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
READY was a randomized, double-blind, phase 3, parallel-group trial evaluating docetaxel plus dasatinib against docetaxel plus placebo in participants with prostatic neoplasms. The registry identifies overall survival as the single primary endpoint and reports a formal stratified log-rank analysis with a hazard ratio.
| Feature | READY |
|---|---|
| Phase | Phase 3 |
| Condition | Prostatic Neoplasms |
| Design | Randomized, double-blind, parallel-group |
| Allocation | Randomized |
| Primary purpose | Treatment |
| Enrollment | 1930 |
| Primary endpoint | Overall Survival: Time From Randomization to Date of Death |
| Primary endpoint type | Time-to-event |
| Primary analysis | Stratified log-rank test with hazard ratio |
| Study period | 2008-10 to 2012-08 |
| Lead sponsor | Bristol-Myers Squibb |
| Sponsor type | Industry |
| ClinicalTrials.gov | NCT00744497 |
2. Clinical Question
The central question was whether adding dasatinib to docetaxel improved overall survival compared with docetaxel plus placebo in participants with prostatic neoplasms enrolled in this phase 3 randomized treatment trial.
Population
Participants with Prostatic Neoplasms enrolled in the randomized phase 3 READY study.
Intervention
Dasatinib plus docetaxel. Prednisone is also listed among the study interventions.
Comparator
Placebo plus docetaxel. Prednisone is also listed among the study interventions.
Primary question
Does the dasatinib-containing regimen change time from randomization to death compared with the placebo-containing regimen?
3. Trial Design
Docetaxel + placebo
- Placebo is listed as a study drug.
- Docetaxel is listed as a study intervention.
- The registry compares this group with the dasatinib group for the primary overall-survival analysis.
Docetaxel + dasatinib
- Dasatinib is listed as a study drug.
- Docetaxel is listed as a study intervention.
- The registry presents hazard ratios for the experimental group relative to the control group for the posted secondary analyses.
4. Primary Endpoint
| Endpoint | Registry definition | Analysis |
|---|---|---|
| Overall Survival: Time From Randomization to Date of Death | Overall survival is defined as time in months from the randomization date to the date of death due to any cause in the randomized population. If the patient did not die, survival was censored on the last date he or she was known to be alive. | From randomization to death or date of last contact; maximum reached: 45 months. Stratified log-rank test with hazard ratio. |
The endpoint is a classic time-to-event outcome. Every randomized participant contributes follow-up from randomization until either death or censoring at the last date known to be alive. This makes the analysis fundamentally different from a simple comparison of proportions at a single time point.
5. Statistical Methodology
Primary analysis population
The primary analysis included all participants who were randomized to receive any treatment. This anchors the efficacy comparison to randomized treatment assignment rather than restricting the primary analysis to participants who completed treatment.
Stratified log-rank test
The registry reports a two-sided log-rank comparison of survival between the placebo and dasatinib groups. The analysis was stratified by bisphosphonate intake at randomization (yes/no) and urinary N-telopeptide category (<60 versus ≥60 nmol/mmol creatinine) defined at randomization.
The reported analysis tested the null hypothesis that survival was equal between the two randomized treatment groups. The registry reports a two-sided alpha level of 0.0447 for the primary log-rank comparison.
Hazard ratio
The primary effect measure was a hazard ratio. The registry reports an estimate of 0.99 with a two-sided 95.53% confidence interval of 0.87 to 1.13.
A hazard ratio near 1 indicates that the estimated instantaneous event rates are similar under the fitted time-to-event comparison. It is a relative measure over follow-up, not a percentage of patients who benefited and not an absolute difference in survival probability.
Confidence intervals for median overall survival
The registry states that confidence intervals for median overall survival were calculated using the Brookmeyer and Crowley method. This method derives uncertainty around a Kaplan-Meier median from the estimated survival curve rather than treating the median as an ordinary sample mean.
Power and event target
The registry states that the power calculations were based on at least 858 deaths producing at least 90% power at the 5% level for rejecting the null hypothesis when the true hazard ratio was assumed to be 0.8.
6. Primary Result: Overall Survival
Hazard ratio for death
95.53% CI: 0.87–1.13 · P = 0.9009
Stratified two-sided log-rank analysis; analysis population: all randomized participants.
| Primary endpoint | Effect measure | Estimate | Confidence interval | P-value |
|---|---|---|---|---|
| Overall Survival: Time From Randomization to Date of Death | Hazard ratio | 0.99 | 95.53% CI 0.87–1.13 | 0.9009 |
The reported hazard ratio of 0.99 is very close to 1. Under the reported survival model, this corresponds to an estimated instantaneous rate of death in the dasatinib group that is approximately 1% lower than the corresponding rate in the placebo group. That is a relative hazard interpretation, not a statement that 1% of patients benefited or that individual patients experienced a 1% reduction in their probability of death.
The 95.53% confidence interval of 0.87 to 1.13 describes uncertainty around the estimated hazard ratio under the analysis framework. It spans 1, so the interval includes hazard ratios corresponding to somewhat lower and somewhat higher hazards in the dasatinib group. The width of the interval also shows that the estimate is not an exact measurement of the underlying treatment effect.
The P-value of 0.9009 measures the compatibility of the observed data with the specified null hypothesis under the statistical testing framework. It does not measure the size of the treatment effect, the probability that the null hypothesis is true, or the probability that the treatment has no clinical effect.
The registry reports a stratified log-rank test with an alpha level of 0.0447. Because the reported P-value is substantially larger than that prespecified significance level, the primary superiority hypothesis was not demonstrated in this analysis. The confidence interval provides additional information beyond the P-value by showing the range of hazard-ratio values compatible with the analysis.
7. Secondary Endpoint Results
The registry contains six posted secondary analyses in addition to the primary overall-survival analysis. The registry states that because superiority for overall survival was not demonstrated, the secondary endpoints were not tested. The reported odds ratios and hazard ratios are therefore descriptive effect estimates rather than a sequence of formally tested confirmatory hypotheses.
Objective Tumor Response
Odds ratio for objective tumor response
95% CI: 0.688–1.271
Participants with at least 1 target lesion at baseline; odds ratio presented for experimental to control group.
An odds ratio of 0.935 compares the odds of objective tumor response between the dasatinib and placebo groups in the registry's response population. An odds ratio of 1 would indicate equal odds of response; 0.935 is close to that reference value.
The 95% CI of 0.688 to 1.271 spans 1. This means the interval is compatible with lower or higher odds of response in the dasatinib group relative to placebo. Because the registry states that secondary endpoints were not tested after the primary superiority endpoint was not demonstrated, the estimate should not be interpreted as an independently confirmed superiority finding.
Time to First Skeletal-related Event
Hazard ratio for time to first SRE
95% CI: 0.64–1.02
From randomization to first SRE or last SRE assessment; maximum reached: 42 months.
The hazard ratio of 0.81 corresponds to an estimated instantaneous SRE rate about 19% lower in the dasatinib group relative to placebo under the reported hazard-ratio interpretation. This is not equivalent to a 19% absolute reduction in the percentage of participants experiencing an SRE.
The 95% CI of 0.64 to 1.02 is relatively informative about the range of plausible relative effects but includes 1. Because this was a secondary endpoint that was not formally tested after the primary endpoint result, the estimate is best regarded as descriptive.
Reduction in Urinary N-telopeptide
Odds ratio for reduction in uNTx
95% CI: 0.930–1.763
Analysis population: participants with baseline urinary N-telopeptide above the specified threshold.
An odds ratio of 1.280 indicates that the estimated odds of the specified urinary N-telopeptide reduction were 28% higher in the dasatinib group than in the placebo group, using the odds ratio presented by the registry.
The 95% CI of 0.930 to 1.763 includes 1. The confidence interval therefore does not establish a clear direction of the underlying comparative odds from this analysis alone. The registry also states that this secondary endpoint was not formally tested following the primary overall-survival result.
Progression-free Survival
Hazard ratio for progression or death
95% CI: 0.82–1.05
From randomization to disease progression or death; maximum reached: approximately 43 months.
The reported PFS hazard ratio of 0.92 corresponds to an estimated instantaneous rate of progression or death about 8% lower in the dasatinib group under the reported hazard-ratio interpretation.
The 95% CI of 0.82 to 1.05 crosses 1, so the estimate is compatible with a range of relative effects that includes little or no difference. Again, the registry identifies this as a secondary endpoint that was not formally tested after the primary overall-survival superiority hypothesis was not demonstrated.
Time to PSA Progression
Hazard ratio for PSA progression
95% CI: 0.79–1.01
From randomization to confirmed PSA progression; maximum reached: 30 months.
The hazard ratio of 0.89 corresponds to an estimated instantaneous rate of PSA progression about 11% lower in the dasatinib group relative to placebo, under the reported hazard-ratio interpretation.
The 95% CI of 0.79 to 1.01 is close to the null value and includes 1. The estimate therefore does not by itself establish a statistically confirmed difference, particularly because the registry states that secondary endpoints were not tested after the primary overall-survival result.
Reduction in Pain Intensity
Odds ratio for reduction in pain intensity
95% CI: 0.594–1.052
Analysis population: participants with baseline pain intensity of 2 or greater.
An odds ratio of 0.791 indicates that the estimated odds of the specified pain-intensity reduction were approximately 21% lower in the dasatinib group relative to placebo, using the odds ratio presented for the experimental-to-control comparison.
The 95% CI of 0.594 to 1.052 includes 1 and is fairly wide. It therefore encompasses both a potentially lower and a potentially similar or higher odds of response. The registry explicitly states that this secondary endpoint was not formally tested following the overall-survival analysis.
| Secondary endpoint | Effect measure | Estimate | 95% CI | Analysis population |
|---|---|---|---|---|
| Objective tumor response | Odds ratio | 0.935 | 0.688–1.271 | Participants with at least 1 target lesion at baseline |
| Time to first skeletal-related event | Hazard ratio | 0.81 | 0.64–1.02 | All randomized participants |
| Reduction in urinary N-telopeptide | Odds ratio | 1.280 | 0.930–1.763 | Participants above the specified baseline uNTx threshold |
| Progression-free survival | Hazard ratio | 0.92 | 0.82–1.05 | All randomized participants |
| Time to PSA progression | Hazard ratio | 0.89 | 0.79–1.01 | All randomized participants |
| Reduction in pain intensity | Odds ratio | 0.791 | 0.594–1.052 | Participants with baseline pain intensity ≥2 |
8. Statistical Methods Explained
Why was a log-rank test used for overall survival?
Overall survival records the time until death, with patients who remain alive at the end of observation treated as censored observations. A log-rank test is designed to compare survival distributions between randomized groups while using information from the entire follow-up period rather than reducing the outcome to a single time point.
Why was the log-rank test stratified?
The registry states that the analysis was stratified by bisphosphonate intake and urinary N-telopeptide category at randomization. Stratification allows the survival comparison to account for these prespecified baseline categories when comparing the treatment groups.
What does an HR of 0.99 mean?
An HR of 0.99 means that the estimated instantaneous rate of death in the dasatinib group was approximately 1% lower than in the placebo group under the reported hazard-ratio model. It does not mean that the absolute probability of death was 1% lower, nor does it mean that every participant experienced a 1% reduction in risk.
Why is the confidence interval important?
The confidence interval communicates uncertainty around the estimated effect. For the primary HR, the 95.53% interval extends from 0.87 to 1.13. This shows that the point estimate of 0.99 should not be treated as an exact value for the underlying treatment effect.
Why does the P-value not measure treatment-effect size?
The P-value describes how compatible the observed evidence is with the specified null hypothesis under the testing framework. It depends on both the observed effect and the amount of statistical information available. A P-value is therefore not a substitute for the hazard ratio or its confidence interval.
Why does the analysis use an event target?
In survival trials, the number of deaths often determines statistical information more directly than the number randomized. The registry's design calculation specified that at least 858 deaths would provide at least 90% power at the 5% level under a true hazard ratio of 0.8.
Why should the secondary results be interpreted cautiously?
The registry states that because superiority of dasatinib was not demonstrated for overall survival, the secondary endpoints were not tested. Their effect estimates and confidence intervals remain useful for understanding the observed data, but they should not be treated as independently confirmed superiority findings.
9. Odds Ratios vs Hazard Ratios
READY illustrates two distinct effect-measure families. Time-to-event endpoints were summarized using hazard ratios, while several binary outcomes were summarized using odds ratios. These quantities answer different statistical questions.
| Measure | Used for | Interpretation |
|---|---|---|
| Hazard ratio | Overall survival, skeletal-related event, progression-free survival, PSA progression | Relative comparison of instantaneous event rates over follow-up. |
| Odds ratio | Objective tumor response, urinary N-telopeptide reduction, pain-intensity reduction | Ratio of the odds of the specified binary outcome between treatment groups. |
An OR of 1 represents equal odds. Values above 1 indicate higher odds in the dasatinib group, while values below 1 indicate lower odds, subject to the exact outcome definition.
Odds ratios should not be read as risk ratios. For example, an odds ratio of 1.280 does not mean that 28% more participants necessarily experienced the event. The difference is particularly important when the underlying outcome is common.
10. Time-to-Event Endpoints and Censoring
Four of the posted analyses are time-to-event outcomes: overall survival, time to first skeletal-related event, progression-free survival, and time to PSA progression. Each has its own event definition and censoring rule.
| Endpoint | Event | Censoring / follow-up rule |
|---|---|---|
| Overall survival | Death due to any cause | If the patient did not die, censored at the last date known to be alive; maximum 45 months. |
| Time to first SRE | First skeletal-related event | Last SRE assessment if subsequent cancer therapy began or no SRE; maximum 42 months. |
| Progression-free survival | Disease progression or death | Last clinical assessment if subsequent cancer therapy started or no progression/death; maximum approximately 43 months. |
| Time to PSA progression | Confirmed PSA progression | Last bone scan assessment if no progression or if cancer therapy started; maximum 30 months. |
Censoring is not simply missing data. A censored participant contributes information about being event-free up to the censoring point. The validity of standard survival methods also depends on assumptions about the relationship between censoring and the event process.
11. Statistical Interpretation of the Primary Result
The primary HR of 0.99 indicates an estimated instantaneous death rate in the dasatinib group that is approximately 1% lower than in the placebo group under the reported model. Because the estimate is close to 1, the point estimate itself represents little relative separation between the groups.
The 95.53% CI of 0.87–1.13 gives a range around the point estimate. It includes the null value of 1 and therefore does not establish a statistically distinguishable difference under the prespecified superiority test.
The reported P = 0.9009 is much larger than the registry-reported two-sided alpha level of 0.0447. Thus the primary superiority hypothesis was not demonstrated. The P-value should not be interpreted as the probability that dasatinib has no effect.
The registry data posted on ClinicalTrials.gov for the primary analysis do not report numerical median overall survival estimates, survival probabilities at specified time points, or an absolute survival difference. The primary evidence therefore centers on the reported hazard ratio, confidence interval, and log-rank P-value.
12. Safety
The registry reports serious adverse events by randomized treatment arm. The counts are presented as affected participants divided by participants at risk.
| Arm | Participants with serious adverse events | Participants at risk |
|---|---|---|
| Placebo | 317 | 757 |
| Dasatinib | 381 | 761 |
Placebo arm
317/757 participants were affected by serious adverse events.
Dasatinib arm
381/761 participants were affected by serious adverse events.
These are safety counts rather than a time-to-event efficacy analysis. The denominator is the number at risk reported for each arm, and the registry does not provide a formal comparative statistical test for these serious-adverse-event counts in the information summarized here.
13. Primary Analysis vs Secondary Analyses
The READY statistical structure illustrates an important distinction between an estimand being reported and a hypothesis being formally tested.
| Analysis | Role | Statistical interpretation |
|---|---|---|
| Overall survival | Primary endpoint | Formal stratified two-sided log-rank test with HR; superiority hypothesis. |
| Objective tumor response | Secondary | OR 0.935 with 95% CI 0.688–1.271; not formally tested. |
| Time to first SRE | Secondary | HR 0.81 with 95% CI 0.64–1.02; not formally tested. |
| uNTx reduction | Secondary | OR 1.280 with 95% CI 0.930–1.763; not formally tested. |
| PFS | Secondary | HR 0.92 with 95% CI 0.82–1.05; not formally tested. |
| PSA progression | Secondary | HR 0.89 with 95% CI 0.79–1.01; not formally tested. |
| Pain-intensity reduction | Secondary | OR 0.791 with 95% CI 0.594–1.052; not formally tested. |
This hierarchy matters because the secondary confidence intervals are descriptive in the context described by the registry. They should not be interpreted as though each endpoint independently passed a prespecified confirmatory testing sequence.
14. What the Registry Reports About Stratification
The primary survival analysis was stratified using two baseline variables defined at randomization:
- Bisphosphonate intake: yes versus no.
- Urinary N-telopeptide category: <60 versus ≥60 nmol/mmol creatinine.
Stratification can improve the alignment between the analysis and the randomized trial design when important prognostic or design-related variables have been specified in advance. Rather than ignoring the strata, the analysis compares survival while accounting for their structure.
A stratified analysis does not mean that the treatment effect is estimated separately and independently within every stratum. Instead, the strata are incorporated into the comparison so that the overall treatment effect reflects the prespecified stratified design.
15. Censoring and Follow-up
The primary endpoint could be followed for a maximum of 45 months according to the registry. Participants who had not died were censored at the last date they were known to be alive.
This structure is important because not every participant necessarily has the same amount of follow-up. Survival analysis allows participants with shorter observation periods to contribute information without requiring the analyst to discard them simply because the event was not observed.
The secondary time-to-event endpoints used different maximum follow-up periods: 42 months for time to first skeletal-related event, approximately 43 months for progression-free survival, and 30 months for time to PSA progression.
16. Statistical Methods Explained in Practice
Kaplan-Meier estimation
A Kaplan-Meier curve estimates the probability of remaining event-free over time while incorporating censored observations. It is the natural descriptive companion to a time-to-event endpoint such as overall survival.
Log-rank comparison
The log-rank test compares the survival experience of randomized groups over follow-up. READY used a stratified version that incorporated the prespecified baseline categories.
Hazard ratio
The HR summarizes the relative instantaneous event rate between groups. An HR of 0.99 is close to the null value of 1 and should not be translated into an absolute survival difference.
Odds ratio
The OR compares odds for binary outcomes. It is appropriate to distinguish it from the hazard ratio because an OR does not incorporate the timing of an event.
17. Limitations
- Primary endpoint scope: the registry identifies overall survival as the single primary endpoint; the remaining posted efficacy analyses are secondary.
- Secondary endpoint testing: the registry states that secondary endpoints were not tested because superiority for overall survival was not demonstrated.
- Limited reported summary data: the registry information summarized here does not report numerical median survival estimates or survival probabilities at selected time points.
- Hazard-ratio interpretation: a single HR is a relative time-to-event measure and should not automatically be interpreted as a constant relative difference in survival probability throughout follow-up.
- Censoring: valid survival inference depends on assumptions concerning the censoring process and the information contributed by censored participants.
- Analysis populations: the primary efficacy analysis used all randomized participants, whereas some secondary endpoints used restricted populations defined by baseline measurements or lesions.
- Multiple secondary outcomes: six secondary effect estimates are reported, but the registry states that they were not tested after the primary endpoint result.
- Safety interpretation: serious adverse-event counts describe affected participants and participants at risk, but the registry information summarized here does not provide a formal comparative safety test.
18. Why This Trial Matters Statistically
READY is a useful teaching example because it combines a large randomized phase 3 design with a classic time-to-event primary endpoint and several secondary endpoints using different effect measures.
| Concept | How it appears in READY |
|---|---|
| Randomization | 1930 participants were enrolled in a randomized parallel-group phase 3 trial. |
| Blinding | The study was double-blind. |
| Time-to-event analysis | Overall survival was measured from randomization to death or censoring. |
| Stratified analysis | The primary survival comparison was stratified by bisphosphonate intake and urinary N-telopeptide category. |
| Log-rank test | The primary endpoint used a stratified log-rank comparison. |
| Hazard ratio | The primary OS effect was reported as HR 0.99. |
| Confidence interval | The primary HR had a 95.53% CI of 0.87–1.13. |
| P-value | The primary analysis reported P = 0.9009. |
| Odds ratio | Binary secondary endpoints were summarized using odds ratios. |
| Event-driven power | The registry based power calculations on at least 858 deaths. |
| Multiple endpoints | Secondary endpoints were not tested after the primary superiority result was not demonstrated. |
19. Related Tutorials
Learn more about the methods used in this trial:
20. Related Calculators
21. Sources
- ClinicalTrials.gov: NCT00744497.
- Linked publication: PubMed record for PMID 24722180.
- Linked publication: PubMed record for PMID 24211163.
Continue through Clinical Biostats
Connect the statistical methods used in READY with deeper tutorials and practical statistical tools.
22. Record Summary
READY provides a clear example of how a randomized phase 3 oncology trial can combine time-to-event methodology with binary secondary outcomes. The primary endpoint was overall survival, defined from randomization to death or censoring at the last date known to be alive. The registry reports a stratified log-rank analysis, a hazard ratio of 0.99, a 95.53% confidence interval of 0.87–1.13, and a P-value of 0.9009.
The secondary analyses illustrate the distinction between hazard ratios and odds ratios. Time-to-event outcomes included time to first skeletal-related event, progression-free survival, and time to PSA progression, while objective tumor response, urinary N-telopeptide reduction, and pain-intensity reduction were summarized using odds ratios. Across these analyses, the registry explicitly states that secondary endpoints were not tested after superiority for overall survival was not demonstrated.
From a statistical perspective, the central lesson is that the treatment estimate, confidence interval, P-value, endpoint definition, analysis population, stratification, censoring rules, and testing hierarchy must be interpreted together. A hazard ratio is not an absolute risk difference, a confidence interval is not a range of individual treatment effects, and a P-value is not a measure of effect size.