Real-World Effectiveness of Palbociclib in Combination With an Aromatase Inhibitor: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Real-World Effectiveness of Palbociclib in Combination With an Aromatase Inhibitor (NCT05361655), a registered phase not stated study. The primary outcome is Overall Survival (OS) in Postmenopausal Female or Male Participants With Metastatic Breast Cancer.

2. Clinical Question

The comparison concerns real-world treatment groups, not groups assigned by a stated randomization procedure. The registry describes the study as retrospective and says that participant characteristics were balanced using stabilized inverse probability of treatment weighting. This approach is intended to make the groups more comparable on measured characteristics, but the record does not provide the detailed covariate list or enough information to assess how completely the groups were balanced.

The registry lists 2888 enrolled participants. It does not report the number of participants in each treatment group or the weighted number analyzed for overall survival. The primary outcome is overall survival in months, with a median and 95% confidence interval as the stated summary. The record describes follow-up from the index date to death or censoring at 30-Sep-2020, for approximately up to 68 months.

The central question is therefore a real-world association between initial treatment group and survival, conditional on the study’s weighting and time-to-event methods. The registry record does not report the estimated medians, confidence intervals, or hazard ratios needed to describe the direction or precision of that comparison.

3. Trial Design

This is a completed retrospective study using de-identified patient information from the Flatiron Health Analytic Database. The brief description identifies the intended comparison as first-line palbociclib plus an aromatase inhibitor versus an aromatase inhibitor alone among postmenopausal women or men with HR+/HER2- metastatic breast cancer in United States clinical practices. The intervention descriptions identify both groups as adults with metastatic breast cancer who initiated their respective treatment as first-line therapy.

The treatment-initiation window in the arm descriptions runs from Feb 3, 2015 to March 31, 2020. The registry lists the study start and primary completion as 2021-09-01 and the status as completed. These registry dates are distinct from the treatment-initiation window described for the database cohorts.

The record lists enrollment of 2888. It does not report the number initially included in each treatment group, the group-specific number retained for analysis, or the number contributing to each outcome. It states that stabilized inverse probability of treatment weighting was used to balance participant characteristics and create balanced virtual groups. As a result, the overall number analyzed after weighting differed from the number of participants enrolled.

This was not described as a randomized comparison; allocation is not stated in the registry record. Treatment choice in a retrospective database study may be associated with patient characteristics, so weighting can address measured differences but does not by itself establish that all relevant differences were controlled. Details of eligibility criteria beyond the described population, the covariates used in weighting, and checks of balance are not reported in the registry record.

The registry lists 10 outcome measures and 7 analyses. The listed secondary outcomes include real-world progression-free survival, tumor response counts and response rate, initial palbociclib dose, time to dose adjustment, duration of treatment, time to next line of treatment, and time to subsequent chemotherapy. The record does not give the analysis-specific results for those outcomes.

4. Participant Flow and Accounting

The registry reports how participants moved through each stage of the study. All counts below are exactly as posted; missing counts are not inferred from other stages.

Before assignment: All participants who were eligible for the retrospective database analysis were included for the study (i.e., 2888).

Overall Study

Analysis population for the primary outcome, as defined in the registry: FAS population comprised of all eligible participants enrolled in the study. Analysis was performed using sIPTW method to balance participant characteristics among reporting groups. sIPTW method created balanced virtual groups and hence overall number of participants analyzed was different from numbers in participant flow (not balanced).

5. Baseline Characteristics

The table reproduces the baseline characteristics posted in the registry, which describe the participants analysed at baseline and help show whether the randomized groups were comparable.

Population: Full analysis set (FAS) population comprised of all eligible participants enrolled in the study.

6. Primary Endpoint

The registry lists the primary outcome as Overall Survival (OS) in Postmenopausal Female or Male Participants With Metastatic Breast Cancer, measured over From index date to death due to any cause or censoring date of 30-Sep-2020 (approximately up to 68 months).

The primary endpoint is overall survival (OS), defined as time from the index date—the start of palbociclib plus an aromatase inhibitor or aromatase inhibitor alone—to death from any cause. Participants who had not died by the end of the study were censored at 30-Sep-2020. The registry describes follow-up as approximately up to 68 months.

Overall survival is a time-to-event outcome: it incorporates both whether an event occurred and how long participants were followed before an event or censoring. The registry specifies the median OS in months, accompanied by a 95% confidence interval, as the primary summary. In general, a median time-to-event estimate describes the time by which the estimated event-free proportion has reached one half; a confidence interval conveys uncertainty around the estimate rather than a guarantee that an individual participant’s outcome lies within that range.

The record says that OS was estimated with the Kaplan-Meier method and adjusted using stabilized inverse probability of treatment weighting. It does not report the group-specific median OS estimates or their confidence intervals. Nor does it provide numerical hazard-ratio results in the information recorded for this outcome. Therefore, the endpoint definition and analysis approach can be described, but the size, direction, and precision of the primary comparison cannot be stated from the registry record.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The methods described for the primary outcome combine Kaplan-Meier estimation with stabilized inverse probability of treatment weighting. These methods address different aspects of the analysis: Kaplan-Meier estimation summarizes event timing in the presence of censoring, while weighting is intended to improve comparability of the treatment groups on measured characteristics.

How does Kaplan-Meier estimation work?

In general, the Kaplan-Meier method estimates the probability of remaining free of an event over time. It updates the estimate as events occur, using the participants still under observation at each event time. Participants who have not had the event by the end of their observed follow-up can contribute information up to their censoring time without being counted as if they had an event. This is useful when participants have different follow-up durations.

For this study, the event is death from any cause, time begins at the index date, and the stated censoring date is 30-Sep-2020. The primary summary is the median in months with a 95% confidence interval. The registry record does not report those estimates. It also does not give the number at risk at successive times or the amount of censoring in each group, so the stability of the survival curves cannot be evaluated from the record.

What does weighting contribute?

In general, inverse probability of treatment weighting assigns different weights to participants according to their probability of receiving the treatment group they entered, conditional on measured characteristics. The goal is to create weighted groups that are more alike on those measured characteristics than the original groups. Stabilized weights modify the weighting scheme to improve its behavior, but their use does not ensure that all confounding has been removed.

The registry says that stabilized inverse probability of treatment weighting was used to balance participant characteristics and create balanced virtual groups. It also notes that the overall number analyzed differed from the number enrolled. The covariates used to calculate weights, the treatment-probability model, weight distribution, handling of extreme weights, and balance results are not reported in the registry record. Without these details, readers cannot assess whether the intended balance was achieved or how sensitive the estimates might be to the weighting choices.

Weighting can address only characteristics represented in the available data and adequately handled by the model. In general, unmeasured or poorly measured differences between groups can remain. That limitation matters in a nonrandomized treatment comparison because treatment selection may be related to patient or clinical factors. The record does not describe allocation as randomized and does not state a randomization procedure.

What does the confidence interval describe?

In general, a confidence interval expresses statistical uncertainty about an estimated quantity under the assumptions of the analysis. A narrower interval indicates greater precision than a wider interval, but the interval does not capture every source of uncertainty, including potential bias from unmeasured confounding or errors in data collection. The registry specifies a 95% confidence interval for median OS, but the interval endpoints are not reported in the registry record.

How should time-to-event assumptions be considered?

In general, time-to-event methods rely on assumptions about how censoring relates to the event process. Kaplan-Meier estimates are most readily interpreted when censoring is not systematically related to future event risk in a way that distorts the observed follow-up. The record states that participants who did not die were censored at the end-of-study date. It does not provide additional censoring diagnostics or information on loss to follow-up, so those aspects cannot be assessed.

The record’s primary endpoint is median OS, not a mean survival time or a fixed-time survival probability. A median can be difficult to estimate precisely if the observed event information is limited around the point where the estimated survival curve crosses one half. Whether that occurred in either group cannot be determined because the estimates and curves are not reported in the registry record.

What can be said about hazard ratios?

A hazard ratio, in general, compares the instantaneous event rates between groups over time, conditional on the model used. It is not the same quantity as a difference between median survival times. The registry’s statistical description identifies median OS and its confidence interval, while the study information does not include numerical sponsor-reported hazard ratios. No hazard-ratio estimate or interpretation can therefore be given. The record also does not state whether a proportional-hazards model was used for the primary analysis.

The sample-size description refers to planning illustrations based on an observed effect, with a two-sided significance level of 0.05, equal allocation, and a normal approximation. It states that time-to-event illustrations concern total events. The numerical planning illustrations are not included in the record information, and they are not evidence of observed power. No sample-size justification or design calculation is reported as the study’s own planning rationale.

Overall, the methods description establishes the broad analytic framework but leaves important implementation details unspecified. Without the numerical survival results, analysis counts, weighting diagnostics, and fuller model information, readers can understand the intended approach but cannot reproduce or fully assess the primary comparison.

10. Multiplicity, Missing Data and Analysis Populations

The registry lists 10 outcome measures and 7 analyses. Overall survival is identified as the primary outcome. The listed secondary outcomes include real-world progression-free survival, real-world tumor response counts and response rate, initial palbociclib dose, time to dose adjustment, duration of treatment, time to next line of treatment, and time to subsequent chemotherapy. The record does not describe a hierarchy among these outcomes or state whether multiplicity was addressed through an adjustment procedure.

Multiplicity matters in general when several hypotheses are tested: the chance of at least one apparently unusual result can increase as the number of tests grows. A prespecified primary endpoint can help focus interpretation, but the registry record does not report a detailed statistical analysis plan or a multiplicity strategy. No conclusions about family-wise error or the confirmatory status of secondary findings can be drawn from the information recorded.

The analysis population is described as the FAS population, comprising all eligible participants enrolled in the study. The record also says that weighting created balanced virtual groups and that the overall number analyzed differed from the number enrolled. Group-specific numbers analyzed, outcome-specific denominators, and participant flow are not reported in the registry record.

Missing-data details are also limited. The record describes censoring of participants who did not die at 30-Sep-2020, but does not report the amount or handling of other missing information, missing covariates, or incomplete outcome records. It does not state whether imputation was used or describe sensitivity analyses for missing data. These omissions limit assessment of how missingness may affect the results.

11. What the Result Does, and Does Not, Mean

The record documents a retrospective real-world comparison of first-line palbociclib plus an aromatase inhibitor with aromatase inhibitor alone. It defines overall survival as time from treatment initiation to death from any cause or censoring and describes a weighted Kaplan-Meier analysis. Those details explain the question and broad method, but the numerical primary outcome results are not reported in the registry record.

Accordingly, the record does not establish the direction or size of an overall-survival difference, the precision of any estimate, or whether an estimate met a statistical significance threshold. A treatment effect cannot be inferred from the treatment names, enrollment total, or death counts alone. The death counts listed in the adverse-event record do not have group denominators and are not identified as treatment-related; they cannot be interpreted as comparative mortality rates.

Even if numerical estimates were available, an observational comparison would describe an association under the study’s design and assumptions. In general, weighting can reduce measured differences between groups, but it cannot guarantee that unmeasured differences have been eliminated. The registry record does not report the covariates and diagnostic information needed to judge the adequacy of that adjustment.

The results, as recorded, also cannot establish how the comparison applies beyond the described database population and treatment-initiation window. The registry describes de-identified data from United States clinical practices and first-line treatment initiated between Feb 3, 2015 and March 31, 2020. It does not provide enough detail to evaluate representativeness or transportability to other settings.

The defensible conclusion is methodological rather than clinical: the registry describes a weighted retrospective survival analysis but does not report its primary estimates. No claim that either treatment was more or less effective is supported by the posted information.

12. Why This Trial Matters Statistically

This study is statistically notable because it asks a comparative effectiveness question using retrospective clinical-practice data rather than a stated randomized allocation. Real-world records can support analyses of treatment patterns and outcomes in routine care, but interpretation depends on how treatment-group differences and follow-up are handled. The registry describes one such strategy—stabilized inverse probability of treatment weighting combined with Kaplan-Meier estimation for overall survival.

The weighting description is important because the enrolled groups may differ before analysis. In general, weighting aims to balance measured characteristics and can change the effective composition of the comparison, so the weighted analysis total need not match the enrollment count. The registry explicitly notes that the overall number analyzed differed from the number enrolled. This makes clear why enrollment alone is not a sufficient description of the population contributing to the estimated outcome.

The record also illustrates the distinction between a named analysis and an interpretable result. It defines the endpoint, follow-up origin, censoring date, and broad analytic approach, but does not report the median survival estimates or confidence intervals. Without those results, neither the effect’s direction nor its precision can be assessed. Similarly, the listed enrollment and death counts do not replace outcome-specific denominators or time-to-event estimates.

Finally, the listed range of secondary outcomes makes transparent that the study considered more than one endpoint, while the record does not describe multiplicity handling. Together, these features underscore why observational methods, analysis populations, outcome reporting, and uncertainty all matter when reading real-world comparative studies. The registry record supports understanding of the study question and framework, but not an estimate of the primary treatment comparison.

13. Limitations

• The posted summary statistics are the only data these results rest on; patient-level data, full protocols and statistical analysis plans are not available.
• Recalculated results are simple, unadjusted calculations from posted summaries. They do not replace the sponsor's prespecified analysis.

The main trial-specific limitation is that the registry record does not report numerical results for the primary overall-survival endpoint. Median estimates, 95% confidence intervals, group-specific analysis counts, and numerical hazard ratios are not reported in the registry record. The direction, magnitude, and precision of the primary comparison therefore cannot be evaluated.

The study is retrospective, and allocation is not stated. The record describes weighting to balance participant characteristics but does not identify the covariates, weighting model, balance diagnostics, or any sensitivity analyses. It is therefore not possible to assess from the record whether measured characteristics were adequately balanced or how much residual confounding may remain.

Although enrollment is listed as 2888, the record does not provide arm-specific enrollment or weighted analysis totals. It notes that the number analyzed differed from the number enrolled after construction of balanced virtual groups. Without participant flow and outcome-specific denominators, the scope of the analyzed population is unclear.

The record gives a censoring date of 30-Sep-2020 and approximately up to 68 months of follow-up, but does not report group-specific follow-up, censoring counts, or missing-data handling. It also does not describe a multiplicity procedure despite listing several outcomes and analyses. These missing details limit assessment of precision, robustness, and the interpretation of secondary results.

The adverse-event record lists death counts but no group denominators or attribution, and no serious or other adverse-event terms. Those entries cannot support comparative safety rates. Finally, the available description identifies the database setting and treatment-initiation window, but does not give enough information to judge how representative the study population is of other settings.

14. Related Tutorials

15. Related Statistical Calculators

16. Sources

• ClinicalTrials.gov: Real-World Effectiveness of Palbociclib in Combination With an Aromatase Inhibitor — NCT05361655.
• Registry record last updated 2024-06-07; status last verified 2024-05.
• Results first posted 2024-06-07.
• Data retrieved from ClinicalTrials.gov on 2026-11-09.
• Group sizes, means, event counts and sponsor-reported p-values are reproduced as posted. Effect estimates, confidence intervals and p-values not attributed to the sponsor are independently calculated from the posted summary statistics.

17. Record Summary

The registry record describes a completed retrospective study of first-line palbociclib plus an aromatase inhibitor versus an aromatase inhibitor alone for metastatic breast cancer. It uses de-identified information from the Flatiron Health Analytic Database and lists enrollment of 2888. Treatment initiation in the described cohorts occurred from Feb 3, 2015 to March 31, 2020.

The primary outcome is overall survival, measured from the index date to death from any cause or censoring at 30-Sep-2020. The stated method is Kaplan-Meier estimation adjusted using stabilized inverse probability of treatment weighting. The record lists 10 outcome measures and 7 analyses.

Numerical primary survival estimates, confidence intervals, group-specific analysis counts, and numerical hazard ratios are not reported in the registry record. As a result, the statistical design can be described, but the primary comparison’s direction and precision cannot be determined from the posted information.