Chemotherapy After Prostatectomy (CAP) For High Risk Prostate Carcinoma: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Chemotherapy After Prostatectomy (CAP) For High Risk Prostate Carcinoma (NCT00132301), a randomized phase 3 study. The primary outcome is Number of Participants With Progression-Free Survival.

2. Clinical Question

The clinical question concerns the timing and addition of chemotherapy after surgery, rather than chemotherapy for disease already documented as progressing. The trial was designed to assess whether early adjuvant treatment changes progression-free survival for patients whose cancer had been surgically removed but who remained at high risk of progression. The primary endpoint description identifies detectable PSA as evidence of progression. The registry describes follow-up as centralized and lasting up to 100 months.

The registry identifies 298 enrolled participants and describes a randomized comparison. Its arm summaries report 66 participants with progression-free survival among 140 in the docetaxel-and-prednisone arm, and 84 among 157 in the standard-care arm. These are posted participant counts; they do not, by themselves, describe when progression occurred or how follow-up time was incorporated. The registry also reports a separate log-rank analysis using a hazard ratio, which addresses the time-to-event comparison. The interpretation therefore depends on distinguishing the endpoint counts from that reported time-to-event analysis, and on recognizing the uncertainty in both summaries.

3. Trial Design

The registry describes a phase 3, randomized, single-masked, parallel-group trial sponsored by the VA Office of Research and Development. The study is titled “Chemotherapy After Prostatectomy (CAP) For High Risk Prostate Carcinoma,” and its official title is “CSP #553 - Adjuvant Therapy in Prostate Carcinoma Treatment.” It is listed as completed. The recorded start is 2006-06 and the primary completion is 2015-10.

The two arms compare chemotherapy after radical prostatectomy with standard care. The chemotherapy arm received docetaxel and prednisone; the comparator arm was standard care. The registry’s description of standard care is surveillance, with androgen deprivation at biochemical relapse. The study summary describes the intended population as Veterans at high risk for progression after prostatectomy and characterizes the intervention as early adjuvant chemotherapy added to standard care.

The enrollment entry is 298. For the primary analysis, the registry says that one participant in the chemotherapy arm was excluded by the Data Monitoring Committee, leaving 140 participants in that arm’s analysis population. The reported standard-care analysis population is 157. The record does not explain the reason for the exclusion beyond the committee’s decision, nor does it describe other exclusions or participant flow in detail.

Single masking is recorded, but who was masked is not reported in the registry record. The planned allocation ratio, the original sample-size justification, and further operational details of the randomization are not reported in the registry record. The posted results include 1 primary outcome measure and 1 reported analysis. No secondary outcomes are posted. The registry record does not give further detail about the analysis plan or specify how the primary endpoint count relates to the timing of events for each participant.

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.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: One participant in Arm 1 was excluded from analysis by the Data Monitoring Committee (DMC), making the analysis population in Arm 1 140 participants.

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: One participant in Arm 1 was excluded from analysis by the Data Monitoring Committee (DMC), making the analysis population in Arm 1 140 participants.

6. Primary Endpoint

The registry lists the primary outcome as Number of Participants With Progression-Free Survival, measured over Up to 100 months (centralized follow-up).

The primary endpoint is titled “Number of Participants With Progression-Free Survival.” It records participants meeting the progression-free survival outcome over a timeframe of up to 100 months, with centralized follow-up. The endpoint description states that the study sought to determine whether early docetaxel plus prednisone, compared with standard care alone, reduced disease progression evidenced by detectable PSA in patients with high-risk prostate cancer after radical prostatectomy.

Progression-free survival is used to describe the time until a specified progression event, or the absence of such an event during follow-up. In general, an endpoint tied to time can retain information about when an event occurs and how long participants are followed. A participant count, in contrast, gives the number recorded in an outcome category but does not itself show event timing or account for unequal follow-up. The registry also reports a log-rank analysis with a hazard ratio, which is the posted time-to-event comparison. These two forms of result should not be treated as interchangeable.

The endpoint name and description do not provide a full operational definition of progression, including all assessment rules or handling of events other than detectable PSA. Those details are not reported in the registry record. The posted counts should therefore be read as the recorded primary outcome counts, while the time-to-event estimate should be interpreted through its reported method and confidence interval.

7. Results Posted in the Registry

• Arm 1: Docetaxel and Prednisone: 47.1% (66/140) versus 53.5% (84/157) on control; risk difference -6.4 percentage points (95% CI -17.7 to 5.0), risk ratio 0.88 (0.70–1.11); unadjusted two-proportion z-test p = 0.274.

8. Statistical Interpretation

9. Statistical Methods Explained

The registry presents two kinds of statistical summary for the primary outcome: counts of participants with progression-free survival and a log-rank time-to-event analysis reported as a hazard ratio. General statistical concepts help clarify why the results are not identical in form and what each can establish.

What does a participant count show?

A count reports how many participants are recorded in an outcome category within an analysis group. The chemotherapy arm has 66 participants with progression-free survival out of 140 in its analysis population; the standard-care arm has 84 out of 157. A count is direct to read, but it does not show when each participant met the outcome, how long each participant was observed, or how censoring was handled. For outcomes where follow-up duration and event timing matter, a count-only comparison can omit important information.

The recalculated p = 0.274 is an unadjusted comparison based on the posted summary counts. In general, an unadjusted analysis uses the summarized data without adjustment for other characteristics. It should not be conflated with an analysis using participant-level follow-up times. The registry does not report covariate adjustment for the posted log-rank analysis.

What does the log-rank analysis address?

In general, the log-rank test compares the pattern of time-to-event experience between groups across follow-up, taking account of the ordering of observed events and the periods in which participants remain under observation. The registry reports a hazard ratio of 0.80 from a log-rank analysis, with a 95% confidence interval from 0.58 to 1.11 and p = 0.40. A hazard ratio summarizes a relative event-hazard contrast under the analysis framework; it is not a ratio of the posted participant counts and is not the same as a risk difference.

The registry does not report the underlying event times, censoring information, detailed model specification, or the assumptions used for the time-to-event analysis. In general, interpretation of a hazard ratio is strongest when the analysis assumptions are appropriate for the data. The registry record does not provide enough detail to independently evaluate those assumptions.

Why are the two analyses different?

The count-based recalculation and the sponsor-reported time-to-event analysis use different summaries and methods. The count-based calculation does not incorporate the timing of events; the log-rank analysis is designed for time-to-event data. Their p-values need not match, even when they concern the same clinical endpoint. The registry does not report a discrepancy between the posted results. The count-based recalculation is best read as a supplementary, unadjusted summary rather than a replacement for the reported analysis.

What does a p-value tell the reader?

In general, a p-value measures how compatible the observed data are with a specified null hypothesis under the assumptions of the analysis. It is not the probability that the null hypothesis is true, the probability that the result occurred by chance, or a measure of clinical importance. The reported p = 0.40 belongs to the log-rank analysis; p = 0.274 belongs to the unadjusted recalculation. Neither value describes the magnitude or practical importance of an effect on its own.

Statistical significance is not a substitute for examining the estimate and its interval. A result may be compatible with a range of effects, and a p-value alone does not show how wide that range is. For this trial, the confidence interval around the hazard ratio is relevant to that uncertainty and includes the no-difference value.

What is over-dispersion, and does it apply?

In general, over-dispersion occurs when observed variability is greater than a model assumes. It can make standard errors too small if a model fails to account for the extra variability. The posted primary result is a binary participant count, and the sponsor-reported analysis is a log-rank comparison; the registry does not report a dispersion estimate or identify an over-dispersion adjustment. It would therefore be inappropriate to assert that over-dispersion was present or that a particular correction was needed.

What cannot be reconstructed from the registry summary?

The registry does not report participant-level records, event dates, the detailed endpoint assessment rules, or a full statistical analysis plan. It also does not report the original sample-size justification or whether there were interim analyses. Without those details, the posted summaries cannot be used to reproduce the time-to-event analysis or determine whether all protocol-specified analyses are represented. The available figures support a careful description of the reported results, not a reconstruction of unreported methods.

10. Confidence Intervals and Precision

The sponsor-reported time-to-event estimate is a hazard ratio of 0.80, with a 95% confidence interval from 0.58 to 1.11. In general, a confidence interval gives a range of estimates compatible with the data and the method, subject to the analysis assumptions. It is not a guarantee that the true effect lies in that range, nor does it represent the range of individual participant outcomes.

The interval includes 1, the no-relative-difference value for a hazard ratio. The interval therefore allows for an estimated lower event hazard in the chemotherapy arm as well as an estimate slightly above the no-difference value. Its breadth indicates that the data do not pin down a narrow range of relative effects. The reported p = 0.40 is consistent with the interval’s inclusion of the no-difference value under the reported analysis.

Precision concerns how much uncertainty surrounds an estimate, rather than whether the point estimate alone appears favorable. The hazard ratio point estimate is below 1, but the interval includes values on both sides of 1. The posted result consequently does not precisely distinguish a relative reduction from no difference or a relative increase in event hazard. The registry does not provide additional interval estimates for the participant counts.

The unadjusted count comparison has a reported recalculated p = 0.274, but the recalculated risk-difference estimate and its interval have more than 3 decimal places in the underlying registry summaries and are not reproduced. In general, counts and p-values do not convey the full range of plausible effect sizes as directly as a confidence interval. The sponsor-reported hazard ratio interval is the principal posted measure of precision for the time-to-event analysis.

Confidence intervals are conditional on the analysis method and its assumptions. The registry record does not give enough detail to independently assess those assumptions, the influence of censoring, or the precision of subgroup effects. No subgroup precision estimates are posted.

11. Planning a Similar Trial: Sample Size

This is a planning illustration, not a re-evaluation of the original trial. It asks: if a new study wanted to detect a difference like the one observed, or a smaller one, how large would it need to be? It is deliberately not “observed power” of the finished trial, which is not a meaningful quantity.

The planning figures are expressed as participants per arm for 80% power and participants per arm for 90% power. They are not total event counts. The assumptions are a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. The effect to detect is set to the stated fraction of the observed effect. The trial’s original design assumptions and original sample-size justification are not reported in the registry record.

For planning at 100% of the observed effect, the illustration gives 969 participants per arm for 80% power and 1,297 participants per arm for 90% power. For planning at 75% of the observed effect, it gives 1,723 participants per arm for 80% power and 2,306 participants per arm for 90% power. For planning at 50% of the observed effect, it gives 3,874 participants per arm for 80% power and 5,186 participants per arm for 90% power.

In plain language, 80% power means that a study planned under the stated assumptions would have an 80% chance of detecting the specified effect at the chosen significance level if that effect were truly present. The 90% column uses a higher target chance of detection, which requires more participants under the same assumptions. Smaller effects are harder to distinguish from random variation and therefore require more participants. That pattern is visible across the planning scenarios.

These are planning illustrations based on the observed effect and the stated assumptions, not observed power. They do not show how much power the completed trial had under its original design, because the original design assumptions are not reported in the registry record. They should not be read as a retrospective judgment about trial quality or as evidence that any particular effect is established.

12. How the Sample Size Is Derived

Power is the probability that a study of a given size rejects the null hypothesis when a specific true effect exists. It depends on four things: the effect to detect, the variability of the outcome, the significance level (here a two-sided α of 0.05, so z_1−α/2 = 1.960) and the desired power (z_1−β = 0.842 for 80% and 1.282 for 90%).

Formula

n is the number of participants in each arm, p₀ and p₁ are the control and treatment proportions, and p̄ is their average.

Worked calculation for this trial

1. Control proportion p₀ = 0.535; treatment proportion p₁ = 0.471; average p̄ = 0.503.
2. First term: z_1−α/2 × √[2p̄(1−p̄)] = 1.96 × 0.7071 = 1.3859.
3. Second term: z_1−β × √[p₀(1−p₀) + p₁(1−p₁)] = 0.842 × 0.7057 = 0.5939.
4. n = (1.3859 + 0.5939)² / (-0.0636)² = 968.9, rounded up to 969 participants per arm for 80% power.

The smaller-effect rows in the table above repeat the same calculation with the effect multiplied by 0.75 and 0.5 (for ratios, applied to the logarithm of the ratio).

R code

alpha <- 0.05
z_a   <- qnorm(1 - alpha / 2)          # two-sided
p0 <- 0.535031847133758                          # control proportion
p1 <- 0.471428571428571                          # treatment proportion
n_per_arm <- function(power, frac = 1) {
  q1 <- p0 + (p1 - p0) * frac               # smaller effect = fraction of the observed difference
  pbar <- (p0 + q1) / 2
  z_b <- qnorm(power)
  ceiling(round((z_a * sqrt(2 * pbar * (1 - pbar)) + z_b * sqrt(p0 * (1 - p0) + q1 * (1 - q1)))^2 / (q1 - p0)^2, 6))
}
for (f in c(1, 0.75, 0.5))
  cat(sprintf("%3.0f%% of observed effect: n per arm = %s (80%% power), %s (90%% power)\n", 100 * f,
              format(n_per_arm(0.80, f), big.mark = ","), format(n_per_arm(0.90, f), big.mark = ",")))

# Check against R's built-in function (unrounded n per arm)
cat("power.prop.test:", round(power.prop.test(p1 = p0, p2 = p1, power = 0.80)$n, 1), "\n")

# Simulation: how often does the z-test reject at the 80%-power sample size?
set.seed(132301)
n    <- n_per_arm(0.80)
reps <- 20000
x0 <- rbinom(reps, n, p0); x1 <- rbinom(reps, n, p1)
ph <- (x0 + x1) / (2 * n)
z  <- ((x1 - x0) / n) / sqrt(ph * (1 - ph) * 2 / n)
cat(sprintf("Simulated power at n = %s per arm: %.3f (%d replicates)\n", format(n, big.mark = ","), mean(abs(z) > z_a, na.rm = TRUE), reps))

Output from running the code

100% of observed effect: n per arm = 969 (80% power), 1,297 (90% power)
 75% of observed effect: n per arm = 1,723 (80% power), 2,306 (90% power)
 50% of observed effect: n per arm = 3,874 (80% power), 5,186 (90% power)
power.prop.test: 968.9 
Simulated power at n = 969 per arm: 0.793 (20000 replicates)

The simulation draws many pairs of event counts at the planned sample size and counts how often the two-proportion z-test rejects the null. It should land close to the target power; at very small required sample sizes it can fall below it, because the normal approximation is rough there.

13. Analyses Reported by the Sponsor

The registry’s reported analysis is a log-rank comparison with a hazard ratio of 0.80, a 95% confidence interval from 0.58 to 1.11, and p = 0.40. This is the trial’s posted time-to-event analysis and is the most direct registry result for interpreting the primary endpoint as progression-free survival over follow-up.

The recalculated comparison uses the posted participant counts and gives p = 0.274. It is a simple unadjusted analysis of those summary counts, not a reproduction of the log-rank method. It does not incorporate event timing in the way a time-to-event analysis does. Differences between the two p-values are therefore understandable from the distinct data summaries and methods; they do not, by themselves, indicate a contradiction.

The registry record does not report a discrepancy between its reported analysis and the recalculation. The recalculation can provide a supplementary view of the posted counts, but it should not displace the trial’s reported time-to-event result. Neither analysis should be interpreted from its p-value alone: the hazard ratio and its interval describe the direction and uncertainty of the reported relative event-hazard estimate, while the counts describe the participant totals recorded in each arm.

The registry does not provide the detailed protocol or analysis plan needed to compare every analytic choice with the original plan. It also does not report participant-level data that would permit an independent reconstruction of the sponsor-reported result.

14. Multiplicity, Missing Data and Analysis Populations

The registry posts 1 primary outcome measure and 1 analysis, with 0 secondary outcomes posted. This is a description of what appears in the registry record, not evidence that the trial had no other planned outcomes or analyses. Whether other outcomes were collected, analyzed, or reported elsewhere is not reported in the registry record.

Multiplicity arises in general when several hypotheses, endpoints, time points, subgroups, or analyses are tested. With repeated opportunities to find a statistically unusual result, the chance of at least one such result can increase unless the analysis plan accounts for the testing structure. The registry record does not report a multiplicity adjustment, an order of testing, or the number of analyses conducted beyond the posted analysis. It is not possible to determine from the registry record whether multiplicity was relevant to the full trial program or how it was handled.

The primary analysis population is described as excluding one participant in the docetaxel-and-prednisone arm by decision of the Data Monitoring Committee, resulting in 140 participants in that arm’s analysis population. The standard-care analysis group contains 157 participants. The registry does not report a fuller account of the analysis population, such as the reasons for any other differences between enrollment and analysis groups, or whether all randomized participants were included in an intention-to-treat analysis.

Missing outcome data, loss to follow-up, censoring rules, and methods for handling missing observations are not reported in the registry record. These details matter because missingness and follow-up patterns can affect estimates, particularly in time-to-event analyses. The registry describes centralized follow-up up to 100 months, but that does not by itself specify the completeness of follow-up or how participants without observed progression were treated analytically.

The registry record also does not report interim analyses, stopping rules, the detailed statistical analysis plan, or the rationale for the Data Monitoring Committee’s exclusion decision. No conclusion about those procedures should be inferred from their absence in the posted record. The available information supports describing the stated analysis population and posted results, while leaving the broader multiplicity and missing-data procedures unresolved.

15. Safety Results

The registry reports serious adverse events as affected participants divided by participants at risk.

Other (non-serious) adverse events

Participants with at least one non-serious event at or above the registry's 0% reporting threshold.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: pneumonia (3.6%), urinary incontinence (3.2%), cornary artery disease (2.1%), myocardial infarction (2.1%), post procedural complication (2.1%).

The serious adverse event summaries list 47 affected participants among 140 at risk in the chemotherapy-agent arm and 50 among 157 at risk in the standard-care arm. These are participant counts, not event counts; the registry does not state how many separate serious events occurred for affected participants. The record also lists other adverse events affecting 79 participants among 140 at risk in the chemotherapy-agent arm and 3 among 157 at risk in the standard-care arm. The record does not provide enough detail to define the full collection or classification procedures for these summaries.

The top listed serious-event terms and reported percentages are Pneumonia, 3.6%; Urinary incontinence, 3.2%; Cornary artery disease, 2.1%; Myocardial infarction, 2.1%; Post procedural complication, 2.1%; and Chest pain, 1.9%. The registry’s spelling of “Cornary artery disease” is retained as posted. These entries describe the listed serious-event terms and are not a complete account of all possible safety outcomes.

In general, adverse-event summaries require attention to the population at risk, the event definition, the period of collection, and whether the figures represent people or event occurrences. The figures should not be compared as if they were adjusted for follow-up or other participant characteristics; such adjustments are not reported in the registry record. Death counts are not reported. These safety summaries are descriptive and do not, on their own, establish a causal relationship between treatment and any event.

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

The posted results show the participant counts recorded for progression-free survival in each arm and a separate time-to-event comparison. The registry reports a hazard ratio of 0.80 for the log-rank analysis, with a 95% confidence interval from 0.58 to 1.11 and p = 0.40. The interval includes 1, so the reported estimate is not sufficiently precise to rule out no relative difference in event hazard or an estimate in the opposite direction. The unadjusted recalculation from the counts gives p = 0.274 and should be understood as a different, supplementary analysis.

These statistics do not prove that the two strategies are equivalent. Demonstrating equivalence or non-inferiority requires a prespecified design and margin, neither of which is reported in the registry record. Nor do the results establish that early chemotherapy reduces progression. The point estimate’s direction is not sufficient to establish a treatment effect when uncertainty remains, and a non-significant p-value does not establish that there is no effect.

The endpoint is described in terms of progression-free survival and detectable PSA, but the registry record does not give a complete endpoint definition or detailed follow-up and censoring information. The posted summaries also do not provide estimates for secondary endpoints, because none are posted. The conclusions should remain tied to the specific outcome and analyses available in the record.

The safety summaries describe affected participants and listed serious-event terms, but they do not establish causation. Differences in counts alone do not account for exposure duration, event recurrence, or other clinical characteristics, and those details are not reported in the registry record. The trial’s findings should therefore be read as statistical evidence with uncertainty, not as individualized medical advice or a treatment recommendation.

17. Why This Trial Matters Statistically

This trial illustrates why the statistical representation of an endpoint matters. Progression-free survival is inherently related to time, yet the registry also gives participant counts in outcome categories. Counts offer an accessible snapshot, while time-to-event analyses use event timing and follow-up information. The reported log-rank hazard ratio and the count-based recalculation answer related but distinct statistical questions, so interpreting them side by side requires attention to their methods.

It also shows why estimates and confidence intervals are more informative than a binary focus on statistical significance. The hazard ratio is below 1, but its 95% confidence interval extends across 1. That interval makes the uncertainty explicit and cautions against treating the point estimate as a settled effect. A p-value alone would not show the range of estimates compatible with the data.

The record also offers a useful distinction between trial planning and observed results. The sample-size figures are illustrations based on the observed effect and specified assumptions. They are not the trial’s original design rationale and do not reveal its observed power. The original sample-size justification is not reported in the registry record, so the planning illustration cannot fill that gap.

Finally, the posted record is incomplete for several questions that often shape interpretation: full outcome definitions, missing-data procedures, multiplicity, interim analyses, and detailed analysis-plan choices. Identifying those limits is part of sound statistical reading. It prevents a narrow set of posted summaries from being treated as a complete account of the trial while still allowing the reported estimates and their uncertainty to be described clearly.

18. 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 registry record provides limited detail for independently evaluating the primary endpoint. It describes progression through detectable PSA and gives a timeframe of up to 100 months, but does not report the full definition of progression-free survival, event adjudication rules, assessment schedule, or handling of events other than detectable PSA. The relationship between the participant counts and the time-to-event analysis is not explained in detail.

Participant-level event times, censoring information, and follow-up completeness are not reported in the registry record. Those details would be needed to reproduce the log-rank analysis or evaluate how follow-up patterns affected the hazard ratio. The record identifies a single-masked design but does not state who was masked. It also does not describe the randomization procedure or allocation ratio.

The analysis population has a stated exclusion: one participant in the docetaxel-and-prednisone arm was excluded by the Data Monitoring Committee, leaving 140 participants in that arm’s analysis population. The reason beyond the committee’s decision, the timing of the decision, and whether the exclusion followed a prespecified rule are not reported in the registry record. The record does not provide a complete participant-flow account.

The registry does not report the original sample-size justification, interim analyses, stopping rules, detailed statistical analysis plan, multiplicity procedures, or methods for handling missing data. No secondary outcomes are posted. The serious adverse-event summaries do not give event counts per participant, deaths, or a detailed collection window. These gaps limit assessment of analysis choices, safety context, and the completeness of reporting. They do not invalidate the posted results, but they restrict how far the results can be independently checked or generalized beyond the population and endpoint described.

19. Related Tutorials

20. Related Statistical Calculators

21. Sources

• ClinicalTrials.gov: Chemotherapy After Prostatectomy (CAP) For High Risk Prostate Carcinoma — NCT00132301.
• Registry record last updated 2018-06-25; status last verified 2018-05.
• Results first posted 2018-06-25.
• 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.

22. Record Summary

The registry describes a completed, phase 3, randomized, single-masked, parallel-group study of adjuvant docetaxel and prednisone after radical prostatectomy compared with standard care for Veterans at high risk of prostate cancer progression. Enrollment is recorded as 298. The primary outcome is the number of participants with progression-free survival, with centralized follow-up up to 100 months.

The posted arm summaries report 66 participants with progression-free survival among 140 in the chemotherapy arm and 84 among 157 in standard care. The reported log-rank analysis gives a hazard ratio of 0.80, a 95% confidence interval from 0.58 to 1.11, and p = 0.40. A simple unadjusted recalculation from the posted counts gives p = 0.274; it is not a replacement for the time-to-event analysis.

The record includes serious and other adverse-event summaries, but important details about outcome definitions, missing data, multiplicity, and the original design assumptions are not reported in the registry record. The results should be interpreted as posted, with their uncertainty and these reporting limits in view.