This page separates information reported in the ClinicalTrials.gov record from statistical interpretation and educational explanation. The registry extract reports that results have been posted, but it contains no formal statistical-analysis records.
1. Trial at a Glance
NSABP B-30 was a randomized, parallel-group, open-label phase 3 treatment trial with 5351 participants. The registry identifies two primary time-to-event endpoints: Overall Survival over 8 years and Disease Free Survival, defined as time to breast cancer recurrence, second primary cancer, or death from any cause as a first event.
| Feature | NSABP B-30 |
|---|---|
| Trial name | NSABP B-30 |
| NCT identifier | NCT00003782 |
| Phase | 3 |
| Status | COMPLETED |
| Therapeutic area | Oncology |
| Condition | Breast Cancer |
| Allocation | RANDOMIZED |
| Design model | PARALLEL |
| Masking | NONE |
| Primary purpose | TREATMENT |
| Enrollment | 5351 |
| Study start | 1999-03 |
| Primary completion | 2008-12 |
| Lead sponsor | NSABP Foundation Inc |
| Sponsor type | NETWORK |
| Results posted | Yes |
2. Clinical Question
The registry describes a randomized phase 3 treatment study in women with stage I, stage II, or stage IIIA (cT1-3, N0-1, M0) breast cancer and positive axillary lymph nodes. The interventions listed in the ClinicalTrials.gov record is cyclophosphamide, docetaxel, and doxorubicin.
Population
Women with stage I, stage II, or stage IIIA (cT1-3, N0-1, M0) breast cancer and positive axillary lymph nodes.
Interventions
Cyclophosphamide, docetaxel, and doxorubicin are the interventions identified in the ClinicalTrials.gov record.
Comparator structure
The registry identifies 3 randomized arms. The ClinicalTrials.gov record provides the arm names associated with the serious-adverse-event summary, but do not provide a separate narrative description of the treatment comparison for each efficacy endpoint.
Primary questions
The registered primary endpoints ask how treatment assignment relates to Overall Survival over 8 years and Disease Free Survival over the specified event-based follow-up.
3. Trial Design
The absence of masking is statistically relevant because treatment assignment is not concealed from participants or investigators after randomization. For objective time-to-event outcomes, randomization remains the central design feature supporting comparison between treatment groups, while lack of masking can matter more strongly for outcomes involving clinical assessment or treatment-related decisions.
Three-arm randomized structure
A 3-arm parallel design allows more than one treatment strategy to be evaluated within a single randomized trial. The statistical consequence is that the number and nature of pairwise comparisons matter. If multiple pairwise efficacy hypotheses are formally tested, the statistical analysis plan must specify how multiplicity is handled. The ClinicalTrials.gov record does not report such a multiplicity procedure, so no particular adjustment is attributed to NSABP B-30 on this page.
Doxorubicin + Cyclophosphamide, Then Doc
- Serious adverse events: 66 affected out of 1748 at risk.
Doxorubicin + Docetaxel
- Serious adverse events: 43 affected out of 1748 at risk.
Doxorubicin + Docetaxel + Cyclophosphami
- Serious adverse events: 46 affected out of 1745 at risk.
Intervention set
- Cyclophosphamide
- Docetaxel
- Doxorubicin
4. Endpoints
| Endpoint | Registered definition / time frame | Endpoint type | Formal analysis in the ClinicalTrials.gov record |
|---|---|---|---|
| Overall Survival | 8 years | Time-to-event | Not in the posted results |
| Disease Free Survival | time to event: breast cancer recurrence; second primary cancer; death from any cause as a first event | Time-to-event | Not in the posted results |
Overall Survival
Overall Survival is registered with a time frame of 8 years. For a conventional time-to-event analysis, the event would be death from any cause, with patients who remain alive at the last available follow-up contributing censored follow-up time.
Disease Free Survival
The registry defines Disease Free Survival as time to breast cancer recurrence, second primary cancer, or death from any cause as a first event. This is a composite time-to-first-event endpoint: the first qualifying event determines the observed event time.
5. Results Status and Planned Analysis
The ClinicalTrials.gov record states that Results Posted is “Yes” and that results are posted for both registered primary endpoints. However, no formal statistical analyses were posted, with 0 analyses omitted for length. Therefore, this page does not report a formal treatment-effect estimate, confidence interval, p-value, event count, or median time for either primary endpoint.
What can be established from the ClinicalTrials.gov record
Overall Survival at 8 years and Disease Free Survival defined by breast cancer recurrence, second primary cancer, or death from any cause as a first event.
The ClinicalTrials.gov record does not provide formal statistical-analysis results for either endpoint.
Planned Analysis: Overall Survival
For an endpoint defined as time to death, a standard analysis would estimate the survival function with the Kaplan-Meier method and compare randomized groups with an appropriate time-to-event test. A Cox proportional-hazards model is commonly used to estimate a hazard ratio and confidence interval when its assumptions are appropriate.
Those are typical statistical methods for this endpoint, not methods attributed to NSABP B-30 in the registry extract. The ClinicalTrials.gov record does not identify the statistical model, comparison test, hazard ratio, confidence interval, p-value, or analysis population used for the posted Overall Survival results.
Planned Analysis: Disease Free Survival
Because Disease Free Survival is a time-to-first-event endpoint, Kaplan-Meier estimation and time-to-event comparison methods would ordinarily be appropriate. A Cox model could also be used to estimate a relative hazard if its assumptions are satisfied.
Again, these methods describe the standard statistical framework for an endpoint of this type. They do not establish that NSABP B-30 used a particular model or test because the ClinicalTrials.gov record does not contain that information.
6. Statistical Methodology
Time-to-event analysis
Both registered primary endpoints are time-to-event outcomes. This changes the statistical question from simply asking whether an event occurred to asking when the first qualifying event occurred while accounting for participants whose event status is not observed during the available follow-up.
The survival function represents the probability that the event time T exceeds a specified time t. In clinical trials, censored participants contribute information up to their censoring time rather than being treated as if an event occurred at that time.
Kaplan-Meier estimation
The Kaplan-Meier estimator is commonly used to describe the distribution of a time-to-event endpoint. It updates the estimated probability of remaining event-free whenever an event occurs and accounts for the number of participants still at risk immediately before that event time.
Here di is the number of events at time ti, and ni is the number at risk immediately before that time.
Hazard ratios
A hazard ratio is a relative measure of the instantaneous event rate between two groups under a specified statistical model. For example, an HR below 1 would correspond to a lower modeled instantaneous event rate in the numerator treatment group relative to the reference group.
The hazard ratio is not an absolute risk difference, not a probability of survival at a particular time, and not a statement that every individual experiences the same proportional change in risk.
Three-arm comparisons
With 3 randomized arms, there can be several scientifically relevant comparisons. The statistical analysis plan should specify which comparison or comparisons address the primary question. If several hypotheses are tested, the interpretation of p-values depends on the prespecified multiplicity strategy.
Randomization
Randomization is the central design mechanism that creates treatment groups whose outcomes can be compared without assigning treatment according to observed baseline characteristics. It does not guarantee that every baseline characteristic will be numerically identical between groups; rather, it provides the foundation for a valid randomized comparison under the trial's design and analysis assumptions.
Unmasked design
The registry identifies the study as having Masking: NONE. In an unmasked trial, treatment assignment is known rather than concealed after randomization. For time-to-event outcomes, the statistical implications depend on how events are defined, assessed, and recorded. If outcome assessment or subsequent clinical management can be influenced by knowledge of treatment assignment, that can introduce considerations that are different from those in a blinded trial.
7. Statistical Methods Explained
Why is randomization important in a 3-arm trial?
Randomization determines treatment assignment rather than allowing treatment choice to depend on patient or investigator characteristics. In a 3-arm trial, it creates multiple randomized groups within the same study, allowing prespecified comparisons while maintaining a common trial population and study infrastructure.
Why are Overall Survival and Disease Free Survival both time-to-event endpoints?
Neither endpoint is defined only by whether an event eventually occurs. The timing of the event is part of the outcome. Participants may also remain event-free for the duration of their observed follow-up, which creates right-censored observations. Time-to-event methods are designed to use the available follow-up without incorrectly treating censored participants as events.
Why is Disease Free Survival a composite endpoint?
The registered definition includes breast cancer recurrence, second primary cancer, and death from any cause as possible first events. Combining these event types produces a single time-to-first-event endpoint. The interpretation therefore depends on the relative contribution and clinical meaning of the component events, not merely on a single undifferentiated count.
What would a hazard ratio tell us?
A hazard ratio would summarize the relative instantaneous event rate between randomized groups under the chosen model. An HR of 1 would represent equal modeled hazards, whereas an HR below 1 would indicate a lower modeled hazard for the numerator group. It would not directly tell us the difference in survival probability at a specific time.
Why does the 8-year time frame matter for Overall Survival?
The registry explicitly specifies an 8-year time frame for Overall Survival. A time-specific survival estimate answers a different question from a median survival time or a hazard ratio. Keeping the registered time frame intact prevents different follow-up summaries from being silently treated as interchangeable measures.
Why does a 3-arm design create a multiplicity issue?
If the scientific question requires several pairwise comparisons among 3 arms, each additional hypothesis test can increase the chance of at least one false-positive result when tests are considered collectively. A prespecified multiplicity strategy can control the relevant familywise error rate or otherwise define how confirmatory conclusions are made. The registry-reported NSABP B-30 data do not identify such a procedure.
What does an empty statistical-analysis record mean?
It means that the ClinicalTrials.gov recordset does not provide the formal statistical-analysis records needed to reproduce the numerical efficacy results. It does not mean that the trial had no statistical analysis. The registry data explicitly state that results were posted, while the registry-reported extract contains no formal statistical analyses. Those two facts should be kept separate.
8. Safety Results
The ClinicalTrials.gov record provides serious adverse-event counts by arm. These are presented exactly as reported in the registry, using the affected and at-risk counts rather than deriving additional percentages.
| Arm | Serious adverse events | Affected / at risk |
|---|---|---|
| Doxorubicin + Cyclophosphamide, Then Doc | Serious adverse events | 66 / 1748 |
| Doxorubicin + Docetaxel | Serious adverse events | 43 / 1748 |
| Doxorubicin + Docetaxel + Cyclophosphami | Serious adverse events | 46 / 1745 |
The raw counts should therefore not be interpreted as evidence of a statistically significant difference between arms. Differences in affected counts can reflect differences in the numbers at risk as well as differences in event occurrence. A formal safety comparison would require the prespecified safety population and analysis framework.
9. Analysis Populations and Missing Data
The registry extract does not specify the intention-to-treat population, per-protocol population, safety population definition, censoring rules, missing-data strategy, or imputation method for the primary efficacy analyses.
What is known
The trial was randomized, had 3 arms, enrolled 5351 participants, and had two registered primary time-to-event endpoints.
What is not specified
The ClinicalTrials.gov record does not identify the exact efficacy analysis population or the rules used for censoring, missing assessments, or protocol deviations.
Censoring
Time-to-event analysis depends on defining when a participant's event-free follow-up ends if the event has not occurred. In general survival analysis, such observations are censored rather than counted as events. The validity of the resulting estimate depends on the appropriateness of the censoring mechanism and the prespecified rules used to determine censoring dates.
Missing data and imputation
The ClinicalTrials.gov record does not identify an imputation method. For time-to-event endpoints, investigators often address incomplete follow-up through censoring rules rather than imputing a continuous outcome value. The exact handling remains a trial-specific methodological question and should not be inferred from the endpoint name alone.
10. Stratification and Covariate Adjustment
The ClinicalTrials.gov record identifies the trial as randomized but do not report stratification factors or covariates used in the primary statistical models.
This distinction matters because a stratified log-rank test and a stratified Cox model make different analytical commitments from their unstratified counterparts. Without the actual statistical-analysis specification, it is not appropriate to claim that a particular stratification scheme was used.
11. Interim Analysis, Multiplicity, and Other Design Features
| Design topic | What the ClinicalTrials.gov record establishes |
|---|---|
| Interim analysis | Not reported in the ClinicalTrials.gov record. |
| Alpha spending | Not reported in the ClinicalTrials.gov record. |
| Multiplicity adjustment | Not reported in the ClinicalTrials.gov record. |
| Non-inferiority margin | Not applicable from the registry-reported endpoint/design information; no non-inferiority design is reported. |
| Crossover | Not reported in the ClinicalTrials.gov record. |
| Factorial design | Not reported; the registry design model is PARALLEL. |
| Bayesian methods | Not reported in the ClinicalTrials.gov record. |
| Missing-data/imputation method | Not reported in the ClinicalTrials.gov record. |
These omissions are analytically important. Trial conclusions depend not only on the endpoint definition but also on the prespecified analysis plan, treatment comparisons, error-control strategy, handling of incomplete follow-up, and analysis population.
12. Interpreting Time-to-Event Evidence Without a Reported Effect Estimate
A hazard ratio would provide a relative comparison of modeled instantaneous event rates between specified treatment groups. It would not by itself provide the probability of being alive or disease-free at 8 years, nor would it describe the experience of every individual participant.
A confidence interval around a hazard ratio would describe statistical uncertainty around the estimated relative effect under the specified model and sampling framework. A narrower interval generally represents greater statistical precision than a wider interval, although precision and clinical importance are separate concepts.
A p-value addresses evidence against a specified null hypothesis under the statistical model. It does not measure the magnitude of the treatment effect, the probability that a treatment is beneficial, or the clinical importance of an observed difference. Effect estimates and their confidence intervals are therefore essential companions to hypothesis tests.
A Cox-model hazard ratio is often interpreted as a relative hazard over follow-up, but the model's interpretation depends on its assumptions. If hazards vary substantially in a non-proportional way, a single hazard ratio may not fully describe the treatment-group survival experience. Time-specific survival estimates and graphical assessments can provide complementary information.
13. Primary Endpoint Interpretation Framework
| Question | Overall Survival | Disease Free Survival |
|---|---|---|
| What is measured? | Time to the registered Overall Survival event over 8 years. | Time to breast cancer recurrence, second primary cancer, or death from any cause as a first event. |
| Statistical type | Time-to-event. | Time-to-event. |
| Typical descriptive analysis | Kaplan-Meier survival estimates. | Kaplan-Meier event-free estimates. |
| Typical relative-effect analysis | Cox-model hazard ratio, if appropriate. | Cox-model hazard ratio, if appropriate. |
| Key interpretation issue | Death is the event; participants still alive may be censored. | Several clinically different event types can define the first event. |
| Results available in registry-reported analysis field? | No formal statistical-analysis record reported. | No formal statistical-analysis record reported. |
14. Limitations
- Formal efficacy estimates are absent from the ClinicalTrials.gov record: although the registry states that results have been posted, no formal statistical analyses were posted.
- No confidence intervals or p-values are available in the registry-reported extract:
- Three-arm multiplicity is unresolved in the ClinicalTrials.gov record: the exact confirmatory comparison structure and any multiplicity adjustment are not reported.
- Analysis populations are not specified: the ClinicalTrials.gov record does not define the precise ITT, per-protocol, or safety populations.
- Censoring rules are not specified: the registry extract does not provide detailed rules for handling incomplete follow-up for either time-to-event endpoint.
- Missing-data methods are not specified: no imputation strategy is reported in the ClinicalTrials.gov record.
- Unmasked design: Masking is listed as NONE, which can be relevant when clinical assessment, treatment decisions, or event ascertainment can be influenced by treatment knowledge.
- Endpoint composition: Disease Free Survival combines breast cancer recurrence, second primary cancer, and death from any cause as a first event; these components are clinically distinct.
- Safety counts are descriptive: the registry-reported serious-adverse-event counts do not include formal between-arm statistical comparisons.
- Registry extract limitation: this page is constrained to the trial data posted on ClinicalTrials.gov for the analysis and therefore does not substitute for a full statistical analysis plan or complete clinical study report.
15. Why This Trial Matters Statistically
NSABP B-30 is a useful statistical teaching case because its registry design combines randomized allocation, 3 parallel treatment arms, and two primary time-to-event endpoints. Those features create several important analysis questions even before any numerical efficacy estimate is considered.
| Concept | How it appears in NSABP B-30 |
|---|---|
| Randomization | The registry identifies RANDOMIZED allocation. |
| Three-arm design | The study contains 3 treatment arms. |
| Parallel design | The registered design model is PARALLEL. |
| Time-to-event analysis | Both primary endpoints are registered as time-to-event outcomes. |
| Overall Survival | Registered with an 8-year time frame. |
| Disease Free Survival | Defined by recurrence, second primary cancer, or death as a first event. |
| Censoring | Relevant to the analysis of time-to-event outcomes, although exact registry censoring rules are not reported. |
| Hazard ratio | A standard potential relative-effect measure for time-to-event outcomes, but no trial-specific HR is reported in the statistical-analysis field. |
| Multiplicity | Potentially important in a 3-arm trial with multiple primary endpoints; the ClinicalTrials.gov record does not specify the adjustment. |
| Unmasked treatment | Masking is listed as NONE. |
| Safety analysis | Serious adverse-event affected/at-risk counts are available for each arm. |
16. Related Tutorials
Learn more about the methods used in this trial:
17. Related Statistical Calculators
18. Sources
- ClinicalTrials.gov: NSABP B-30, NCT00003782.
- PubMed record: PMID 18302020.
- PubMed record: PMID 28954297.
- PubMed record: PMID 20519679.
Continue through the Clinical Biostats trial-analysis pathway
Use the trial record to connect randomized design, time-to-event endpoints, survival analysis, and clinical-trial methodology.
19. Record Summary
NSABP B-30 is a completed randomized phase 3 trial with 5351 participants, 3 parallel treatment arms, and two registered primary time-to-event endpoints. The registry defines Overall Survival over 8 years and Disease Free Survival as time to breast cancer recurrence, second primary cancer, or death from any cause as a first event. The ClinicalTrials.gov record also provides serious-adverse-event counts for each of the 3 arms.
The statistical structure is therefore centered on randomized multi-arm comparison and time-to-event methodology. Kaplan-Meier estimation, appropriate time-to-event hypothesis testing, and hazard-ratio modeling are natural analytical frameworks for endpoints of this type, but the registry extract does not identify the formal statistical methods used for the posted efficacy results. It also does not provide the corresponding effect estimates, confidence intervals, or p-values.