PF-07304814 for Inpatients With COVID-19 (An ACTIV-3/TICO Treatment Trial): Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for PF-07304814 for Inpatients With COVID-19 (An ACTIV-3/TICO Treatment Trial) (NCT05780541), a randomized phase 3 study. The primary outcome is Number of Participants With Sustained Recovery.

2. Clinical Question

This question is narrower than whether a participant improved at any point during hospitalization. It combines discharge with a subsequent period alive and at home, so the endpoint represents a sustained state rather than a single observation. The registry record also lists deaths and safety outcomes, which provide additional perspectives on events during follow-up. They do not change the definition of the primary outcome.

The comparison evaluates the two treatment strategies as described in the record: PF-07304814 added to standard care, or placebo added to standard care. Since remdesivir was available to participants in both groups unless contraindicated, the contrast is not between PF-07304814 and no treatment. The registry does not report how many participants received remdesivir or whether use differed between groups.

3. Trial Design

The registry identifies this as a phase 3, randomized, triple-masked, parallel-group trial. Its official title describes a multicenter, adaptive, randomized, blinded controlled trial; the record does not provide further detail about the adaptive features. Enrollment is listed as 58. The trial is marked terminated. The reason for termination is not reported in the registry record.

The record lists a start date of 2021-09-15 and a primary completion date of 2022-04-06. The sponsor is the National Institute of Allergy and Infectious Diseases (NIAID). The trial is identified as ACTIV-3/TICO Treatment Trial H6 and concerns PF-07304814 for people hospitalized with COVID-19.

Participants were assigned to PF-07304814 plus standard of care or placebo plus standard of care. The PF-07304814 regimen is described as 250 mg per day for 5 days, given by constant-rate intravenous infusion. Placebo was given by intravenous infusion. Remdesivir was provided to all study participants as standard of care unless it was contraindicated for an individual participant, and was administered by intravenous infusion.

The primary outcome was sustained recovery through Day 90. The record defines this as discharge from the index hospitalization followed by being alive and at home for 14 consecutive days before Day 90. Other listed outcomes include death from all causes and safety outcomes through Day 5, Day 28, and Day 90. Although the registry lists six outcome measures, it does not report detailed numerical results for every listed outcome.

The record does not describe the allocation procedure, the masking roles, the intended enrollment target, or the trial's original sample-size assumptions. It also does not explain how termination affected follow-up or outcome collection. These design details should not be inferred from the trial title or the observed enrollment.

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: Of 58 subjects enrolled, 58 met inclusion criteria and were randomized to treatment.

Overall Study

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.

6. Primary Endpoint

The registry lists the primary outcome as Number of Participants With Sustained Recovery, measured over Through Day 90.

The primary endpoint counts participants who achieved sustained recovery through Day 90. To qualify, a participant had to be discharged from the hospitalization that began the study and then be alive and at home for 14 consecutive days before Day 90. The endpoint therefore combines a transition out of hospital with a period of continued survival and time at home.

This definition is useful because a single discharge event does not establish that recovery was sustained: a participant might return to hospital or die later in follow-up. Requiring a consecutive period alive and at home makes the outcome reflect more than the timing of discharge alone. This is a general rationale for sustained endpoints, not a claim about how the trial designers selected this definition; that rationale is not reported in the registry record.

The endpoint is reported as a count of participants meeting the definition in each group. The registry lists 23 of 32 participants in the PF-07304814 plus standard-of-care group and 13 of 26 in the placebo plus standard-of-care group. The comparison is a summary of these observed group outcomes. The record does not describe the handling of participants whose recovery status could not be determined, or provide a separate analysis population definition.

7. Results Posted in the Registry

• Placebo Plus SOC: 50.0% (13/26) versus 71.9% (23/32) on control; risk difference -21.9 percentage points (95% CI -46.6 to 2.9), risk ratio 0.70 (0.45–1.08); unadjusted two-proportion z-test p = 0.088.

8. Secondary Efficacy Endpoints

The registry also posts results for secondary outcomes. For each outcome below, the endpoint type, the test applied to the posted summary statistics, the result and its interpretation are given in turn. Secondary endpoints are usually not adjusted for multiple testing, so they support rather than confirm the primary finding.

Number of Participants Who Died From All Causes

• Endpoint type: Binary: the number of participants with the event, compared as proportions.
• Test: Unadjusted two-proportion z-test, with a Wald 95% confidence interval for the risk difference.
• Result:
  • Placebo Plus SOC: 23.1% (6/26) versus 28.1% (9/32) on control; risk difference -5.0 percentage points (95% CI -27.5 to 17.4), risk ratio 0.82 (0.34–2.01); unadjusted two-proportion z-test p = 0.662.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Placebo Plus SOC and the control. The interval gives the range of effects that remain compatible with the data.

The secondary outcomes add context to the sustained-recovery endpoint by covering mortality and safety over different follow-up windows. Death from all causes through Day 90 is the only secondary outcome for which the registry record includes a numerical group comparison. Its uncertainty interval includes no difference. Detailed numerical results for safety through Day 5, Day 28, and Day 90 are not reported in the registry record, so the overall secondary-outcome pattern cannot be characterized with the same level of numerical detail as the primary result.

The secondary outcomes address related but distinct questions. Mortality is not identical to sustained recovery, because the primary definition also depends on discharge and being alive and at home for a consecutive period. Safety outcomes concern adverse events rather than recovery. Even when outcomes are clinically related, a result for one does not automatically establish a result for another. A consistent interpretation therefore requires considering each endpoint's definition and available data, rather than treating the set as interchangeable evidence about a single effect.

Unadjusted secondary findings generally carry less confirmatory weight when several outcomes are examined. General statistical principle: considering multiple outcomes creates more opportunities for apparently notable findings to arise by chance. Adjustment or a prespecified testing hierarchy can address some of this issue, but the registry does not report such a strategy. It also lists no reported secondary analyses. The numerical death comparison should therefore be read as a secondary, unadjusted result, not as an independent confirmation of the primary comparison.

The record lists six outcome measures overall, but does not provide numerical results for every one. Missing result details limit the ability to assess whether secondary outcomes broadly support, qualify, or differ from the primary pattern. Their absence from the registry record should not be interpreted as evidence that an outcome did or did not occur.

Secondary outcomes registered for this trial, for reference (only those above were recalculated):

• Number of Participants Who Died From All Causes
• Number of Participants With a Safety Outcome Through Day 5
• Number of Participants With a Safety Outcome Through Day 28
• Number of Participants With a Safety Outcome Through Day 90

9. Statistical Interpretation

10. Statistical Methods Explained

The registry reports the primary outcome as a binary participant outcome: each participant either met the sustained-recovery definition or did not. The displayed counts give the number meeting the definition and the number at risk in each randomized group. For PF-07304814 plus standard of care, the record lists 23 of 32; for placebo plus standard of care, it lists 13 of 26. These are observed group summaries, not estimates adjusted for participant characteristics.

What does a risk difference express?

A risk difference subtracts one group's event proportion from the other group's event proportion. In this comparison, the posted contrast is placebo plus standard of care minus PF-07304814 plus standard of care. A negative value therefore means the observed proportion meeting the recovery definition was lower in the placebo group. A value of zero would correspond to equal observed proportions. The registry provides the risk-difference estimate and an interval around it, but the specific formula used for the interval is not reported in the registry record.

Risk differences use the absolute scale: they describe the gap between group proportions rather than how many times larger one proportion is than the other. This can make the estimate easier to connect to the outcome in each group. Other effect measures, such as risk ratios and odds ratios, express the contrast on different scales. The registry lists alternative measures in the comparison data, but the interpretation above focuses on the posted risk difference. Measures on different scales are not interchangeable, and their numerical values should not be directly compared as if they had the same meaning.

How should the confidence interval and p-value be read?

A confidence interval describes a range of effect values that are reasonably compatible with the observed data under the interval's statistical procedure. In repeated use of a procedure, intervals constructed in this way have a stated coverage property. A particular interval does not assign a probability to each possible value, and it does not mean that the true value has a specified probability of lying within its endpoints. The primary interval crosses no difference, so the posted data do not rule out no difference at the interval's stated level.

The registry reports an unadjusted p-value of 0.088 for the primary comparison. A p-value measures the compatibility of the observed data, or data more extreme, with a specified null model, assuming the model and test conditions apply. It is not the probability that the null model is true. The registry does not identify the test procedure, its exact assumptions, or whether any analysis other than the posted unadjusted comparison was conducted. No multiplicity-adjusted primary p-value is reported.

Why does limited information affect interpretation?

General statistical principle: estimates from a small dataset tend to be less precise than estimates from a larger dataset when other features are comparable. With relatively few observed outcomes, a change in the outcome status of a small number of participants can materially alter a group proportion. This does not establish that a particular result is biased; it explains why the uncertainty interval and the underlying counts deserve attention alongside the point estimate.

General statistical principle: random assignment is intended to make the groups comparable on average, but it does not ensure that the realized groups are identical in a particular trial. The registry identifies randomized allocation, but does not provide baseline characteristics, allocation details, or information about analysis adjustments. It is therefore not possible to assess balance or adjustment choices from the registry record alone.

What is known about the analysis?

The record identifies the primary measure as a count of participants and supplies the group counts and comparison summaries. It does not report an analysis population, missing-data strategy, prespecified statistical model, or methods for the confidence interval. The primary population description is blank. The record lists no reported analyses in the secondary-outcome analysis field. That absence means the registry record does not document those analysis details; it does not by itself establish that no analysis was performed.

The registry marks the trial as adaptive in its official title, but does not report how adaptation was implemented or how it affected inference. Details about interim looks, stopping boundaries, or changes to allocation or sample size are not reported in the registry record. Without those details, the statistical implications of the adaptive design cannot be evaluated. The same caution applies to the terminated status: the record does not state when or why the trial was terminated, or whether termination affected endpoint ascertainment.

11. Confidence Intervals and Precision

The primary risk-difference interval extends from a negative value to a value slightly above no difference. Its span indicates substantial uncertainty about the size and even the direction of the underlying contrast. The point estimate favors a higher observed recovery proportion in the PF-07304814 group, but the interval also allows for no difference and for a small contrast in the opposite direction. The record's unadjusted p-value is 0.088, which is consistent with the interval not excluding no difference.

Confidence intervals are useful because they show the range of estimates compatible with the data under a specified procedure, rather than giving only a single point estimate. A wide interval signals limited precision: the data do not distinguish closely among multiple possible effects. It should not be read as proof that all values in the interval are equally likely, or that values outside it are impossible. The registry does not state the method used to construct the primary interval, so method-specific properties cannot be assessed from the record.

The group counts are modest, and the outcome counts are limited. General statistical principle: estimates based on fewer observations often have more sampling variability, particularly for binary outcomes. That helps explain why an apparently sizable difference in the observed proportions can still be accompanied by an interval crossing no difference. This is a general explanation of precision, not a claim about the exact contribution of any one design feature in this trial.

The mortality comparison is also imprecise: its confidence interval includes no difference and allows effects in either direction. The registry does not report confidence intervals for every listed safety outcome. Overall, the posted intervals support cautious interpretation of the estimated contrasts rather than a precise claim about their magnitude.

12. 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 sample-size illustrations are planning calculations based on the observed primary effect, not the trial's original design assumptions. The record states a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. The calculations consider detecting the stated fraction of the observed effect. The trial's original sample-size justification and assumptions are not reported in the registry record.

Each row describes planning for a specified fraction of the observed effect. At 100% of the observed effect, the illustration lists 77 participants per arm for 80% power and 103 participants per arm for 90% power. At 75% of the observed effect, it lists 134 participants per arm for 80% power and 179 participants per arm for 90% power. At 50% of the observed effect, it lists 292 participants per arm for 80% power and 390 participants per arm for 90% power.

In plain language, the 80% power column gives the participants per arm estimated to provide an 80% chance of detecting the specified effect under the stated assumptions, if that effect were the true one and the calculation's model applied. The 90% power column uses a higher target chance of detecting that effect, and consequently requires more participants per arm in each listed illustration. Smaller effects are harder to distinguish from random variation, so planning to detect a smaller fraction of the observed effect requires more participants.

These figures are not observed power. They are hypothetical planning illustrations anchored to an effect observed in a small trial, and should not be mistaken for a guarantee that a future study will obtain a particular result. The record does not say that these calculations informed the trial's actual enrollment or design.

13. 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 = 0.719; treatment proportion p₁ = 0.500; average p̄ = 0.609.
2. First term: z_1−α/2 × √[2p̄(1−p̄)] = 1.96 × 0.6900 = 1.3523.
3. Second term: z_1−β × √[p₀(1−p₀) + p₁(1−p₁)] = 0.842 × 0.6724 = 0.5659.
4. n = (1.3523 + 0.5659)² / (-0.2188)² = 76.9, rounded up to 77 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 <- 0.71875                          # control proportion
p1 <- 0.5                          # 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(5780541)
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 = 77 (80% power), 103 (90% power)
 75% of observed effect: n per arm = 134 (80% power), 179 (90% power)
 50% of observed effect: n per arm = 292 (80% power), 390 (90% power)
power.prop.test: 76.9 
Simulated power at n = 77 per arm: 0.808 (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.

14. Multiplicity, Missing Data and Analysis Populations

The registry record does not report how multiplicity was addressed. It lists a primary endpoint and several secondary outcomes, but does not describe a hierarchical testing sequence, family-wise error procedure, or other adjustment for examining multiple outcomes. The reported primary p-value is explicitly unadjusted, and the secondary mortality p-value is also presented as unadjusted. Without a stated analysis plan, the confirmatory status of the secondary comparisons cannot be determined.

General statistical principle: when researchers examine multiple outcomes, each additional test creates another opportunity for a result to appear unusual under a null model. Adjustment procedures can control the chance of at least one false positive across a defined family of tests, while a testing hierarchy can specify which comparisons are intended to support formal claims. The appropriate approach depends on the prespecified trial objectives and analysis plan; those details are not reported in the registry record.

The registry lists no reported analyses for the secondary outcomes. It does not provide the protocol-defined analysis populations, such as whether the primary analysis included all randomized participants or a different subset. The primary population description is blank. The record also does not report whether participants were analyzed according to assigned group or how deviations from assigned treatment were handled.

Missing-data methods are not reported in the registry record. In particular, it does not state how unavailable Day 90 status was handled, whether any outcome values were imputed, or how incomplete follow-up was considered. The endpoint definition requires a recovery status extending over time, and interpretation depends on how that status was determined for participants without complete follow-up. No assumptions about missing outcomes should be inferred.

The trial is described as adaptive in its official title, but details of interim analyses, decision rules, or adjustments to statistical thresholds are not reported. The record also marks the trial as terminated without giving the reason or explaining whether termination altered analysis plans. These omissions prevent a full assessment of the inferential framework. They do not demonstrate that appropriate procedures were absent; they mean that the registry record does not document them.

15. Safety Results

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

Deaths

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: small intestinal obstruction (3.8%), pneumothorax (3.1%), deep vein thrombosis (3.1%).

The registry lists serious adverse events for 2 of 32 participants in the PF-07304814 plus standard-of-care group and 1 of 26 in the placebo plus standard-of-care group. It also lists other adverse events for 11 of 32 and 9 of 26, respectively. These are participant counts in the record; they should be read separately from the primary recovery outcome and from the count of deaths.

The record identifies small intestinal obstruction (3.8%), pneumothorax (3.1%), and deep vein thrombosis (3.1%) among the top serious terms. It does not provide arm-specific rates for these listed terms in the information reported. The registry record does not explain how the event terms were selected or provide additional detail on severity, timing, or attribution.

These descriptive counts do not establish that events were caused by a study treatment, nor do they by themselves provide a precise comparison of safety between groups. Interpretation is limited by the number of participants, the absence of detailed event-level information, and the lack of reported analysis methods for these safety comparisons.

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

The posted primary data show that sustained recovery through Day 90 was recorded more often in the PF-07304814 plus standard-of-care group than in the placebo plus standard-of-care group. The estimated risk difference points in that direction, but its confidence interval includes no difference and extends to a value on the other side of no difference. The unadjusted p-value is 0.088. Together, these findings describe a potentially meaningful observed contrast with considerable uncertainty; they do not establish a precise underlying effect.

The result does not prove that PF-07304814 caused the observed group difference, and it does not prove that the groups have identical outcomes. Statistical uncertainty works in both directions: the data are not precise enough to rule out no difference, but failure to rule out no difference is not evidence that the true difference must be zero. The posted estimate should therefore be interpreted with its interval and the trial context, not as a binary verdict based only on a p-value.

The comparison is between PF-07304814 plus standard of care and placebo plus standard of care. Remdesivir was provided to participants as standard of care unless contraindicated for an individual participant. The results therefore describe the recorded strategies in that context; they are not a comparison of PF-07304814 against no standard care. The record does not report remdesivir use by group or other details that would permit further assessment of background treatment.

The mortality and safety information provides additional observations, but it does not resolve the uncertainty around the primary outcome. The mortality comparison is also imprecise, and detailed results for the listed safety time points are not reported in the registry record. The record does not report an explanation for termination, missing-data procedures, or full statistical methods. Each limits how comprehensively the posted results can be evaluated.

These results are not medical advice and do not determine what treatment an individual should receive. They describe what was recorded among participants in this trial. Broader judgments about treatment effects would require consideration of other evidence and the details of the full analysis, neither of which is established by this registry record alone.

17. Why This Trial Matters Statistically

This trial illustrates why an observed difference and the precision of that difference must be considered together. The group counts show a higher observed sustained-recovery proportion in the PF-07304814 group, while the confidence interval for the risk difference still includes no difference. A point estimate can look substantial without being estimated precisely. Reporting both the event counts and uncertainty is essential to avoid treating the point estimate as a settled answer.

It also illustrates the importance of defining outcomes carefully. Sustained recovery is not simply discharge: it requires being alive and at home for 14 consecutive days before Day 90 after discharge from the index hospitalization. That definition affects what the event count represents and what kind of clinical course the endpoint captures. Other outcomes, such as mortality or safety events, address different aspects of the trial and cannot substitute for the primary endpoint.

The record highlights how limited registry detail can constrain statistical interpretation. It identifies a randomized, triple-masked, parallel design and reports primary counts and a comparison, but it does not provide the analysis population, missing-data approach, multiplicity strategy, or detailed adaptive-design methods. Such information is important for understanding how an estimate was produced and which conclusions were intended to be confirmatory.

Finally, the trial is marked terminated and reports enrollment of 58. The record does not state why it ended or how termination affected the analysis. The statistical lesson is not to infer a reason from the outcome summaries, but to keep the limits of the documented evidence visible when describing the results.

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.

Several trial-specific limitations narrow what can be concluded from the registry record. Enrollment is listed as 58, and the primary group counts are 23 of 32 and 13 of 26. With limited counts, uncertainty around the comparison is substantial. The primary confidence interval includes no difference, so the observed direction should not be presented as a precise estimate of the underlying contrast.

The trial is marked terminated, but the reason for termination is not reported in the registry record. Nor does the record say whether the stopping decision affected recruitment, follow-up, endpoint collection, or the planned analyses. Because the official title describes an adaptive trial, the absence of details on adaptation and interim decision rules is an additional barrier to evaluating the statistical context.

The registry does not report the primary analysis population, missing-outcome handling, protocol-defined model, or the method used to construct the confidence interval. It also does not report a multiplicity strategy. These omissions limit the ability to assess robustness, reproduce the analysis, or determine the intended status of secondary findings. The record does not give baseline characteristics or group-specific remdesivir use, so comparability on those features cannot be reviewed from these results.

Numerical results are incomplete across the listed outcomes. Death is reported as a secondary result, but detailed results for safety outcomes through Day 5, Day 28, and Day 90 are not reported in the registry record. The listed serious-event terms do not include arm-specific counts in the information reported. These limitations do not show that unreported outcomes were absent; they limit the conclusions available from the record.

19. Related Tutorials

20. Related Statistical Calculators

21. Sources

• ClinicalTrials.gov: PF-07304814 for Inpatients With COVID-19 (An ACTIV-3/TICO Treatment Trial) — NCT05780541.
• Registry record last updated 2024-02-09; status last verified 2024-02.
• Results first posted 2024-02-09.
• 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

This is a phase 3 randomized, triple-masked, parallel-group trial of PF-07304814 plus standard of care versus placebo plus standard of care in hospitalized people with COVID-19. The registry lists 58 enrolled participants, a start date of 2021-09-15, a primary completion date of 2022-04-06, and a terminated status. The reason for termination is not reported in the registry record.

The primary outcome was sustained recovery through Day 90. It was recorded for 23 of 32 participants assigned to PF-07304814 plus standard of care and 13 of 26 assigned to placebo plus standard of care. The reported risk-difference interval includes no difference, and the unadjusted p-value is 0.088. The record also lists mortality and safety outcomes, but detailed results are not available for every listed outcome.

The registry record does not report the original sample-size justification, full analysis plan, missing-data methods, analysis populations, or details of the adaptive design. These limits are relevant when interpreting the observed contrasts and their precision.