This page provides an independent statistical analysis and educational interpretation of publicly reported results. ClinicalTrials.gov provides the official trial registry record. Numerical trial facts on this page are restricted to the registry-reported ROADSTER trial data.
1. Trial at a Glance
ROADSTER was a completed phase 3, single-group study of the MICHI NPS+f device in patients with carotid artery disease. The registered primary endpoint was a hierarchical composite of stroke, myocardial infarction, and death during a 30-day post-procedural period.
| Feature | ROADSTER |
|---|---|
| Trial name | ROADSTER |
| ClinicalTrials.gov identifier | NCT01685567 |
| Phase | Phase 3 |
| Status | Completed |
| Condition | Carotid Artery Disease |
| Allocation | NA |
| Design model | Single group |
| Masking | None |
| Primary purpose | Prevention |
| Enrollment | 219 |
| Intervention | MICHI NPS+f (device) |
| Lead sponsor | Silk Road Medical |
| Sponsor type | Industry |
2. Clinical Question
The registered study evaluated the use of the MICHI NPS+f device during a carotid artery stenting procedure in patients with carotid artery disease. The primary statistical question is centered on the occurrence of a hierarchical composite of clinically important events during the specified 30-day post-procedural period.
Population
The trial enrolled 219 participants in a phase 3 study for carotid artery disease. The primary endpoint definition specifies an ITT population comprising pivotal and extended enrollment subjects deemed to be at high risk for complications from CEA.
Intervention
MICHI NPS+f, identified in the registry data as a device intervention.
Comparator
No separate comparator arm was registered. The design model was single group and masking was none.
Primary question
What was observed for the registered hierarchical composite of stroke, myocardial infarction, and death during the 30-day post-procedural period?
3. Trial Design
MICHI NPS+f
- Device intervention
- Used during a carotid artery stenting procedure
- Study design: single group
- Masking: none
- Primary purpose: prevention
Interpretive consequence
- No randomized control group is identified in the ClinicalTrials.gov record
- No allocation ratio is reported
- No between-arm hazard ratio or risk ratio is available in the statistical analyses posted on ClinicalTrials.gov
- Formal comparative treatment inference is therefore not reported here
The study began in 2012-11 and had a primary completion date of 2016-03. The registry status in the ClinicalTrials.gov record is COMPLETED.
4. Primary Endpoint
| Endpoint | Registered definition | Time frame |
|---|---|---|
| Hierarchical Composite of Stroke, Myocardial Infarction, and Death | The primary endpoint was a hierarchical composite of any stroke, myocardial infarction and death during a 30-day post-procedural period in the ITT (pivotal and extended enrollment) population comprised of subjects deemed to be high risk for complications from CEA. | 30-day post-procedure |
The wording is important. This is not simply a list of three unrelated outcomes. The registered endpoint is explicitly described as a hierarchical composite, meaning that the endpoint structure gives the component outcomes an ordered role in the composite analysis. The ClinicalTrials.gov record does not provide the hierarchy's numerical scoring rules or a formal statistical-analysis record, so this page does not reconstruct an unreported scoring algorithm.
For a hierarchical composite, the statistical analysis normally specifies how patients with different component events are ordered and compared. The exact ROADSTER analysis specification is not contained in the posted results.
5. Results
Primary Endpoint Results
The ClinicalTrials.gov record identifies the registered primary endpoint and confirm that results were posted for it. However, no formal statistical analyses were posted. No primary-endpoint estimate, confidence interval, p-value, or formal comparison is provided in the ClinicalTrials.gov record.
| Reported item | What the ClinicalTrials.gov record establishes |
|---|---|
| Results posted | Yes |
| Primary endpoint | Hierarchical composite of stroke, myocardial infarction, and death |
| Primary endpoint time frame | 30-day post-procedure |
| Primary analysis population | ITT (pivotal and extended enrollment) population |
| Formal statistical analyses reported | None |
| Reported effect estimate reported | None |
| Reported confidence interval reported | None |
| Reported p-value reported | None |
The absence of a registry-reported hazard ratio, risk ratio, odds ratio, confidence interval, or p-value means that no quantitative comparative effect should be inferred from the registry summary provided here. A single-group study can describe how often an endpoint occurred, but that is different from demonstrating a relative treatment effect against a concurrent randomized comparator.
A confidence interval, if reported for a single-group event proportion, would quantify statistical uncertainty around the estimated event probability. It would not establish superiority or non-inferiority to an external comparator unless a prespecified benchmark and corresponding statistical framework were provided.
A p-value, if reported, would test a specific null hypothesis defined by the analysis plan. It would not itself measure the size or clinical importance of an observed effect. No such p-value is reported in the ClinicalTrials.gov record for ROADSTER.
Serious Adverse Events
the ClinicalTrials.gov record reports serious adverse events for the intention-to-treat population as 29/219.
Serious adverse events
Reported for the Intention-to-Treat population
The reported value identifies the number affected and the population at risk: 29 participants out of 219 in the ITT population were reported as having serious adverse events. The ClinicalTrials.gov record does not provide a comparator-arm rate, a confidence interval, a p-value, individual serious-adverse-event categories, or a formal statistical comparison.
Because ROADSTER is registered as a single-group study, this safety result is descriptive in the ClinicalTrials.gov recordset. It should not be converted into a claim that the device increases or decreases serious adverse events relative to another intervention without an appropriate comparator and analysis.
6. Planned Analysis
Because the ClinicalTrials.gov record contains results but no formal statistical analyses, the most useful statistical interpretation is to describe how an endpoint of this type would normally be analysed while clearly separating that methodological explanation from what ROADSTER actually reports in the ClinicalTrials.gov record.
How a hierarchical composite would normally be analysed
A hierarchical composite generally requires an analysis rule that determines how the individual component outcomes are ordered. Depending on the prespecified design, investigators may compare patients according to the most important event they experienced, or use a formal rank-based method that gives priority to more clinically consequential outcomes. The appropriate method depends on the exact protocol and statistical analysis plan.
For ROADSTER, the registry defines the primary endpoint as a hierarchical composite of any stroke, myocardial infarction and death over a 30-day post-procedural period. The ClinicalTrials.gov record does not identify the formal statistical test or effect measure used for this endpoint.
How a single-group analysis differs from a randomized comparison
In a randomized two-arm trial, a composite endpoint can be compared between treatment groups using a risk ratio, odds ratio, time-to-event method, difference in proportions, or another prespecified measure. In a single-group study, there is no randomized concurrent comparator against which to estimate such a treatment effect.
Descriptive question
How many participants experienced the endpoint during the specified follow-up period?
Comparative question
Is the event experience different from that under another treatment or prespecified benchmark?
ROADSTER data reported
The ClinicalTrials.gov record supports the study design and endpoint definition, but do not supply a formal comparative analysis.
Statistical implication
Observed results should be interpreted descriptively unless a valid prespecified comparator or benchmark is available.
7. Statistical Methodology
Intention-to-treat population
The registered primary endpoint definition specifies an ITT population, described as the pivotal and extended enrollment population comprised of subjects deemed to be high risk for complications from CEA. The ITT principle generally preserves the population defined by study enrollment rather than restricting the primary analysis to participants who completed treatment exactly as planned.
For a single-group device study, an ITT analysis can be useful because excluding participants after enrollment may introduce selection into the endpoint analysis. However, the ClinicalTrials.gov record does not provide enough detail to determine whether any participants were excluded from particular analyses or how missing endpoint assessments were handled.
Hierarchical composite endpoints
A hierarchical composite differs from a simple composite in which every component event is treated identically. The statistical method must define how the ordering of stroke, myocardial infarction, and death contributes to the final analysis.
The exact ROADSTER hierarchy and analysis statistic are not included in the ClinicalTrials.gov record. A precise reconstruction therefore requires the underlying protocol or statistical analysis plan rather than an inferred method.
30-day follow-up
The primary endpoint has a fixed 30-day post-procedural time frame. A fixed assessment window makes the endpoint a bounded post-procedure outcome rather than an indefinitely observed time-to-event endpoint.
If every participant had complete ascertainment through the full 30-day period, a simple event proportion could describe the frequency of the composite. If follow-up were incomplete, the statistical method would need to specify how incomplete observations and censoring were handled. The registry-reported ROADSTER data do not provide that information.
Safety analysis
The ClinicalTrials.gov record is reported for the ITT population as 29/219 serious adverse events. The data do not identify a separate as-treated safety population, exposure-adjusted rate, confidence interval, or comparative test.
8. Statistical Methods Explained
Why does the hierarchical composite matter?
A conventional composite endpoint combines several clinical outcomes into one endpoint, but a hierarchical composite adds an ordering principle. This can preserve more information about the relative clinical importance of different outcomes than simply counting whether any component occurred. The validity of the result depends on the hierarchy and statistical procedure being prespecified and appropriately analyzed.
Why is the 30-day time frame important?
The 30-day post-procedural window defines the period during which the primary endpoint is evaluated. It prevents later events from being silently incorporated into the primary outcome. It also means that the primary endpoint is specifically a short-term post-procedural assessment rather than a statement about lifetime risk.
Why is a single-group design different from a randomized trial?
Randomization creates a concurrent comparison group whose outcomes provide the counterfactual reference for treatment-effect estimation. In a single-group study, that randomized counterfactual is absent. Observed event experience can therefore be described, but a causal treatment effect requires an external or prespecified reference with an appropriate statistical framework.
What would a confidence interval tell us?
For an estimated event proportion, a confidence interval would describe statistical uncertainty around the population quantity being estimated. For a comparative effect measure, it would describe uncertainty around the treatment contrast. In either case, the interval is not a prediction interval for an individual patient and does not automatically establish clinical importance.
Why does the p-value not measure effect size?
A p-value evaluates compatibility with a specified null hypothesis under the statistical model. It does not tell us how large an effect is. Effect size is conveyed by an appropriate measure such as a difference, ratio, or other prespecified statistic, together with its uncertainty.
What does the absence of a formal statistical analysis mean here?
It means that the ClinicalTrials.gov record does not provide the numerical inferential analysis needed to report a formal treatment effect. It does not mean that no analysis occurred during the actual study. It means only that no entries are present in the posted results, so this page does not reconstruct unreported numerical analyses.
9. Secondary Endpoints and Additional Results
the ClinicalTrials.gov record states that 5 outcome measures were posted. However, the ClinicalTrials.gov record does not provide the definitions, time frames, numerical results, or formal statistical analyses for those additional outcome measures.
| Registry result information | registry data |
|---|---|
| Outcome measures posted | 5 |
| Primary endpoints registered | 1 |
| Primary endpoint results posted | Yes |
| Formal statistical analyses reported | None |
| Additional endpoint estimates reported | None in the provided statistical-analysis data |
Accordingly, no additional endpoint estimate, confidence interval, p-value, subgroup result, or efficacy claim is added here. The distinction is important because the presence of posted outcome measures does not by itself provide enough information to reconstruct their statistical results.
10. Design Topics Not Reported in the ClinicalTrials.gov record
| Topic | What can be stated from the ClinicalTrials.gov record |
|---|---|
| Non-inferiority margin | Not reported |
| Crossover | Not reported |
| Factorial design | Not reported |
| Interim analysis | Not reported |
| Missing-data or imputation method | Not reported |
| Stratification factors | Not reported |
| Bayesian methods | Not reported |
| Formal multiplicity strategy | Not reported |
| Statistical-analysis entries | None reported |
11. Safety
The ClinicalTrials.gov record provides one safety summary: serious adverse events in the ITT population were reported as 29/219.
| Safety measure | Population | Affected / at risk |
|---|---|---|
| Serious adverse events | Intention-to-Treat | 29/219 |
No arm-to-arm safety comparison is possible because the registry-reported design is single group. The ClinicalTrials.gov record also do not distinguish individual serious-adverse-event categories, severity distributions, timing, treatment relationship, or discontinuations.
12. Interpreting the ROADSTER Evidence
The ROADSTER registry record describes a completed phase 3, single-group study of MICHI NPS+f in carotid artery disease, with a registered 30-day hierarchical composite of stroke, myocardial infarction, and death as the primary endpoint.
The ClinicalTrials.gov record establishes an enrollment of 219 and report serious adverse events affecting 29 of 219 participants in the ITT population. Results were posted, but no formal statistical-analysis entries were reported.
A single-group design does not create a randomized comparator simply because a numerical outcome is observed.
13. Limitations
- Single-group design: the ClinicalTrials.gov record identifies one registered study arm, so there is no randomized concurrent comparator in the ClinicalTrials.gov record.
- No formal statistical analyses reported: no formal statistical analyses were posted, preventing reporting of an inferential effect estimate, confidence interval, or p-value.
- Composite endpoint detail: the registry definition identifies the endpoint as hierarchical but the ClinicalTrials.gov record does not specify the numerical hierarchy or scoring procedure.
- Limited outcome detail: five outcome measures were posted, but the ClinicalTrials.gov record does not provide the additional outcome definitions or numerical results.
- Safety detail: the ClinicalTrials.gov record consists of serious adverse events reported as 29/219 and does not provide category-level or comparative safety analyses.
- Missing statistical-design information: the ClinicalTrials.gov record does not report the handling of missing data, interim analyses, multiplicity, stratification, Bayesian methods, or other potentially important statistical features.
- Interpretation of observed events: an observed event count describes experience within the study population but does not by itself quantify a causal treatment effect.
14. Why This Trial Matters Statistically
ROADSTER is useful as a teaching case because it illustrates an important distinction in clinical-trial statistics: a phase 3 designation does not by itself determine whether the primary analysis is randomized, comparative, or based on a particular inferential framework.
| Concept | How it appears in ROADSTER |
|---|---|
| Single-group design | The registry identifies a SINGLE_GROUP design with one registered arm. |
| Device evaluation | MICHI NPS+f is the registered intervention. |
| Hierarchical composite | The primary endpoint combines stroke, myocardial infarction, and death hierarchically. |
| Fixed follow-up window | The primary endpoint is evaluated during a 30-day post-procedural period. |
| ITT analysis | The primary endpoint definition specifies the ITT pivotal and extended enrollment population. |
| Descriptive safety | Serious adverse events are reported as 29/219 in the ITT population. |
| Inferential restraint | No formal statistical-analysis entries are reported, so comparative effect measures are not reconstructed. |
| Composite-endpoint methodology | The analysis requires attention to component ordering rather than treating the endpoint as an undifferentiated count. |
The most important lesson is methodological: the statistical question must match the design. A randomized comparison can estimate a treatment contrast because randomization supplies a concurrent counterfactual. A single-group study instead requires careful descriptive analysis or a prespecified external benchmark if the objective is to make a formal comparative claim.
15. Related Statistical Concepts
Learn more about the methods used in this trial:
16. Related Statistical Calculators
17. Sources
- ClinicalTrials.gov: NCT01685567 — ROADSTER.
- PubMed: PMID 26506270.
- PubMed: PMID 28335706.
- PubMed: PMID 36563712.
- PubMed: PMID 30611582.
Continue with Clinical Biostats
Explore statistical tutorials and calculators that explain the endpoint, design, and inference concepts used in clinical trials.
18. Record Summary
ROADSTER provides a useful statistical case study in how trial design determines what can legitimately be inferred from observed clinical data. The study was a completed phase 3, single-group evaluation of MICHI NPS+f in carotid artery disease, with 219 participants enrolled and a registered primary endpoint consisting of a hierarchical composite of stroke, myocardial infarction, and death during a 30-day post-procedural period. The primary endpoint was defined in the ITT pivotal and extended enrollment population. The ClinicalTrials.gov record reports serious adverse events as 29/219 in the ITT population.
The central statistical limitation is equally important: although results were posted, no formal statistical analyses were posted. Instead, it distinguishes the observed study information from the inferential methods that would normally be required for a hierarchical composite endpoint and emphasizes the difference between descriptive evidence from a single-group study and causal comparative evidence from a randomized trial.