Study to Evaluate Efficacy and Safety of PF-04965842 in Subjects Aged 12 Years And Older With Moderate to Severe Atopic Dermatitis: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Study to Evaluate Efficacy and Safety of PF-04965842 in Subjects Aged 12 Years And Older With Moderate to Severe Atopic Dermatitis (NCT03349060), a randomized phase 3 study. The primary outcome is Percentage of Participants Achieving Investigator's Global Assessment (IGA) Response of Clear (0) or Almost Clear (1) and Greater Than or Equal to 2 Points Improvement From Baseline at Week 12.

2. Clinical Question

The primary endpoint combines a final disease-severity state with improvement from baseline. It therefore asks whether participants reached a specified IGA category and had a minimum change in score, rather than relying on either the endpoint score or the change alone. The registry lists other measures of itch, symptoms, and skin response as secondary outcomes, but those do not replace the designated primary question.

3. Trial Design

This was a Phase 3 randomized, double-blind, placebo-controlled, parallel-group, multicenter study of PF-04965842 monotherapy in subjects aged 12 years and older with moderate to severe atopic dermatitis. The registry lists an enrollment of 387 and marks the study completed. The recorded start date is 2017-12-07, and the primary completion date is 2019-03-26.

The three interventions were PF-04965842 100 mg, PF-04965842 200 mg, and placebo. The brief summary describes both active doses as oral, once-daily treatment and states that efficacy and safety were evaluated relative to placebo over 12 weeks of study participation. Eligible patients could enter a long-term extension study after completing the treatment period; the registry record does not report extension results in the primary endpoint data.

The study is listed as randomized, with quadruple masking and a parallel model. The registry record does not explain the allocation procedure, the roles covered by quadruple masking, or the operational details of the multicenter conduct. Nor does it report the original sample-size justification, the planned statistical analysis in full, or whether interim analyses were planned.

The primary analysis population is described as the full analysis set, consisting of randomized participants who received at least 1 dose of study medication. For this outcome, the registry notes that the analyzed number means participants evaluable for the measure. The reported endpoint counts are 156 for PF-04965842 100 mg, 153 for PF-04965842 200 mg, and 76 for placebo. These outcome-specific analyzed counts should not be confused with the study’s enrollment figure; the record does not explain in the primary outcome description why the analyzed counts differ across groups.

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: This study was conducted from 07-December-2017 to 26-March-2019 at 76 sites in 8 countries.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: Full analysis set included all randomized participants who received at least 1 dose of study medication. Here, "Overall Number of Participants Analyzed" signifies participants evaluable for this outcome measure.

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: The full analysis set (FAS) included all randomized participants who received at least 1 dose of study medication.

6. Primary Endpoint

The registry lists the primary outcome as Percentage of Participants Achieving Investigator's Global Assessment (IGA) Response of Clear (0) or Almost Clear (1) and Greater Than or Equal to 2 Points Improvement From Baseline at Week 12, measured over Baseline, Week 12.

The primary endpoint is the percentage of participants achieving an Investigator’s Global Assessment response of clear or almost clear and at least a 2-point improvement from baseline at Week 12. The IGA rates atopic dermatitis severity on a 5-point scale from 0 to 4, with higher scores indicating greater severity. A score of 0 means clear, and a score of 1 means almost clear. The response definition requires both reaching one of those categories and improving by the specified amount from baseline.

This combined definition is useful because a participant’s final category and the amount of change from baseline describe related but different aspects of response. Requiring both makes the endpoint more specific than simply recording a low final severity score or any improvement. The scale’s category descriptions distinguish clear or almost-clear skin from mild, moderate, and severe disease. The assessment excluded the sole, palms, and scalp.

The registry identifies the primary outcome timeframe as baseline to Week 12 and reports it as a percentage of participants. Its primary outcome population is the full analysis set, defined as randomized participants who received at least 1 dose of study medication; the analyzed count for the measure refers to participants evaluable for it. The endpoint is a binary response summarized as a group percentage. It does not by itself describe the range of individual changes, the timing of response, or the course beyond the stated assessment.

7. Results Posted in the Registry

• PF-04965842 100 mg: 23.7% (37/156) versus 7.9% (6/76) on control; risk difference 15.8 percentage points (95% CI 6.8 to 24.8), risk ratio 3.00 (1.33–6.81); unadjusted two-proportion z-test p = 0.004.
• PF-04965842 200 mg: 43.8% (67/153) versus 7.9% (6/76) on control; risk difference 35.9 percentage points (95% CI 26.0 to 45.8), risk ratio 5.55 (2.52–12.21); unadjusted two-proportion z-test p = <0.001.

8. Statistical Interpretation

9. Statistical Methods Explained

The primary outcome is binary at the participant level: each evaluated participant either met the response definition or did not. A group percentage expresses the proportion meeting that definition. The registry reports the response percentages alongside the analyzed counts, allowing readers to see the event frequency and the denominator used for each group. A percentage alone does not convey how uncertain the estimate is, so the comparison is also reported with a confidence interval.

What does a difference in percentage mean?

A difference in percentage compares the response proportion in an active-treatment group with the response proportion in the placebo group. The sponsor’s analysis reports the contrast in percentage points. A positive difference in this comparison means the reported response proportion was higher in the active-treatment group than in placebo. This is a group-level contrast; it does not mean that every participant in one group had a different outcome from every participant in the other.

The posted re-calculation uses the displayed summary statistics and is unadjusted. Such a calculation is useful as a check on the relationship between the observed counts and the estimated contrast, but it need not reproduce an analysis that uses additional features of the design or analysis. The registry reports the sponsor’s method as Cochran-Mantel-Haenszel. The record does not specify the details needed to describe any stratification or adjustment beyond naming that method, so those details should not be inferred.

What does the Cochran-Mantel-Haenszel result describe?

As a general statistical explanation, the Cochran-Mantel-Haenszel approach is used to estimate or test an association across strata, with the stratum structure taken into account. In this trial, the registry identifies the method for the reported difference in percentage, but the registry record does not report the specific strata or provide enough detail to reconstruct the sponsor’s analysis. The reported estimate, interval, and p-value should therefore be understood as the sponsor’s stated analysis, rather than treated as a simple unadjusted calculation.

The sponsor labels the comparisons as superiority analyses. A superiority test evaluates whether the data are inconsistent with the null comparison specified for the analysis, in the direction of a difference. The registry reports p = 0.004 for the 100 mg comparison and p < 0.001 for the 200 mg comparison. Those p-values do not measure the size or practical importance of the difference. The corresponding estimates and confidence intervals are needed to describe magnitude and precision.

How should the confidence interval be understood?

In general, a confidence interval is a range generated by a statistical procedure that reflects the precision of an estimate under the model and assumptions used. A wider interval indicates less precision than a narrower interval on the same scale. The 95% intervals reported for the sponsor’s comparisons describe uncertainty in the estimated percentage-point differences. They are not ranges containing a fixed percentage of individual participant outcomes, and they do not mean there is a 95% probability that the true effect lies in the particular interval after it has been calculated.

Interpretation also depends on the population and outcome definition. Here, the primary analysis set consists of randomized participants who received at least 1 dose, and the analyzed count means participants evaluable for the endpoint. The registry record does not report the missing-data procedure or give a complete account of how evaluability was determined. Those omissions limit independent assessment of how missing outcome data were handled.

Why do binary outcomes need careful interpretation?

A binary endpoint simplifies a more detailed clinical assessment into whether a threshold was met. This makes a response percentage readily interpretable, but it discards information about how far participants who did not meet the response definition improved, and it treats participants just short of the threshold differently from those just meeting it. The endpoint also depends on the specified IGA categories and improvement requirement. A different definition or assessment time could produce a different summary, and the registry’s primary result should not be generalized to such alternatives.

For binary outcomes, confidence intervals and tests depend on the data and statistical method used. General statistical practice recognizes that sparse event counts can make estimates less precise and that the selected method matters. The placebo response count is 6, and the reported intervals reflect uncertainty around each comparison. The registry’s stated method should be kept distinct from the simpler unadjusted re-calculation; the latter is a check based on summary information, not a replacement for the trial analysis.

The outcome list includes other binary responses, such as itch improvement and skin response, alongside changes in symptom scores and time-to-event measures. These outcomes answer different questions and may use different analysis populations or time points. Their presence does not change the definition of the primary endpoint, and findings across multiple outcomes require attention to multiplicity and the analysis plan. The registry record does not report a multiplicity adjustment for the primary and secondary analyses.

10. Confidence Intervals and Precision

The sponsor reports a 95% confidence interval from 6.8 to 24.8 percentage points for the PF-04965842 100 mg comparison with placebo. For the PF-04965842 200 mg comparison, the reported 95% confidence interval is from 26.2 to 45.7 percentage points. These intervals accompany the reported differences of 15.8 and 36.0 percentage points, respectively.

A confidence interval conveys the precision of an estimate under the analysis used. It is not an interval for individual responses and does not guarantee that the true effect lies within the particular bounds. The 100 mg interval spans values above zero, as does the 200 mg interval. The interval for the 200 mg comparison is wider in absolute percentage-point terms than the interval for the 100 mg comparison, while the registry’s estimates also differ in magnitude. Comparing precision should therefore consider both the interval width and the scale of the estimated contrast, not just the endpoints in isolation.

The outcome-specific analyzed counts were 156 for 100 mg, 153 for 200 mg, and 76 for placebo. The registry record does not explain why these analyzed counts differ from the enrollment figure or provide a missing-data analysis for this endpoint. Those details matter because the number of evaluable outcomes and their handling contribute to uncertainty and can affect the interpretation of an estimate.

The p-values add evidence about compatibility with the null test, but they are not substitutes for the interval. A small p-value does not make an estimate precise by itself, and a confidence interval does not address every source of uncertainty, including limitations in the study conduct, endpoint definition, or reporting. The registry’s 95% intervals describe sampling uncertainty under the sponsor’s reported method; they do not resolve those broader limitations.

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 sample-size figures are planning illustrations based on the observed effect, not the original trial design. For these binary endpoint comparisons, the figures are participants per arm. The stated planning assumptions are a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation; each illustration uses the stated fraction of the observed effect to detect. The original trial’s own sample-size assumptions and justification are not reported in the registry record.

The column for 80% power gives the illustrative number of participants per arm under the stated assumptions for a design with an 80% chance of detecting the assumed effect if that effect were the true one. The 90% power column describes the corresponding planning target for a 90% chance. These are design calculations, not evidence that the completed trial had a particular probability of detecting an effect after the results were known. This is not observed power.

Rows based on 100% of the observed effect use the full observed contrast as the planning target; rows based on 75% or 50% use smaller assumed effects. The 100 mg comparison lists 83 participants per arm for 80% power and 110 for 90% power when planning for 100% of the observed effect; at 75%, the figures are 132 and 176, and at 50%, they are 261 and 349. The 200 mg comparison lists 23 and 30 for 100%, 36 and 47 for 75%, and 68 and 90 for 50%.

The pattern illustrates a general planning principle: smaller effects require more participants to detect under otherwise fixed assumptions. These calculations do not show that the trial was designed around the observed result, nor do they establish whether its original design was adequate. The record does not report the trial’s original design assumptions.

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 = 0.079; treatment proportion p₁ = 0.237; average p̄ = 0.158.
2. First term: z_1−α/2 × √[2p̄(1−p̄)] = 1.96 × 0.5159 = 1.0112.
3. Second term: z_1−β × √[p₀(1−p₀) + p₁(1−p₁)] = 0.842 × 0.5036 = 0.4239.
4. n = (1.0112 + 0.4239)² / (0.1582)² = 82.2, rounded up to 83 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 <- 0.0789473684210526                          # control proportion
p1 <- 0.237179487179487                          # 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(3349060)
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 = 83 (80% power), 110 (90% power)
 75% of observed effect: n per arm = 132 (80% power), 176 (90% power)
 50% of observed effect: n per arm = 261 (80% power), 349 (90% power)
power.prop.test: 82.2 
Simulated power at n = 83 per arm: 0.817 (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 sponsor’s reported analysis compares each PF-04965842 group with placebo using the Cochran-Mantel-Haenszel method. It reports a difference in percentage of 15.8 for 100 mg versus placebo, with a 95% confidence interval from 6.8 to 24.8 and p = 0.004. For 200 mg versus placebo, it reports a difference in percentage of 36.0, with a 95% confidence interval from 26.2 to 45.7 and p < 0.001. Both are identified as superiority analyses.

A separate re-calculation from the posted summary statistics is an unadjusted comparison. The sponsor’s method and the re-calculation are not identical descriptions of an analysis: the sponsor names Cochran-Mantel-Haenszel, while the re-calculation uses the posted summary statistics without adjustment. The registry’s discrepancy indicator is false, so no discrepancy is flagged. The re-calculation should be read as a secondary arithmetic check, not as a substitute for the sponsor-reported comparison.

The registry record does not report enough detail to reconstruct the sponsor’s full analysis, including any strata used. When describing the trial result, the sponsor’s reported estimates, confidence intervals, p-values, and named method are the appropriate figures to cite. The re-calculated comparison helps distinguish the simple summary-based calculation from the analysis reported by the sponsor; it does not add a separate randomized comparison.

14. Multiplicity, Missing Data and Analysis Populations

The primary endpoint is one outcome within a broader set of registered measures. The registry lists 45 outcome measures and 298 analyses, and the count of posted secondary outcomes is 43. Secondary measures include itch response at several visits, changes in the Pruritus and Symptoms Assessment for Atopic Dermatitis total score, time to itch improvement, Eczema Area and Severity Index response, and additional IGA outcomes. The different endpoints and time points address distinct questions.

When many outcomes or comparisons are examined, the chance of at least one apparently notable result can increase. As a general statistical principle, multiplicity procedures can control or describe that issue, depending on the analysis plan. The registry record does not report whether a multiplicity adjustment was applied, how outcomes were ordered for formal testing, or whether the primary comparison was protected within a prespecified testing strategy. The reported p-values should therefore be read as the stated comparison results, without assuming an unreported adjustment or testing hierarchy.

The primary population is described as the full analysis set, including randomized participants who received at least 1 dose of study medication. For the outcome, the analyzed number denotes those evaluable. Some secondary outcomes are also listed for a per protocol analysis set. These populations answer analyses based on different inclusion definitions, and the registry record does not give a complete account of every population used for the posted analyses.

The primary outcome description does not state how missing Week 12 assessments were handled, whether participants with missing data were counted as nonresponders, or what imputation procedure was used. The registry record does not report these details. It also does not report the original statistical analysis plan, interim analyses, or a sample-size justification. These absences restrict the ability to judge the full prespecification and robustness of the reported comparisons from the registry record alone.

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 5% reporting threshold.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: appendicitis (1.3%), meniscal degeneration (1.3%), asthma (1.3%), dermatitis atopic (1.3%), inflammatory bowel disease (0.6%).

The registry reports serious adverse events for the safety groups as counts of affected participants and participants at risk. Serious adverse events were reported for 5 of 156 participants in the PF-04965842 100 mg group, 5 of 154 in the 200 mg group, and 3 of 77 in the placebo group. The record lists 0 deaths in each group.

The serious-event terms listed include Appendicitis, Meniscal degeneration, Asthma, and Dermatitis atopic, each reported as 1.3%, and Inflammatory bowel disease and Peritonsillitis, each reported as 0.6%. The registry record does not establish that these events were caused by study treatment. These term percentages should be read as the registry’s reported figures, not as comparative risk estimates.

Serious adverse events are distinct from the broader category of other adverse events. The record reports 65 participants affected by other adverse events in each active-treatment group and 26 in the placebo group. The safety denominators differ from the primary endpoint analyzed counts, and the safety summaries should not be combined with the efficacy outcome counts. These descriptive figures do not by themselves establish a comparative safety profile; interpretation depends on event definitions, exposure, follow-up, and analyses that are not fully described in the registry record.

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

The posted primary results show the percentage meeting the specified Week 12 IGA response was 23.7% with PF-04965842 100 mg, 43.8% with PF-04965842 200 mg, and 7.9% with placebo. The sponsor-reported differences versus placebo were 15.8 percentage points and 36.0 percentage points, respectively, with the stated 95% confidence intervals and p-values. These are the results for this endpoint, population, and assessment time.

The result does not mean that every participant in an active-treatment group achieved the response, or that every placebo participant failed to achieve it. The group percentages describe the proportion meeting a threshold that combines the IGA category with improvement from baseline. Participants who did not meet the threshold may have had changes that the binary endpoint does not capture. The outcome also does not describe the durability of response beyond Week 12.

The registry record does not report full details of missing-data handling, multiplicity control, or the original statistical analysis plan. Those omissions limit assessment of how sensitive the estimates may be to analytic decisions and how the primary comparison fits within the full set of analyses. The trial’s results should not be extended automatically to other populations, treatment durations, or outcomes, including the listed secondary measures.

The study also reports serious adverse events and other adverse events, but the counts alone do not establish causation or settle comparative safety. The registry record does not report detailed exposure-adjusted comparisons in the information. Overall, the statistical findings describe the observed trial groups and the sponsor’s specified comparison; they are not medical advice and do not determine what an individual should do.

17. Why This Trial Matters Statistically

This trial illustrates how a clinically defined response can be translated into a binary statistical endpoint. The primary outcome required both a clear or almost-clear IGA score and at least a 2-point improvement from baseline. That combination is more informative than reporting only a final category or only a change score, but it also creates a threshold: people just on either side of the definition are classified differently even if their underlying assessments are similar.

The three-arm design permits separate comparisons of two dose strengths with placebo. The registry reports each active-dose comparison independently, with group response percentages, differences, confidence intervals, and p-values. Looking at estimates and intervals alongside p-values is important because statistical evidence against a null comparison does not by itself communicate the size or precision of the estimated contrast.

The registry also makes visible the distinction between sponsor analysis and a simple re-calculation. The sponsor reports a Cochran-Mantel-Haenszel analysis, while the re-calculation is unadjusted and based on posted summary statistics. Keeping those approaches distinct is central to sound interpretation, particularly when the full details of the sponsor’s analysis are not reported.

Finally, this record demonstrates why efficacy results must be read alongside analysis-population definitions, missing-data reporting, multiplicity, and safety summaries. The primary endpoint has a defined full analysis set, but important details about missing outcomes and the overall analysis plan are not reported in the registry record. Statistical interpretation is strongest when estimates are read together with those design and reporting limits rather than in isolation.

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.
• Participant counts were reconstructed from posted percentages and arm sizes, so small rounding differences are possible.

The primary outcome is a threshold-based binary response at Week 12. It does not preserve the full range of IGA changes, and the registry’s reported primary result does not describe response after that visit. The IGA scale excludes the sole, palms, and scalp, according to the endpoint description, so the primary assessment does not cover those sites.

The registry record defines the primary analysis population as randomized participants who received at least 1 dose and says the analyzed count represents participants evaluable for the outcome. It does not explain why the group-specific outcome counts differ from one another or from the enrollment figure. It also does not report the missing-data method, reasons for unevaluable outcomes, or sensitivity analyses for missing observations.

The sponsor’s named Cochran-Mantel-Haenszel method is reported, but the registry record does not provide the strata or enough details to reconstruct the full analysis. The record also does not report the original sample-size justification, the full statistical analysis plan, interim analyses, or whether multiplicity was controlled across the many outcomes and analyses. These are reporting limitations; their absence does not establish that such procedures were not used.

The trial evaluated two dose strengths against placebo over the stated treatment period. Although the brief summary says eligible patients could enter a long-term extension study, results from that extension are not included in the primary outcome information. The registry’s adverse-event summaries are descriptive counts and do not, on their own, establish causality or provide a complete comparative safety analysis.

19. Related Tutorials

20. Related Statistical Calculators

21. Sources

• ClinicalTrials.gov: Study to Evaluate Efficacy and Safety of PF-04965842 in Subjects Aged 12 Years And Older With Moderate to Severe Atopic Dermatitis — NCT03349060.
• Registry record last updated 2019-12-10; status last verified 2019-11.
• Results first posted 2019-12-10.
• 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

NCT03349060 is a completed Phase 3 randomized, double-blind, placebo-controlled, parallel-group, multicenter study of PF-04965842 monotherapy for moderate to severe atopic dermatitis in subjects aged 12 years and older. The registry lists enrollment of 387 and a 12-week comparison of PF-04965842 100 mg, PF-04965842 200 mg, and placebo.

The primary endpoint was the percentage achieving a clear or almost-clear IGA response with at least a 2-point improvement from baseline at Week 12. Reported response percentages were 23.7% for 100 mg, 43.8% for 200 mg, and 7.9% for placebo. Sponsor-reported comparisons versus placebo used the Cochran-Mantel-Haenszel method. The registry record does not report the original sample-size justification, missing-data strategy, or multiplicity approach.