APaT population.
Confidence intervalsAll seven registry-reported analyses report two-sided 95% confidence intervals.
P-valuesP = 0.0848 is reported for the primary efficacy endpoint and P = 0.0205 for one secondary endpoint.
MultiplicityThree primary and four secondary analyses are reported, requiring careful interpretation of nominal P-values.

22. A Practical Reading of the MOVe-AHEAD Statistical Record

A disciplined reading starts with the endpoint rather than the P-value. The first question is: What exactly was counted? For the principal efficacy endpoint, the answer is symptomatic, laboratory-confirmed COVID-19 through Day 14 among participants whose baseline NP swabs had undetectable SARS-CoV-2.

The second question is: Who was analyzed? The efficacy analysis was conducted in the mITT population. The safety analyses were conducted in APaT.

The third question is: What was the effect measure? The analyses posted on ClinicalTrials.gov report adjusted differences in percentages. This places the result on an absolute percentage-point scale.

The fourth question is: How uncertain is the estimate? The answer comes from the two-sided 95% confidence interval. For the primary efficacy endpoint, the interval is -5 to 0.9.

Only after these questions should the P-value be considered. P = 0.0848 is one component of the statistical evidence, not a replacement for the estimated difference and its confidence interval.

A useful hierarchy for interpretation
Endpoint → Analysis population → Effect measure → Confidence interval → P-value → Multiplicity context

This sequence helps prevent a common statistical error: treating a single P-value as if it contains all the information needed to understand a clinical-trial result.

23. Related Tutorials

Learn more about the methods used in this trial:

24. Related Statistical Calculators

25. Sources

Continue through Clinical Biostats

Explore the statistical concepts behind randomized trials, binary endpoints, confidence intervals, and stratified analyses.

26. Record Summary

MOVe-AHEAD provides a useful example of statistical analysis for a randomized prevention trial with binary clinical and safety endpoints. The three registered primary endpoints were analyzed using the Miettinen & Nurminen method, with adjusted differences and two-sided 95% confidence intervals. The primary efficacy estimate was -2 with a 95% CI of -5 to 0.9 and P = 0.0848. The two primary safety estimates were -1.4 with a 95% CI of -4.8 to 2.0 and 0.3 with a 95% CI of -0.4 to 1.0.

The secondary analyses demonstrate why endpoint definitions and statistical context matter. The secondary Day 14 COVID-19 endpoint reported an adjusted difference of -3.2 with a 95% CI of -6.3 to -0.1 and P = 0.0205, while the other secondary confidence intervals reported in the registry extend across zero. These results should be interpreted in the context of multiple endpoints, analysis populations, stratification, and the distinction between nominal statistical evidence and a prespecified confirmatory testing strategy.

Clinical Biostats methodology: A trial-results page should not merely reproduce numerical outputs. The goal is to identify the estimand represented by each result, explain the statistical method used to obtain it, show the uncertainty around the estimate, and distinguish reported evidence from educational interpretation.