This page separates reported registry results from statistical interpretation. The ClinicalTrials.gov record identifies the registered primary endpoint and safety results, but do not provide formal statistical analyses for the primary endpoint.
1. Trial at a Glance
ENSEMBLE-2 was a randomized, double-blind, parallel phase 3 prevention trial evaluating Ad26.COV2.S versus placebo for prevention of SARS-CoV-2-mediated COVID-19 in adults, including participants with or without stable co-morbidities associated with progression to severe COVID-19.
| Feature | ENSEMBLE-2 |
|---|---|
| Trial name | ENSEMBLE-2 |
| NCT identifier | NCT04614948 |
| Phase | 3 |
| Status | COMPLETED |
| Therapeutic area | Infectious Disease |
| Population | Participants with or without stable co-morbidities associated with progression to severe COVID-19 |
| Allocation | RANDOMIZED |
| Design model | PARALLEL |
| Masking | DOUBLE |
| Primary purpose | PREVENTION |
| Enrollment | 31835.0 |
| Interventions | Ad26.COV2.S (biological); Placebo (other) |
| Lead sponsor | Janssen Vaccines & Prevention B.V. |
| Sponsor type | INDUSTRY |
| Trial start | 2020-11-12 |
| Primary completion | 2023-06-18 |
2. Clinical Question
The central statistical question was whether vaccination with Ad26.COV2.S, compared with placebo, affected the occurrence of the registered primary endpoint: the first occurrence of molecularly confirmed moderate to severe/critical COVID-19 in seronegative participants with onset at least 14 days after the second vaccination.
Population
Adults with or without stable co-morbidities associated with progression to severe COVID-19.
Intervention
Ad26.COV2.S, classified in the registry as a biological intervention.
Comparator
Placebo, classified in the registry as an other intervention.
Primary question
How does Ad26.COV2.S compare with placebo for the registered binary COVID-19 endpoint during the double-blind phase?
3. Trial Design
Ad26.COV2.S
- Intervention classified as biological
- Randomized allocation
- Double-blind phase
- Evaluated for prevention of the registered COVID-19 endpoint
Placebo
- Intervention classified as other
- Randomized allocation
- Double-blind phase
- Comparator for the registered COVID-19 endpoint
The registry describes the study as randomized, parallel, double-blind, and prevention-focused. Those design features are important statistically because randomization establishes the intended basis for comparing the two treatment assignments, while masking is intended to reduce the influence of treatment knowledge on participant behavior, outcome assessment, and other trial processes.
Planned crossover versus later vaccination sequences
The registry caveat states that there was no planned cross-over. It explains that the various complex vaccination sequences resulted from ethically allowing placebo participants to unblind themselves and receive active vaccination in the open-label phase.
4. Randomization and Blinding
Randomization was a registered design feature, and the study was double masked. Together, these features are central to the validity of the primary comparison.
Participants were allocated through a randomized trial design rather than an observational comparison.
The registry classifies the trial as a parallel-group study.
The registry identifies the study as double masked.
The trial's primary purpose was prevention rather than treatment of established disease.
For statistical interpretation, randomization is particularly important because the treatment assignment is determined independently of the participant's subsequent outcome. That makes a between-arm comparison interpretable as a randomized treatment effect under the assumptions of the analysis, rather than simply an association between vaccination status and disease occurrence.
5. Primary Endpoint
| Endpoint | Registered definition / time frame | Type |
|---|---|---|
| Number of Participants With First Occurrence of Molecularly Confirmed Moderate to Severe/Critical COVID-19 in Seronegative Participants With Onset at Least 14 Days After the Second Vaccination | From at least 14 days after second vaccination on Day 57 (Day 71) up to end of double blind phase (7.2 months) | Binary |
The endpoint is explicitly registered as a binary outcome. Its definition requires molecular confirmation and focuses on the first occurrence of moderate to severe/critical COVID-19 in seronegative participants, with onset beginning at least 14 days after the second vaccination.
6. Statistical Methodology
Binary endpoint analysis
Because the registered primary endpoint is binary, a primary analysis would normally compare the proportion of participants experiencing the first qualifying COVID-19 event between the randomized treatment groups.
For a binary endpoint, these measures describe relative and absolute differences in event probability. A confidence interval is needed to quantify statistical uncertainty around an estimated treatment effect.
Depending on the prespecified statistical analysis plan, a randomized binary endpoint can also be evaluated using a model-based approach such as logistic regression or another appropriate regression framework. The ClinicalTrials.gov record does not provide the formal statistical-analysis specification for this primary endpoint, so a particular model should not be presented as the ENSEMBLE-2 registry's reported analysis.
Why the timing definition is statistically important
The time frame begins at least 14 days after the second vaccination on Day 57, identified in the registry as Day 71, and continues to the end of the double-blind phase. This creates a defined risk period rather than treating every post-enrollment infection as part of the primary endpoint.
First occurrence
The endpoint counts the first occurrence of the qualifying outcome. This prevents multiple episodes in the same participant from being counted as independent primary events. The resulting analysis is therefore naturally participant-based: each participant either experiences the qualifying first event during the specified period or does not.
Seronegative population
The endpoint is specifically restricted to seronegative participants. This restriction defines the population to which the primary endpoint applies and is part of the estimand being studied; it should not be silently replaced with an analysis of all enrolled participants.
7. Results Posted on ClinicalTrials.gov
The ClinicalTrials.gov record states that Results Posted: Yes and that 22 outcome measures are posted. However, no formal statistical analyses were posted. Consequently, the available trial data do not provide a formal treatment-effect estimate, confidence interval, or p-value for the registered primary endpoint.
Primary efficacy comparison
The ClinicalTrials.gov record identifies the primary endpoint and its follow-up period but do not provide a reported effect estimate, confidence interval, or p-value in the posted results.
What would normally be reported?
For this binary primary endpoint, a complete statistical result would ordinarily identify the number and proportion of participants with a qualifying event in each randomized group, an appropriate treatment-effect measure such as a risk ratio or risk difference, a confidence interval, and a hypothesis-test result when specified by the statistical analysis plan.
What an effect estimate would mean: a risk ratio would compare the probability of the qualifying first COVID-19 event between the Ad26.COV2.S and placebo groups, while a risk difference would express the absolute difference in event probability.
What it would not mean: neither measure would describe the experience of every individual participant or establish that every participant receiving the intervention would have the same outcome.
Why a confidence interval matters: the interval would quantify uncertainty around the estimated treatment effect. A narrow interval would indicate greater statistical precision than a wide interval, all else equal.
Why a p-value is not an effect size: a p-value addresses evidence against a specified null hypothesis under the analysis framework. It does not tell the reader how large or clinically important an effect is.
Important caution: the appropriate analysis must respect the registered endpoint definition and its specified follow-up period. Substituting a different population, endpoint definition, or follow-up window could answer a different statistical question.
8. Planned Analysis
Because the ClinicalTrials.gov record contains posted outcome measures but no formal statistical analyses, the appropriate role of this section is to explain how an endpoint of this type would ordinarily be analyzed without attributing an unreported method or result to the registry.
Step 1: Define the analysis population
The primary endpoint itself specifies seronegative participants and requires the qualifying event to occur within the registered post-vaccination period.
Step 2: Count first events
For each eligible participant, determine whether the first qualifying molecularly confirmed moderate to severe/critical COVID-19 event occurred during the defined interval.
Step 3: Compare randomized groups
Compare the event probability between the Ad26.COV2.S and placebo assignments using the prespecified binary-outcome method.
Step 4: Quantify uncertainty
Report an effect estimate with an appropriate confidence interval and, when specified, a hypothesis-test result.
ClinicalTrials.gov does not report a formal statistical comparison in the posted results. Therefore, this page does not state an efficacy estimate, confidence interval, or p-value that is not present in the provided record.
9. Safety Results
The ClinicalTrials.gov record includes serious adverse events by arm for the double-blind phase, as well as a combined double-blind and open-label figure. These safety results should be kept distinct from the primary efficacy endpoint because they answer a different question: whether serious adverse events occurred among participants during the specified safety period.
| Safety population / phase | Arm or population | Affected / at risk |
|---|---|---|
| Double Blind Phase | Ad26.COV2.S | 104/15705 |
| Double Blind Phase | Placebo | 136/15588 |
| Combined Double-blind and Open-label Phase | All Participants | 1583/31705 |
| Double Blind Phase: Non-FAS Post-Dose 1 | Registry-reported population | 0/3 |
| Double Blind Phase: Non-FAS Post-Dose 1 | Registry-reported population | 0/4 |
How to read the serious-adverse-event counts
The double-blind phase figures of 104/15705 and 136/15588 report the number affected and the number at risk for the Ad26.COV2.S and placebo groups, respectively. They should not be treated as a primary efficacy analysis because serious adverse events are a safety outcome rather than the registered COVID-19 endpoint.
The combined double-blind and open-label figure of 1583/31705 is presented separately because it does not have the same population definition as either double-blind treatment arm. Combining it directly with the arm-specific double-blind figures would obscure the distinction between randomized treatment groups and later open-label exposure.
10. Statistical Methods Explained
Why is the primary endpoint treated as binary?
The registry classifies the primary endpoint as binary because the endpoint asks whether a participant experienced the first qualifying molecularly confirmed moderate to severe/critical COVID-19 event during the specified period. Each eligible participant can therefore be classified according to occurrence or non-occurrence of the endpoint.
Why does the endpoint specify "first occurrence"?
Counting the first qualifying event gives each participant a single primary-event status. Repeated episodes would otherwise create a more complicated recurrent-event problem in which multiple events from the same participant could not reasonably be treated as independent observations.
Why does the endpoint specify at least 14 days after the second vaccination?
The timing criterion creates a prespecified post-vaccination evaluation period. It prevents the primary endpoint from mixing events occurring before the defined interval with events occurring during the interval that the trial was designed to evaluate.
Why does serostatus matter to the endpoint?
The primary endpoint is explicitly defined for seronegative participants. That means baseline serostatus is part of the endpoint definition rather than merely an optional subgroup characteristic. An analysis of a different serostatus population could therefore answer a different question.
Why does randomization matter for a binary comparison?
Randomization provides the principal basis for comparing the intervention and placebo groups. If the randomized groups are analyzed according to their assigned treatment under the prespecified framework, differences in the primary outcome can be interpreted within the causal structure created by random assignment, subject to the trial's analysis assumptions and population definition.
Why should a p-value not be confused with treatment effect?
A p-value and an effect estimate answer different questions. The effect estimate describes the magnitude and direction of the observed comparison, while a p-value quantifies the compatibility of the data with a specified null hypothesis under the chosen statistical model. A complete statistical report therefore needs both effect magnitude and uncertainty rather than relying on a p-value alone.
Why is the open-label phase not simply a crossover analysis?
The registry specifically states that there was no planned cross-over. The later vaccination sequences resulted from ethically allowing placebo participants to unblind themselves and receive active vaccination during the open-label phase. That is different from prospectively randomizing participants to treatment sequences for a conventional crossover analysis.
11. Analysis Population and Follow-up
The primary endpoint is defined specifically for seronegative participants and begins from at least 14 days after the second vaccination on Day 57, identified in the registry time frame as Day 71. Follow-up continues to the end of the double-blind phase, specified as 7.2 months.
| Element | Registered information |
|---|---|
| Endpoint population | Seronegative participants |
| Endpoint event | First occurrence of molecularly confirmed moderate to severe/critical COVID-19 |
| Start of endpoint window | At least 14 days after second vaccination on Day 57 (Day 71) |
| End of endpoint window | End of double blind phase |
| Registered time frame | 7.2 months |
| Endpoint type | Binary |
This distinction between enrollment and endpoint follow-up is statistically important. The enrollment number of 31835.0 describes the size of the trial, whereas the primary endpoint applies to a specifically defined population and time interval. Those quantities should not be assumed to be interchangeable.
12. Multiplicity, Interim Analysis, and Missing Data
The ClinicalTrials.gov record does not provide a formal statistical-analysis entry describing multiplicity adjustment, interim-analysis boundaries, alpha spending, missing-data imputation, or a hierarchy among multiple efficacy hypotheses.
| Design topic | What the ClinicalTrials.gov record supports |
|---|---|
| Multiplicity | No formal multiplicity procedure is in the posted results. |
| Interim analysis | No formal interim-analysis method or stopping boundary is reported. |
| Missing data / imputation | No imputation method is posted on ClinicalTrials.gov for the primary endpoint. |
| Bayesian methods | No Bayesian analysis is reported. |
| Non-inferiority margin | No non-inferiority margin is reported; the registry identifies the study as a prevention trial with a binary primary endpoint. |
| Stratification | No formal statistical stratification variables are in the posted results. |
| Crossover | The registry caveat states there was no planned cross-over. |
These omissions should not be filled with assumptions based on common vaccine-trial practice. A statistical analysis plan could specify additional procedures, but those procedures are not contained in the ClinicalTrials.gov record and therefore are not presented here as ENSEMBLE-2 results.
13. Safety and the Difference Between Efficacy and Safety Analyses
The ENSEMBLE-2 registry data illustrate why efficacy and safety should be kept as separate statistical questions. The primary endpoint is a narrowly defined COVID-19 event in a specified population and time window. Serious adverse events, by contrast, describe safety outcomes during specified treatment phases and populations.
Efficacy question
Did the participant experience the registered first occurrence of molecularly confirmed moderate to severe/critical COVID-19 during the defined primary endpoint window?
Safety question
Did the participant experience a serious adverse event during the specified safety period?
Different denominators
The safety figures use their reported at-risk populations, which differ across the double-blind and combined double-blind/open-label summaries.
Different interpretation
An efficacy endpoint and a safety endpoint should not be collapsed into one undifferentiated statistical measure.
14. Registry History and Trial Status
Trial start
The ClinicalTrials.gov record lists 2020-11-12 as the study start date.
Randomized prevention study
The study is classified as phase 3, randomized, parallel, double blind, and prevention-focused.
Outcome measures available
The registry profile states that results are posted and that 22 outcome measures are available.
Primary completion
The ClinicalTrials.gov record lists 2023-06-18 as the primary completion date.
Current registry status in the ClinicalTrials.gov record
the ClinicalTrials.gov record identifies the study as completed.
15. Understanding the Registered Endpoint Statistically
The most direct statistical representation of the binary endpoint is the proportion of eligible randomized participants who experience the first qualifying event during the registered follow-up window.
A relative measure such as a risk ratio would describe how the event probability under Ad26.COV2.S compares with the event probability under placebo. A value below 1 would correspond to a lower observed event probability in the Ad26.COV2.S group.
A risk difference would express the absolute difference in event probabilities. This can be more directly interpretable for prevention questions because it describes the difference in the probability of experiencing the specified event over the defined period.
A confidence interval around the effect estimate is essential because the observed difference is an estimate from a finite study population. The ClinicalTrials.gov record does not provide such an interval for the primary endpoint.
A p-value, if reported under the prespecified analysis, would address the compatibility of the observed data with a specified null hypothesis. It would not replace the effect estimate or confidence interval.
16. Limitations and Interpretation Issues
- No registry-reported formal primary analysis: no formal statistical analyses were posted, so this page cannot report an efficacy effect estimate, confidence interval, or p-value for the primary endpoint.
- Endpoint-specific population: the primary endpoint applies to seronegative participants rather than automatically to every enrolled participant.
- Endpoint-specific timing: the primary endpoint begins at least 14 days after the second vaccination and ends at the close of the double-blind phase.
- Open-label vaccination: later vaccination sequences should not be interpreted as a conventional planned crossover because the registry explicitly states there was no planned cross-over.
- Safety populations differ: the registry-reported serious-adverse-event figures use different populations and phases, so the combined double-blind/open-label figure should not be directly substituted for an arm-specific double-blind result.
- Unreported statistical details: the ClinicalTrials.gov record does not specify formal multiplicity control, interim-analysis boundaries, missing-data methods, Bayesian methods, or a stratified analysis plan.
- Results posted does not equal formal analysis reported: the registry profile says results are posted and 22 outcome measures are available, but the provided statistical-analysis structure contains no formal analyses.
- Incomplete registry safety caveat: the registry-reported caveat ends with an incomplete statement after noting that the data could not be analyzed. No additional interpretation is inferred from that incomplete text.
17. Why This Trial Matters Statistically
ENSEMBLE-2 is a useful statistical teaching case because it combines randomized prevention design with a tightly specified binary clinical endpoint, while also illustrating the distinction between a posted outcome measure and a formally reported statistical comparison.
| Concept | How it appears in ENSEMBLE-2 |
|---|---|
| Randomization | The registry classifies allocation as RANDOMIZED. |
| Blinding | The study is classified as DOUBLE masked. |
| Parallel design | The registry identifies the design model as PARALLEL. |
| Prevention endpoint | The primary purpose is PREVENTION. |
| Binary outcome | The registered primary endpoint is classified as Binary. |
| First-event analysis | The primary endpoint specifies the first occurrence of the qualifying COVID-19 event. |
| Defined risk window | The primary endpoint begins at least 14 days after the second vaccination and continues to the end of the double-blind phase. |
| Population definition | The primary endpoint is defined for seronegative participants. |
| Safety analysis | Serious adverse events are reported using affected/at-risk counts for specified populations and phases. |
| Open-label transition | The registry distinguishes later vaccination sequences from a planned crossover design. |
18. Statistical Methods Explained: A Deeper Walk-Through
Why might a risk difference be especially useful for a prevention endpoint?
A prevention trial ultimately concerns whether the probability of experiencing an event changes over a defined period. A risk difference directly expresses that absolute change. Relative measures can describe proportional differences, but they do not by themselves communicate the absolute event probability in either group.
Why should the primary endpoint window be prespecified?
If investigators could freely select the most favorable follow-up interval after examining the data, the estimated treatment effect could become biased by selective reporting. A prespecified window establishes which events belong to the primary question before the outcome comparison is interpreted.
What does "molecularly confirmed" contribute statistically?
It establishes an objective eligibility criterion for the primary event. The endpoint is not simply a report of suspected COVID-19; the registry requires a molecularly confirmed event as part of the primary endpoint definition.
Why is "moderate to severe/critical" important?
The severity criterion narrows the endpoint to clinically specified disease rather than every possible infection. This affects both the event rate and the clinical interpretation of any treatment-effect estimate.
Why would an odds ratio not automatically be the same as a risk ratio?
An odds ratio compares odds rather than probabilities. When an event is uncommon, the two measures can be numerically similar, but they are not mathematically identical. The appropriate effect measure should therefore follow the prespecified statistical analysis rather than being chosen after seeing the results.
Why should the open-label phase be separated from the randomized double-blind comparison?
Once participants can unblind and receive active vaccination outside the original randomized assignment, subsequent exposure is no longer represented by the original two-arm randomized comparison in the same way. The registry caveat therefore matters when interpreting safety and any later outcome measures.
Why can posted outcome measures and statistical analyses be different?
A clinical-trial registry can contain outcome-measure records without supplying a formal statistical-analysis object for every outcome. In this the ClinicalTrials.gov recordset, that distinction is explicit: results are posted and 22 outcome measures are listed, but no formal statistical analyses were posted.
19. Related Tutorials
Learn more about the methods used in this trial:
20. Related Statistical Calculators
21. Sources
- ClinicalTrials.gov: NCT04614948 — ENSEMBLE-2.
- Linked publication: PubMed record for PMID 36113538.
Continue through the Clinical Biostats statistical library
Explore clinical-trial methods, statistical concepts, and calculators that provide deeper treatment of the design and analysis principles illustrated by ENSEMBLE-2.
22. Record Summary
ENSEMBLE-2 is a completed phase 3, randomized, double-blind, parallel prevention trial with 31835.0 enrolled participants and two intervention groups: Ad26.COV2.S and placebo. Its registered primary endpoint is a binary measure of the first occurrence of molecularly confirmed moderate to severe/critical COVID-19 in seronegative participants, beginning at least 14 days after the second vaccination on Day 57 (Day 71) and continuing through the end of the double-blind phase, specified as 7.2 months.
The ClinicalTrials.gov record states that results are posted and contains 22 outcome measures, but it provides no formal statistical analyses. The statistically responsible interpretation is therefore to distinguish reported outcome data from a formal comparative efficacy analysis.