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
BILCAP was a randomized phase 3 treatment trial evaluating capecitabine or clinical observation after surgery in patients with biliary tract cancer. The registry lists survival at 2 years as the primary endpoint and reports an enrollment of 360 participants.
| Feature | BILCAP |
|---|---|
| Acronym | BILCAP |
| ClinicalTrials.gov identifier | NCT00363584 |
| Phase | Phase 3 |
| Status | Completed |
| Start date | 2006-03 |
| Primary completion date | 2013-08 |
| Lead sponsor | University Hospital Southampton NHS Foundation Trust |
| Conditions | Extrahepatic Bile Duct Cancer; Gallbladder Cancer; Liver Cancer |
| Allocation | Randomized |
| Primary purpose | Treatment |
| Enrollment | 360 |
2. Clinical Question
The registry describes BILCAP as a study of capecitabine or observation after surgery in patients with biliary tract cancer. Statistically, the central question is whether survival at 2 years differs between the randomized treatment strategies.
Population
Patients with biliary tract cancer, including extrahepatic bile duct cancer, gallbladder cancer, and liver cancer, enrolled in a randomized phase 3 treatment trial after surgery.
Intervention
Capecitabine, listed in the registry as a drug intervention.
Comparator
Clinical observation, listed in the registry as the other intervention.
Primary question
Does the randomized comparison of capecitabine versus clinical observation differ with respect to survival at 2 years?
3. Trial Design
Capecitabine
- Capecitabine was listed as the drug intervention.
- The trial title identifies capecitabine as the post-surgical treatment strategy being evaluated.
Clinical observation
- Clinical observation was listed as the comparator intervention.
- The trial title identifies observation after surgery as the comparison strategy.
4. Trial Timeline
Trial start
The registry records March 2006 as the study start date.
Primary completion
The registry records August 2013 as the primary completion date.
Study status
The ClinicalTrials.gov record lists the study status as completed.
5. Endpoints
| Endpoint | Registry definition | Statistical role |
|---|---|---|
| Survival at 2 years | Survival at 2 years | Registered primary endpoint |
The registry identifies survival at 2 years as the primary endpoint. The endpoint is expressed as a fixed follow-up time rather than as a median survival time or a hazard ratio in the registry information available here.
6. Statistical Methodology
The principal statistical structure of BILCAP follows directly from its randomized design and its registered 2-year survival endpoint. Randomization creates the basis for comparing survival outcomes between the assigned treatment strategies. The appropriate analysis depends on whether the underlying patient-level time-to-event information is available and on how the registry operationalized the 2-year endpoint.
Time-to-event framework
Survival is naturally a time-to-event outcome because the event of interest is death and patients may be followed for different lengths of time. When individual event times and censoring times are available, Kaplan-Meier estimation is a standard way to describe survival over time.
Here, di is the number of events at time ti, while ni is the number of participants at risk immediately before that time.
Fixed-time survival
For a primary endpoint defined specifically at 2 years, the treatment groups can be compared by estimating the proportion of participants alive at the 2-year landmark. A confidence interval can then quantify uncertainty around each estimated survival proportion, and an appropriate comparison can assess the difference between randomized groups.
Time-to-event alternative
If the analysis uses the full survival follow-up rather than only the 2-year landmark, a Kaplan-Meier curve can summarize the complete survival experience and a Cox proportional-hazards model can provide a hazard ratio. Such a hazard ratio would answer a different question from the fixed 2-year survival proportion.
A 2-year survival estimate describes the probability of remaining alive beyond the specified 2-year time point. It does not by itself summarize the relative event rate throughout the entire follow-up period.
Intention-to-treat principle
Because the trial is randomized, an intention-to-treat analysis would preserve the randomized comparison by analyzing participants according to their assigned treatment strategy. This protects the interpretability provided by randomization and avoids redefining treatment groups based on what happened after allocation.
7. Planned Analysis
The registry identifies survival at 2 years as the primary endpoint. For an endpoint of this type, the analysis would ordinarily estimate the proportion of participants alive at the 2-year landmark in each randomized group and quantify uncertainty around those estimates. If the underlying follow-up times are available, survival would typically also be characterized using Kaplan-Meier methods, allowing censoring before 2 years to be handled appropriately.
A formal comparison could be expressed as an absolute difference in 2-year survival probabilities, a ratio of survival probabilities, or through a time-to-event model if the prespecified analysis uses the complete survival experience. The choice should follow the statistical analysis plan and the precise endpoint definition used by the trial.
What the primary endpoint would quantify
Absolute survival probability
The percentage of randomized participants estimated to be alive at the 2-year landmark in each treatment group.
Between-group difference
The absolute difference in 2-year survival probabilities directly describes how far apart the two randomized groups are at that time point.
Confidence interval
A confidence interval would describe statistical uncertainty around the estimated survival probability or treatment contrast.
Time-to-event analysis
If full follow-up is analyzed, Kaplan-Meier and Cox methods can use information from the complete survival history rather than only the 2-year landmark.
8. Statistical Methods Explained
Why is survival a time-to-event endpoint?
Survival incorporates both whether an event occurred and when it occurred. Two participants can have different follow-up experiences even if neither has died by a particular calendar date. Time-to-event methods preserve that timing information and can accommodate right censoring.
What does “survival at 2 years” mean?
It refers to the probability of being alive at the 2-year time point. In mathematical terms, if T represents time to death, the quantity is P(T > 2 years). It is a landmark measure rather than a median survival estimate.
Why might Kaplan-Meier estimation be useful?
Kaplan-Meier estimation uses the observed event times and censoring information to estimate survival over time. It allows participants who have not experienced the event by the end of their available follow-up to contribute information without treating them as if the event occurred.
What would a confidence interval tell us?
A confidence interval around a 2-year survival estimate describes uncertainty arising from sampling variability under the statistical model and analysis framework. A narrower interval indicates greater statistical precision; a wider interval indicates greater uncertainty.
What would a hazard ratio mean if one were reported?
A hazard ratio compares estimated instantaneous event rates between groups over the analyzed follow-up under a proportional-hazards model. A hazard ratio below 1 would indicate a lower estimated instantaneous event rate in the numerator treatment group. It would not be equivalent to the difference in 2-year survival probabilities.
Why should the 2-year endpoint and a hazard ratio not be conflated?
The 2-year survival probability describes one clinically defined time point. A hazard ratio is a relative time-to-event measure that summarizes the modeled event-rate relationship over follow-up. They can provide complementary information, but they answer different statistical questions.
9. Randomization and Causal Interpretation
Randomization is central to the causal interpretation of BILCAP. When participants are randomly assigned to capecitabine or clinical observation, the groups are created by an allocation mechanism rather than by investigators selecting treatment according to prognosis. This allows differences in subsequent outcomes to be interpreted as treatment-group differences within the randomized study framework.
| Design feature | Registry information | Statistical implication |
|---|---|---|
| Allocation | Randomized | Supports comparison of outcomes according to assigned treatment. |
| Phase | Phase 3 | Represents a confirmatory-stage clinical trial design. |
| Primary purpose | Treatment | The study is structured to evaluate a treatment strategy. |
| Enrollment | 360 | Defines the overall randomized study population size. |
Randomization does not guarantee that every baseline characteristic will be numerically identical between groups. Its principal statistical value is that, in expectation, prognostic factors are balanced by the allocation process rather than being selected systematically into one treatment group.
10. Interpretation of a Two-Year Survival Estimate
A 2-year survival estimate represents the estimated proportion of the randomized population expected to remain alive at the specified 2-year landmark.
It does not mean that every participant has the same individual probability of survival, nor does it describe the median survival time. It also does not describe what happens after the 2-year landmark.
An absolute survival probability is directly interpretable at the selected time point. If a treatment contrast is reported, the absolute difference can be easier to understand clinically than a relative time-to-event measure alone.
A fixed-time survival estimate depends on the availability and quality of follow-up through the landmark. Participants who have not experienced the event by their last assessment require appropriate handling of censoring rather than being treated as deaths or as guaranteed survivors.
11. Statistical Interpretation vs Clinical Interpretation
Statistical interpretation
The trial uses randomized allocation and identifies survival at 2 years as its primary endpoint. The appropriate analysis should preserve the randomized comparison and account appropriately for time-to-event follow-up and censoring.
Clinical interpretation
The clinical meaning of the study depends on how survival at the 2-year landmark differs between the randomized treatment strategies and how that difference relates to the broader treatment context.
A statistically estimated difference in survival is not, by itself, a complete assessment of a treatment strategy. Interpretation also depends on the endpoint definition, follow-up, uncertainty, treatment exposure, and the population represented by the randomized trial.
12. Limitations
- Limited posted statistical detail: The ClinicalTrials.gov record identifies survival at 2 years as the primary endpoint but does not provide posted statistical analyses in the registry information considered here.
- Endpoint specification: The registry wording is simply “Survival at 2 years.” Additional operational details about censoring, estimands, statistical tests, and confidence-interval construction are not reported in the available registry data.
- No posted effect estimate: The registry information does not report a 2-year survival estimate, treatment contrast, confidence interval, or p-value.
- No arm-level enrollment counts: The available record identifies total enrollment of 360 but does not provide treatment-group sample sizes here.
- No baseline characteristics: Baseline demographic and disease characteristics are not included in the available trial data.
- No subgroup results: The available record does not provide subgroup estimates or interaction analyses.
- No safety results: Serious adverse events by treatment arm are not reported in the available registry information.
- No additional endpoint results: The registered primary endpoint is survival at 2 years, and no posted statistical analyses are listed.
- Time-to-event assumptions: If a Cox proportional-hazards analysis is used, its interpretation depends on the proportional-hazards assumption and the prespecified analysis framework.
13. What Would Be Needed for a Full Survival Analysis?
A complete statistical analysis of the 2-year survival endpoint would require more than the total enrollment count and endpoint label. At minimum, the analysis would benefit from the number of participants randomized to each treatment strategy, individual follow-up or sufficient aggregate survival information, deaths and censoring information, and the prespecified definition of the analysis population.
| Information | Why it matters |
|---|---|
| Randomized sample size by arm | Defines the denominator for each treatment-group survival estimate. |
| Death times | Provides the event information needed for time-to-event estimation. |
| Censoring times | Determines how participants contribute follow-up without an observed death. |
| 2-year survival estimate | Directly answers the registered primary endpoint. |
| Confidence interval | Quantifies uncertainty around the estimated treatment effect or survival probability. |
| Prespecified analysis method | Defines how the primary hypothesis is tested and how uncertainty is calculated. |
These elements distinguish a reproducible statistical analysis from a qualitative description of the study design. The registry provides the trial identity, design classification, enrollment, interventions, and registered primary endpoint, while the statistical analysis plan determines the precise inferential framework.
14. Why This Trial Matters Statistically
BILCAP is a useful teaching example because it illustrates an important distinction in clinical-trial statistics: a trial can be randomized and phase 3 while its primary endpoint is expressed as a specific fixed-time survival measure rather than as a median survival or hazard ratio.
| Concept | How it appears in BILCAP |
|---|---|
| Randomization | The ClinicalTrials.gov record classifies the allocation as randomized. |
| Phase 3 design | The study is registered as phase 3. |
| Time-to-event outcome | The primary endpoint is survival at 2 years. |
| Landmark survival | The primary endpoint focuses on a specified 2-year time point. |
| Kaplan-Meier estimation | A standard method for estimating survival over time when individual follow-up and censoring information are available. |
| Censoring | Relevant to survival analysis when participants have incomplete event follow-up. |
| Confidence intervals | Needed to quantify uncertainty around survival estimates or treatment contrasts. |
| Hazard ratios | A possible complementary time-to-event measure, but not interchangeable with 2-year survival. |
| Intention-to-treat analysis | A natural framework for preserving the randomized comparison in efficacy analysis. |
15. A Practical Statistical Reading of the Registry
The most important statistical information in the registry is the combination of four elements: randomized allocation, phase 3 status, enrollment of 360, and survival at 2 years as the primary endpoint.
That combination tells the reader what the study is designed to compare, but it does not establish the magnitude or statistical significance of the treatment effect. The endpoint label identifies the target quantity; the observed data and prespecified statistical analysis are required to estimate that quantity and make an inferential comparison.
The primary statistical question
The relevant comparison is the survival probability at the 2-year landmark between the randomized capecitabine and clinical-observation strategies.
This distinction is particularly important when reading clinical-trial registries. A registry entry can establish that an endpoint was prospectively registered without providing the numerical result needed to evaluate the endpoint.
16. Sources
- ClinicalTrials.gov: BILCAP, NCT00363584.
- PubMed: PubMed record associated with the BILCAP evidence base.
Continue through Clinical Biostats
Explore additional clinical trial analyses, statistical tutorials, and statistical calculators for understanding randomized clinical-trial methods.
17. Record Summary
BILCAP is a randomized phase 3 treatment trial with 360 enrolled participants evaluating capecitabine or clinical observation after surgery in patients with biliary tract cancer. The ClinicalTrials.gov record identifies survival at 2 years as the primary endpoint. From a statistical perspective, the key analytical distinction is between a fixed-time survival probability at 2 years and broader time-to-event measures such as median survival or a hazard ratio. A rigorous analysis would preserve the randomized comparison, account appropriately for censoring, estimate survival at the specified landmark, and quantify uncertainty around the treatment contrast.