← Clinical Trials
Head and Neck Cancer Phase 3 Completed NCT02105636

CheckMate-141: Complete Statistical Analysis of Nivolumab in Recurrent or Metastatic Head and Neck Carcinoma

An independent statistical review of the randomized phase 3 CheckMate-141 trial comparing nivolumab with investigator's choice of cetuximab, methotrexate, or docetaxel in squamous cell carcinoma of the head and neck.

Randomized phase 3  ·  Enrollment 361  ·  Primary endpoint: Overall Survival
Scope of this record

This page separates reported trial results from statistical interpretation. Numerical results are restricted to the information provided in the ClinicalTrials.gov record. Where the registry does not report a formal statistical method or additional result, no method or estimate is inferred as a reported trial result.

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

CheckMate-141 was a randomized, parallel-group, open-label phase 3 trial evaluating nivolumab versus investigator's choice therapy in recurrent or metastatic squamous cell carcinoma of the head and neck. The registry reports 361 enrolled participants, two treatment arms, and overall survival as the registered primary endpoint.

361
Enrolled
Phase 3
2
Arms
Parallel design
0.70
Primary OS HR
95% CI 0.53–0.92
0.0101
OS P-value
Two-sided
FeatureCheckMate-141
Trial nameCheckMate-141
Brief titleTrial of Nivolumab vs Therapy of Investigator's Choice in Recurrent or Metastatic Head and Neck Carcinoma
PhasePhase 3
ConditionSquamous Cell Carcinoma of the Head and Neck
AllocationRandomized
Design modelParallel
MaskingNone
Primary purposeTreatment
Enrollment361
Arms2
Primary endpointOverall Survival (OS)
Primary endpoint typeTime-to-event
Hypothesis typeSuperiority
Trial statusCompleted
Start2014-05-29
Primary completion2015-11-06
Lead sponsorBristol-Myers Squibb

2. Clinical Question

The central question was whether nivolumab improved overall survival compared with therapy selected by the investigator among participants with recurrent or metastatic squamous cell carcinoma of the head and neck.

Population

Participants with squamous cell carcinoma of the head and neck in the registered CheckMate-141 trial.

Intervention

Nivolumab 3mg/kg.

Comparator

Investigator's choice of cetuximab, methotrexate, or docetaxel.

Primary question

Does nivolumab improve overall survival relative to investigator's choice therapy?

3. Trial Design

01
Randomize361 enrolled
02
Parallel arms2 treatment groups
03
TreatmentNivolumab or investigator's choice
04
Follow-upTime-to-event assessment
05
AnalysisOverall survival
ARM A

Nivolumab

  • Nivolumab 3mg/kg
  • Randomized treatment group
  • Primary comparison based on overall survival
ARM B

Investigator's Choice

  • Cetuximab
  • Methotrexate
  • Docetaxel
Allocation
Randomized allocation was used to compare nivolumab with investigator's choice therapy.
Masking
The registry describes the trial as having no masking.
Design model
The trial used a parallel design with two arms.
Primary purpose
The registered primary purpose was treatment.

4. Endpoints

EndpointRegistry definitionType
Overall Survival (OS) OS was defined as the time from randomization to the date of death from any cause. Participants were censored at the date they were last known to be alive and at the date of randomization if they were randomized but had no follow-up. Median OS time was calculated using Kaplan-Meier (KM) method. Time-to-event
Investigator-Assessed Progression-Free Survival (PFS) From date of randomization to date of disease progression or death, whichever occurs first (Up to approximately 87 months) Time-to-event
Investigator-Assessed Objective Response Rate (ORR) From date of randomization to date of disease progression or study drug is discontinued, whichever occurs first (Up to approximately 87 months) Binary

The registered primary endpoint is overall survival. The posted results also contain investigator-assessed PFS and investigator-assessed ORR analyses, making it possible to examine how the statistical treatment of time-to-event and binary outcomes differs within the same randomized trial.

5. Statistical Methodology

Kaplan-Meier estimation

The registry explicitly states that median overall survival was calculated using the Kaplan-Meier method. Kaplan-Meier estimation is designed for time-to-event data with right censoring, allowing participants who have not yet experienced death at their last known follow-up to contribute information up to that point.

Kaplan-Meier survival function
S(t) = ∏ti ≤ t (1 − di/ni)

Here, di represents the number of events at an event time and ni represents the number at risk immediately before that time.

Log-rank test

The posted primary OS analysis used a log-rank method. A log-rank test compares the observed and expected numbers of events between randomized groups over the follow-up period. The method is therefore naturally suited to the trial's primary time-to-event endpoint.

Stratified Cox proportional-hazards model

The reported OS analysis used a stratified Cox proportional-hazards model. The registry states that the hazard ratio was estimated from a stratified Cox proportional hazards model for the distribution of OS in each randomized arm.

Hazard ratio
HR = hazard in nivolumab group / hazard in investigator's choice group

An HR below 1 indicates a lower estimated instantaneous event rate in the nivolumab group relative to investigator's choice under the fitted model. The hazard ratio is not itself an absolute survival probability or a measure of the percentage of participants who benefit.

Cochran-Mantel-Haenszel analysis

The ORR analyses used the Cochran-Mantel-Haenszel approach. For the reported difference in ORR, the registry describes a stratum-adjusted difference in response rates, with nivolumab minus investigator's choice, based on Cochran-Mantel-Haenszel weighting.

For the common odds ratio analysis, the registry states that the comparison was stratified by prior cetuximab, recorded in the IVRS. The resulting odds ratio was a stratum-adjusted estimate using the Mantel-Haenszel method.

Analysis population

The posted analyses identify the analysis population as all randomized participants. This is particularly important for interpreting the primary efficacy analysis because the treatment comparison remains anchored to the randomized assignment rather than only to participants who completed treatment.

6. Primary Result: Overall Survival

The primary endpoint was overall survival, defined from randomization to death from any cause. The posted analysis compared nivolumab 3mg/kg with investigator's choice therapy using a log-rank test, with the hazard ratio and confidence interval estimated using a stratified Cox proportional-hazards model.

Hazard ratio for overall survival

0.70

95% CI: 0.53–0.92   ·   P = 0.0101

Analysis population: all randomized participants  ·  Superiority hypothesis

Primary endpointNivolumab 3mg/kg vs investigator's choice
EndpointOverall Survival (OS)
Time frameFrom date of randomization to date of death (Up to approximately 18 months)
Analysis methodLog Rank
Model for HRStratified Cox proportional hazard model
Hazard ratio0.70
95% CI0.53–0.92
P-value0.0101
Clinical Biostats interpretation

What the estimate means: An OS hazard ratio of 0.70 means that, under the stratified Cox model, the estimated instantaneous rate of death in the nivolumab group was 30% lower than in the investigator's choice group over the analyzed follow-up.

What it does not mean: It does not mean that 30% of participants avoided death, that survival increased by 30%, or that every individual participant experienced exactly a 30% reduction in risk. A hazard ratio is a relative model-based measure of event rates over time.

What the confidence interval says: The 95% CI of 0.53–0.92 describes uncertainty around the estimated hazard ratio under the statistical model and sampling framework. It does not describe the range of effects experienced by individual patients.

Why the p-value is different: The p-value of 0.0101 addresses evidence against the null hypothesis under the specified testing framework. It does not measure the magnitude of the treatment effect or tell us how clinically important an HR of 0.70 is.

Important modeling caution: The estimate comes from a stratified Cox proportional-hazards model. Interpretation of a single Cox hazard ratio is most straightforward when the proportional-hazards assumption is reasonably appropriate over the analyzed period. The ClinicalTrials.gov record does not provide a diagnostic assessment of that assumption.

7. Secondary Result: Investigator-Assessed Progression-Free Survival

Investigator-assessed PFS was defined as the time from randomization to disease progression or death, whichever occurred first. The posted comparison used all randomized participants and reported a hazard ratio, although the ClinicalTrials.gov record does not identify a formal statistical method for this particular analysis.

Hazard ratio for progression or death

0.86

95% CI: 0.68–1.10

Analysis population: all randomized participants

Secondary endpointReported result
EndpointInvestigator-Assessed Progression-Free Survival (PFS)
Time frameFrom date of randomization to date of disease progression or death, whichever occurs first (Up to approximately 87 months)
Analysis populationAll randomized participants
Effect measureHazard Ratio (HR)
Estimate0.86
95% CI0.68–1.10
Formal method reported in the ClinicalTrials.gov recordNot reported
Clinical Biostats interpretation

What the estimate means: The reported HR of 0.86 corresponds to a lower estimated instantaneous rate of progression or death in the nivolumab group under a hazard-ratio interpretation, relative to investigator's choice.

What it does not mean: It does not mean that 14% of participants avoided progression or death, nor does it describe an absolute difference in PFS probability at a particular time.

Precision: The 95% CI of 0.68–1.10 expresses uncertainty around the reported estimate. Because the interval extends above 1, the ClinicalTrials.gov record does not establish a directionally lower hazard at conventional two-sided confidence-interval interpretation.

Methodological limitation: The ClinicalTrials.gov record does not report the formal statistical method for this particular PFS analysis. A time-to-event endpoint such as PFS would commonly be analyzed using Kaplan-Meier estimation, a log-rank comparison, and a Cox model, but those methods should not be presented here as the registry's reported PFS analysis.

8. Secondary Results: Objective Response Rate

Investigator-assessed objective response rate was analyzed as a binary endpoint. The ClinicalTrials.gov record contains two related stratum-adjusted effect measures: a difference in response rates and a common odds ratio.

Stratum-Adjusted Difference in ORR

Difference in objective response rate

7.6

95% CI: 1.5–13.6

Difference defined as nivolumab minus investigator's choice

MeasureReported result
EndpointInvestigator-Assessed Objective Response Rate (ORR)
Effect measureDifference in ORR
Estimate7.6 percentage points
95% CI1.5–13.6
MethodCochran-Mantel-Haenszel test
AdjustmentStratum-adjusted difference in response rates, nivolumab minus investigator's choice
Clinical Biostats interpretation

The reported difference of 7.6 represents the stratum-adjusted difference in response rates, defined as nivolumab minus investigator's choice. The corresponding 95% CI of 1.5–13.6 describes uncertainty around that difference.

This is an absolute difference, not an odds ratio or hazard ratio. It therefore answers a different statistical question from the OS and PFS hazard ratios. The ClinicalTrials.gov record does not provide the separate response-rate percentages for the two groups, so those percentages should not be reconstructed from the reported difference.

Stratified Common Odds Ratio

CMH estimate of common odds ratio

2.49

95% CI: 1.07–5.82

Stratified by prior cetuximab (yes, no)

MeasureReported result
EndpointInvestigator-Assessed Objective Response Rate (ORR)
Effect measureCMH Estimate of Common Odds Ratio
Estimate2.49
95% CI1.07–5.82
MethodCochran-Mantel-Haenszel test
Stratification factorPrior Cetuximab (yes, no), as recorded in the IVRS
Clinical Biostats interpretation

An odds ratio of 2.49 means that the estimated odds of objective response were 2.49 times as high with nivolumab as with investigator's choice after the reported Mantel-Haenszel stratification by prior cetuximab.

The odds ratio is not a risk ratio and should not be read as "2.49 times as many patients responded." Odds depend on both the probability of response and the probability of nonresponse.

The 95% CI of 1.07–5.82 describes uncertainty around the common odds-ratio estimate. The width of this interval also illustrates why an effect estimate should be interpreted together with its confidence interval rather than in isolation.

9. Post-Hoc Extended Overall Survival Analysis

The ClinicalTrials.gov record includes a post-hoc extended collection for overall survival, with follow-up extending to approximately 87 months. This analysis is distinct from the registered primary endpoint analysis and is explicitly classified as post-hoc in the ClinicalTrials.gov record.

Extended overall survival hazard ratio

0.68

95% CI: 0.54–0.85

Stratified Cox proportional-hazards model

FeatureExtended OS analysis
EndpointOverall Survival (OS) - Extended Collection
RolePost-hoc
Time frameFrom date of randomization to date of death (Up to approximately 87 months)
Analysis populationAll randomized participants
Effect measureHazard ratio
Estimate0.68
95% CI0.54–0.85
MethodStratified Cox proportional-hazards model
Clinical Biostats interpretation

The extended OS HR of 0.68 corresponds to an approximately 32% lower estimated instantaneous rate of death in the nivolumab group relative to investigator's choice under the reported stratified Cox model.

The 95% CI of 0.54–0.85 indicates the statistical uncertainty surrounding this post-hoc estimate. It is important to keep this analysis separate from the registered primary analysis: a later or post-hoc estimate should not automatically be treated as though it were the original confirmatory endpoint analysis.

10. Safety

The ClinicalTrials.gov record reports serious adverse events by treatment arm as affected participants divided by participants at risk. These figures provide a direct safety summary without requiring reconstruction of an event table that is not contained in the ClinicalTrials.gov record.

Treatment groupSerious adverse eventsAt risk
Nivolumab 3mg/kg165236
Investigator's Choice87111
Serious adverse events: affected / at risk
Nivolumab 3mg/kg
165 / 236
Investigator's Choice
87 / 111
Safety interpretation: The ClinicalTrials.gov record identifies serious adverse events as affected participants divided by participants at risk. They do not provide a formal between-arm statistical comparison, confidence interval, attribution assessment, exposure-adjusted rate, or detailed adverse-event classification. Those quantities should not be inferred from the affected/at-risk counts alone.

11. Statistical Methods Explained

Why was a log-rank test used for overall survival?

Overall survival is a time-to-event endpoint because both the timing of death and censoring matter. A log-rank test compares the survival experience of randomized groups across follow-up rather than reducing the outcome to a single binary status at one arbitrary time point.

What does an OS hazard ratio of 0.70 mean?

Under the reported stratified Cox model, an HR of 0.70 means the estimated instantaneous rate of death was 30% lower in the nivolumab group than in the investigator's choice group. It does not mean that 30% more participants survived, nor does it translate directly into a fixed difference in survival probability at every time point.

Why is the confidence interval important?

The estimate is only one point on a range of statistically plausible values under the model. For the primary OS result, the 95% CI of 0.53–0.92 shows that the estimated treatment effect is uncertain even though the point estimate is 0.70.

Why doesn't the p-value measure effect size?

The OS p-value of 0.0101 describes the evidence against the null hypothesis under the specified testing framework. It does not tell us whether the estimated effect is large or small. The HR and its confidence interval provide the information about magnitude and precision.

Why was the Cochran-Mantel-Haenszel method useful for ORR?

ORR is binary rather than time-to-event. The Cochran-Mantel-Haenszel framework allows treatment effects to be estimated while accounting for a stratification structure. In the registry-reported analysis, it was used to obtain a stratum-adjusted difference in response rates and a common odds ratio.

What is the difference between an odds ratio and a hazard ratio?

An odds ratio compares the odds of a binary outcome such as response. A hazard ratio compares estimated instantaneous event rates over time. The ORR odds ratio of 2.49 and the OS hazard ratio of 0.70 therefore describe fundamentally different quantities and should not be compared numerically.

Why does randomization matter for interpretation?

Randomization creates the basis for comparing outcomes between treatment assignments while reducing systematic differences caused by measured and unmeasured baseline factors. The registry-reported primary OS analysis used all randomized participants, preserving that randomized comparison for the efficacy endpoint.

12. Understanding the Primary OS Analysis

Relative effect

HR 0.70 summarizes the relative difference in the modeled instantaneous rate of death between randomized groups.

Precision

The 95% CI of 0.53–0.92 shows uncertainty around the HR estimate.

Evidence against the null

The reported two-sided p-value was 0.0101 under the superiority analysis.

Analysis population

The primary analysis population was all randomized participants.

Three quantities, three questions
Effect size → How large is the estimated relative effect?
Confidence interval → How precise is that estimate?
P-value → How compatible are the data with the specified null hypothesis?

Keeping these questions separate prevents a common statistical error: treating a small p-value as if it were a measure of clinical magnitude.

13. Stratified Analysis

Stratification appears explicitly in the registry-reported CheckMate-141 analyses. The primary OS result used a stratified Cox proportional-hazards model, while the ORR common odds-ratio analysis was stratified by prior cetuximab.

AnalysisStratification information in the ClinicalTrials.gov recordMethod
Primary OSStratified analysis stated; specific strata are not identified in the registry-reported analysis textStratified Cox proportional-hazards model
ORR differenceStratum-adjusted response ratesCochran-Mantel-Haenszel
ORR common odds ratioPrior Cetuximab (yes, no)Mantel-Haenszel / Cochran-Mantel-Haenszel
Extended OSStratified analysis stated; specific strata are not identified in the registry-reported analysis textStratified Cox proportional-hazards model

A stratified analysis is different from simply adjusting for a covariate in an ordinary regression model. The analysis preserves a comparison across defined strata while producing an overall treatment-effect estimate. The exact meaning of the strata depends on the trial's prespecified design and analysis specification; where those factors are not provided in the ClinicalTrials.gov record, they should not be reconstructed.

14. Confidence Intervals Across the Trial

EndpointEffect95% CI
Overall SurvivalHR 0.700.53–0.92
Investigator-Assessed PFSHR 0.860.68–1.10
ORR difference7.61.5–13.6
ORR common odds ratio2.491.07–5.82
Extended OSHR 0.680.54–0.85

The table illustrates why confidence intervals should be interpreted on the scale of the effect measure. A confidence interval around a hazard ratio is not directly comparable in meaning to one around an absolute response-rate difference. Each interval quantifies uncertainty around its own underlying parameter.

15. Multiplicity, Interim Analysis, and Other Design Features

The ClinicalTrials.gov record identifies a superiority hypothesis and five posted statistical analyses, including one primary analysis, three secondary analyses, and one post-hoc extended OS analysis. They do not provide a multiplicity adjustment plan, interim-analysis schedule, alpha-spending strategy, or formal familywise-error procedure.

Design topicWhat the ClinicalTrials.gov record supports
HypothesisSuperiority
Primary endpointOverall Survival
Secondary analysesInvestigator-assessed PFS and ORR
Post-hoc analysisExtended OS
Multiplicity adjustmentNot reported in the ClinicalTrials.gov record
Interim analysisNot reported in the ClinicalTrials.gov record
Alpha spendingNot reported in the ClinicalTrials.gov record
Missing-data/imputation strategyNot reported in the ClinicalTrials.gov record
Bayesian methodsNot reported in the ClinicalTrials.gov record
Interpretation boundary: The primary OS analysis has a reported p-value because a formal primary statistical analysis was posted. The ClinicalTrials.gov record does not state whether that p-value was adjusted for any other endpoints or interim looks. Therefore, no additional multiplicity claim should be inferred.

16. Time-to-Event Analysis: Why Censoring Matters

Both OS and PFS are time-to-event endpoints. The registry's OS definition explicitly states that participants were censored at the date they were last known to be alive and at the date of randomization if they were randomized but had no follow-up.

Censoring means that the exact event time is not observed for some participants during the relevant observation period. Kaplan-Meier and Cox methods are designed to use the information available up to censoring rather than treating censored participants as though they had experienced the event.

Observed event

The participant experiences the endpoint during follow-up, providing an observed event time.

Censored observation

The participant contributes follow-up information up to the censoring time without an observed event at that point.

This distinction is central to why survival analysis cannot be replaced by simply comparing the proportion of participants who died by an arbitrary cutoff without considering follow-up time.

17. Primary vs Secondary vs Post-Hoc Evidence

EndpointRoleReported effectStatistical interpretation
Overall SurvivalPrimaryHR 0.70 (95% CI 0.53–0.92), P = 0.0101Formal primary superiority analysis
Investigator-Assessed PFSSecondaryHR 0.86 (95% CI 0.68–1.10)Secondary time-to-event result; formal method not reported in the ClinicalTrials.gov record
Investigator-Assessed ORRSecondaryDifference 7.6 (95% CI 1.5–13.6)Stratum-adjusted binary endpoint analysis
Investigator-Assessed ORRSecondaryOR 2.49 (95% CI 1.07–5.82)Stratified common odds ratio
Extended OSPost-hocHR 0.68 (95% CI 0.54–0.85)Later analysis that should remain distinct from the registered primary analysis

This hierarchy matters because a primary endpoint is defined before examining the eventual results, while secondary and post-hoc analyses can answer additional questions without necessarily carrying the same confirmatory status. The statistical estimate itself does not determine its evidentiary role; the prespecified trial design does.

18. Important Limitations and Interpretation Issues

19. Why This Trial Matters Statistically

CheckMate-141 is a useful statistical teaching case because the ClinicalTrials.gov record place several common clinical-trial methods side by side: randomized treatment comparison, Kaplan-Meier estimation, log-rank testing, stratified Cox modeling, hazard ratios, confidence intervals, Cochran-Mantel-Haenszel analysis, odds ratios, absolute response-rate differences, and a post-hoc extended survival analysis.

ConceptHow it appears in CheckMate-141
RandomizationRandomized allocation with two parallel treatment arms
Time-to-event endpointOverall survival is the registered primary endpoint
Kaplan-Meier estimationThe registry states that median OS was calculated using the KM method
Log-rank testReported method for the primary OS comparison
Hazard ratioPrimary OS HR 0.70; PFS HR 0.86; extended OS HR 0.68
Cox modelStratified Cox proportional-hazards model used for OS hazard-ratio estimation
Confidence intervalReported for OS, PFS, ORR difference, ORR odds ratio, and extended OS
Binary endpointObjective response rate
Cochran-Mantel-HaenszelUsed for stratum-adjusted ORR analyses
Odds ratioCMH common OR 2.49 for ORR
Stratified analysisAppears in OS and ORR analyses
Post-hoc analysisExtended OS analysis classified as post-hoc

20. A Practical Reading of the Reported Evidence

Start with the primary endpoint

The primary OS comparison reports an HR of 0.70, a 95% CI of 0.53–0.92, and a two-sided p-value of 0.0101. The three quantities should be read together rather than treating any one of them as the complete result.

Then examine secondary endpoints

The investigator-assessed PFS analysis reports HR 0.86 with 95% CI 0.68–1.10. ORR is reported using both an absolute difference of 7.6 and a stratified common odds ratio of 2.49. These estimates answer different questions.

Keep later analyses separate

The post-hoc extended OS analysis reports HR 0.68 with 95% CI 0.54–0.85. Its classification as post-hoc is important when interpreting its statistical role relative to the registered primary analysis.

Separate efficacy from safety

The ClinicalTrials.gov record reports serious adverse events of 165/236 for nivolumab 3mg/kg and 87/111 for investigator's choice. These are safety counts, not efficacy endpoints, and the ClinicalTrials.gov record does not provide a formal statistical comparison.

21. Statistical Interpretation vs Clinical Interpretation

Statistical interpretation

The primary OS analysis reports a hazard ratio below 1 with a 95% confidence interval of 0.53–0.92 and a two-sided p-value of 0.0101. The analysis used a log-rank comparison and a stratified Cox model for the hazard ratio.

Clinical interpretation

The ClinicalTrials.gov record establishes the statistical estimates and their uncertainty, but do not by themselves provide a complete clinical-benefit assessment. Clinical interpretation requires considering the endpoint definition, absolute outcomes when available, safety, follow-up, and the intended patient population.

This distinction is important. Statistical evidence describes the strength, magnitude, and precision of an observed randomized comparison. Clinical interpretation asks what that evidence means in the broader context of patient outcomes. The two perspectives overlap, but neither should be substituted for the other.

22. Trial Timeline

2014-05-29

Trial start

The registered trial start date was 2014-05-29.

2015-11-06

Primary completion

The registered primary completion date was 2015-11-06.

Completed

Registry status

The ClinicalTrials.gov record classifies the study as completed and indicate that results were posted.

Extended analysis

Longer OS collection

A post-hoc extended OS analysis was reported with a time frame extending to approximately 87 months.

23. Related Tutorials

Learn more about the methods used in this trial:

24. Related Calculators

25. Sources

Continue through the Clinical Biostats statistical pathway

Connect the endpoints and methods used in CheckMate-141 with deeper tutorials and statistical calculators.

26. Record Summary

CheckMate-141 provides a compact example of several core clinical-trial statistical methods. Its registered primary endpoint was overall survival, analyzed using a log-rank approach with a stratified Cox proportional-hazards model producing an HR of 0.70 (95% CI 0.53–0.92; P = 0.0101). Secondary analyses extended the statistical framework to investigator-assessed PFS and ORR, including both a stratum-adjusted response-rate difference and a Mantel-Haenszel common odds ratio. A separate post-hoc extended OS analysis reported an HR of 0.68 (95% CI 0.54–0.85).

The most important statistical lesson is that these numbers cannot be treated as interchangeable. Hazard ratios describe relative time-to-event effects, odds ratios describe binary-outcome odds, absolute response-rate differences describe differences on a probability scale, confidence intervals quantify uncertainty, and p-values address compatibility with a specified null hypothesis. The analysis population, stratification, endpoint role, and post-hoc status also determine how each result should be interpreted.

Clinical Biostats methodology: A trial-results page should not merely repeat an abstract or list isolated estimates. The goal is to reconstruct the statistical story of the trial while clearly separating registered design features, reported analyses, and educational interpretation.