Study of the Efficacy and Multiple-Dose Plasma Concentration-Time Profiles of Armodafinil and PROVIGIL: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Study of the Efficacy and Multiple-Dose Plasma Concentration-Time Profiles of Armodafinil and PROVIGIL (NCT00236080), a randomized phase 3 study. The primary outcome is Multiple Sleep Latency Test (MSLT).

2. Clinical Question

The clinical question is about group-level differences in a continuous sleepiness measure over the study period. It is not a question about the effect on every individual participant, nor does a difference in group means establish how a participant’s experience changed. The registry gives mean changes and standard deviations by arm, along with an overall sponsor analysis and recalculated, unadjusted comparisons against placebo. Those results address different statistical questions and should not be treated as interchangeable.

3. Trial Design

The registry describes a Phase 3 randomized, double-blind, parallel-group, placebo-controlled study sponsored by Cephalon. The condition under study was chronic Shift Work Sleep Disorder. The record lists enrollment as 136 and status as completed. The study start is recorded as 2005-08; a completion date is not reported in the registry record.

Five arms are listed: PROVIGIL 200 mg/day, armodafinil 250 mg/day, armodafinil 200 mg/day, armodafinil 150 mg/day, and placebo. The primary outcome data give an analyzed group size for each arm, but the record does not provide a complete account of how those outcome populations relate to all enrolled participants. In particular, the endpoint description notes a placebo patient who completed other requirements without an MSLT, and a patient in the armodafinil 200 mg/day group who had an MSLT but discontinued study drug before endpoint. The latter description is cut off in the registry record, so the subsequent handling of that patient cannot be established from this entry.

Randomization and double masking describe important design features intended to reduce systematic differences in treatment assignment and assessment. A parallel-group design means participants were assigned to one of the listed arms rather than receiving a sequence of study treatments. The registry does not report the randomization method, the masking procedures in detail, or the original statistical design assumptions for the trial. It also does not report the planned sample-size justification or whether interim analyses were conducted. The planning illustrations derived from the observed data are separate from those original design decisions.

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: 2 male participants withdrew after randomization but prior to receiving study drug (1 for noncompliance and 1 at the request of the sponsor)

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: * 1 Placebo Patient did not have an MSLT but did complete the other Primary Measure (PVT) and other requirements. This patient was termed a "Completer". * 1 Patient in the "Armodafinil 200 mg/day group" had an MSLT performed but then discontinued the study drug before reaching the study endpoint and was termed a "Non-Completer" for the 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 Multiple Sleep Latency Test (MSLT), measured over Endpoint (Visit 4) change from baseline (Visit 2).

The primary endpoint was change from baseline in the Multiple Sleep Latency Test (MSLT) at endpoint, measured in minutes. Baseline was Visit 2 and endpoint was Visit 4. The MSLT is described as an objective assessment of sleepiness that measures the likelihood of falling asleep. At each assessment, five scheduled naps were conducted, each lasting up to 20 minutes. The scheduled times were 2300, 0100, 0300, 0500, and 0700. A nap ended after 20 minutes if sleep did not occur. Sleep latency was the elapsed time from lights out to the first epoch scored as sleep.

Using a continuous measure allows the analysis to compare average changes across randomized groups while retaining information about the magnitude of measured sleep latency rather than reducing it to a yes-or-no outcome. A positive mean change indicates an increase in sleep latency from baseline on this measure; it describes the average recorded change in the group, not the change for every participant.

The registry reports the MSLT outcome as a mean with standard deviation. Standard deviation describes the spread of participant measurements around the group mean. The endpoint is relevant to the study’s stated question about overnight sleepiness, but it is only one measured outcome. The registry does not post results for the plasma concentration-time profiles mentioned in the study purpose, and no secondary results are posted.

7. Results Posted in the Registry

• PROVIGIL 200 mg/Day: mean 2.00 versus 1.10 on control; difference 0.90 (95% CI -0.97 to 2.77), standardized effect d = 0.26; z-test p = 0.345.
• Armodafinil 250 mg/Day: mean 3.70 versus 1.10 on control; difference 2.60 (95% CI 0.40 to 4.80), standardized effect d = 0.64; z-test p = 0.021.
• Armodafinil 200 mg/Day: mean 3.70 versus 1.10 on control; difference 2.60 (95% CI 0.06 to 5.14), standardized effect d = 0.59; z-test p = 0.045.
• Armodafinil 150 mg/Day: mean 2.70 versus 1.10 on control; difference 1.60 (95% CI -0.67 to 3.87), standardized effect d = 0.40; z-test p = 0.168.

8. Statistical Interpretation

9. Statistical Methods Explained

The primary measure is continuous: it is recorded in minutes, and the registry reports the mean and standard deviation for each arm. A mean is a summary of the group’s average observed change. It can conceal differences among participants, so it is useful to read it alongside the standard deviation, which describes how dispersed the measurements are. A larger standard deviation means the individual measurements are more spread out around the average; it does not by itself show whether the average is precisely estimated.

What does a mean difference represent?

The recalculated comparisons express the difference between an active-treatment group’s mean change and the placebo group’s mean change. A positive estimate means the active group’s reported mean change was larger. For example, the recalculated estimate for armodafinil 250 mg/day is 2.6 minutes, and the estimate for PROVIGIL 200 mg/day is 0.9 minutes. These contrasts are simple, unadjusted comparisons from the posted group summaries. They do not account for covariates or reproduce the sponsor’s ANCOVA.

The recalculations use the reported group sizes and standard deviations to estimate the uncertainty around each mean difference. This is distinct from the sponsor’s reported overall analysis, which is identified as ANCOVA. In general, ANCOVA compares outcomes while adjusting for specified covariates, often including a baseline measure. The registry record does not identify the covariates used, provide the model specification, or report adjusted effect estimates. It is therefore not possible to determine how the sponsor’s adjusted estimates compare numerically with the simple recalculations.

How should a confidence interval be read?

A confidence interval expresses uncertainty about an estimated difference under the method used to calculate it. It gives a range of values compatible with the data and the calculation’s assumptions; it is not a range containing a specified proportion of individual participant outcomes. A narrower interval indicates greater precision than a wider interval, while an interval that includes zero is compatible with no mean difference under the interval procedure. These general interpretations do not resolve clinical importance, which requires context and a meaningful-change criterion that is not reported in the registry record.

The recalculated confidence intervals are reported with the unadjusted estimates. They should not be mistaken for intervals from the sponsor’s ANCOVA, because the registry does not report sponsor confidence limits. The interval for armodafinil 200 mg/day extends from 0.057 to 5.143 minutes, while that for armodafinil 250 mg/day extends from 0.4 to 4.8 minutes. The width of an interval depends on both the variability in the data and the amount of information available for that contrast. General statistical reasoning says that more variability or less information tends to produce less precise estimates.

Why does the analysis method matter?

The registry identifies the sponsor’s primary analysis as ANCOVA and gives p = 0.124 for an analysis involving all five groups, categorized as non-inferiority or equivalence. An overall analysis across groups addresses a different question from separate active-treatment-versus-placebo contrasts. The registry does not give the ANCOVA’s contrast estimates, its confidence intervals, or its non-inferiority or equivalence margin. Without those details, the result cannot be fully interpreted as a test of a particular margin or reconstructed from the arm summaries.

The recalculated pairwise p-values are unadjusted. When multiple comparisons are examined, the chance of at least one apparently unusual result can increase if no adjustment is made. This is the general reason multiplicity matters. The registry does not state whether the reported analysis used a multiplicity procedure, nor does it report an adjustment for the recalculated contrasts. The p-values should therefore be read as individual, unadjusted results rather than as a complete family-wise inference.

General statistical interpretation also distinguishes evidence against a null hypothesis from evidence of equivalence. A non-significant difference alone does not demonstrate that treatments are equivalent; equivalence or non-inferiority conclusions depend on a prespecified margin, an appropriate design, and suitable estimates and confidence intervals. The margin and supporting estimates are not reported in the registry record. Accordingly, the recorded analysis classification and p-value can be described, but its clinical or formal inferential conclusion cannot be independently assessed from the registry entry.

What can the summaries not establish?

Summary statistics do not reveal the distribution of individual changes, the number of participants with particular responses, or whether the results were affected by missing measurements. The endpoint notes indicate that at least one placebo participant completed other requirements without an MSLT and that an armodafinil 200 mg/day participant discontinued study drug before endpoint after having an MSLT. The remaining description is incomplete, and the registry does not fully describe the analysis population or missing-data procedure. These omissions limit assessment of how the reported outcome groups were formed.

The registry gives no secondary outcome results, no details of interim analyses, and no original sample-size justification. Those items should not be inferred from the posted summaries or from the planning illustrations. The illustrations based on observed effects answer a hypothetical planning question under stated assumptions; they do not report the trial’s design calculations or its observed probability of detecting an effect.

10. Confidence Intervals and Precision

The recalculated mean differences are accompanied by confidence intervals that show the uncertainty around each unadjusted comparison with placebo. For PROVIGIL 200 mg/day, the estimate is 0.9 minutes and the interval runs from -0.970 to 2.77. For armodafinil 250 mg/day, the estimate is 2.6 minutes and the interval runs from 0.4 to 4.8. For armodafinil 200 mg/day, the estimate is 2.6 minutes and the interval runs from 0.057 to 5.143. For armodafinil 150 mg/day, the estimate is 1.6 minutes and the interval runs from -0.675 to 3.875.

The intervals differ in width, reflecting different levels of precision for the separate comparisons. The intervals for PROVIGIL 200 mg/day and armodafinil 150 mg/day include zero; the intervals for armodafinil 250 mg/day and armodafinil 200 mg/day do not. This describes the recalculated intervals, not the sponsor’s adjusted ANCOVA results. The registry does not report confidence intervals for the sponsor’s analysis.

Precision should not be confused with clinical importance. A confidence interval can be informative about the range of differences compatible with the data and still leave uncertainty about whether values in that range matter in practice. The registry does not report a threshold for a clinically meaningful change in MSLT. Nor does the confidence interval account for every limitation, such as missing endpoint measurements, model choices, or multiplicity. The posted interval estimates are unadjusted recalculations and should be interpreted in that context.

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 participants per arm for each active-treatment-versus-placebo comparison, not total events. The calculations assume a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. The effect used for each illustration is the stated fraction of that comparison’s observed effect: 100%, 75%, or 50%. The original trial’s own sample-size assumptions are not reported in the registry record.

The 80% power column gives the estimated participants per arm for a hypothetical study planned to have an 80% chance of detecting the specified effect under those assumptions, if that effect and the calculation assumptions were correct. The 90% power column gives the corresponding planning figure for a higher target chance of detecting that effect. These are planning targets, not guarantees about a future study. The listed figures differ by comparison because each illustration is based on that comparison’s observed effect and variability.

Smaller effects require more participants to detect under otherwise similar assumptions. Thus, the rows planning for 75% or 50% of the observed effect call for larger group sizes than planning for the full observed effect. These illustrations are not observed power: they are hypothetical sample-size calculations based on the observed estimates, not a measure of the probability that this completed trial detected a true effect. They also do not establish that the observed effects will recur in another study.

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, Δ is the difference in means to detect and σ is the common standard deviation.

Worked calculation for this trial

1. Observed difference Δ = 0.900; pooled standard deviation σ = 3.409.
2. n = 2 × (1.96 + 0.842)² × 3.409² / 0.900² = 225.3, rounded up to 226 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
delta <- 0.9                      # observed difference in means
sigma <- 3.40930315460506                     # pooled standard deviation
n_per_arm <- function(power, frac = 1) {
  ceiling(round(2 * (z_a + qnorm(power))^2 * sigma^2 / (delta * frac)^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 (t-based, so very slightly larger)
cat("power.t.test:", round(power.t.test(delta = delta, sd = sigma, power = 0.80)$n, 1), "\n")

# Simulation: exact two-sample t-test using simulated means and variances
set.seed(236080)
n    <- n_per_arm(0.80)
reps <- 20000
m_diff <- rnorm(reps, delta, sigma * sqrt(2 / n))             # difference in sample means
s2     <- sigma^2 * rchisq(reps, 2 * (n - 1)) / (2 * (n - 1))   # pooled sample variance
tstat  <- m_diff / sqrt(s2 * 2 / n)
cat(sprintf("Simulated power at n = %s per arm: %.3f (%d replicates)\n", format(n, big.mark = ","), mean(abs(tstat) > qt(0.975, 2 * (n - 1))), reps))

Output from running the code

100% of observed effect: n per arm = 226 (80% power), 302 (90% power)
 75% of observed effect: n per arm = 401 (80% power), 537 (90% power)
 50% of observed effect: n per arm = 902 (80% power), 1,207 (90% power)
power.t.test: 226.2 
Simulated power at n = 226 per arm: 0.805 (20000 replicates)

The simulation draws many sets of group means and variances at the planned sample size and counts how often the two-sample t-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 is an ANCOVA across the five groups, with p = 0.124, and is classified in the registry as non-inferiority or equivalence. The registry does not provide a corresponding adjusted estimate, confidence interval, model details, or the margin used for a non-inferiority or equivalence assessment. The reported p-value alone is therefore insufficient to reconstruct the sponsor’s conclusion or understand the magnitude and precision of the adjusted contrasts.

The separate recalculated results are unadjusted comparisons of each active arm with placebo, using posted group means, standard deviations, and group sizes. Their estimates and p-values do not reproduce the ANCOVA. The registry record indicates a discrepancy between the reported analysis and the recalculations; the sponsor’s own analysis takes precedence. The recalculations remain useful as descriptive, transparent comparisons of the posted summaries, but should not replace the sponsor’s analysis or be presented as its adjusted results.

The registry does not report enough information to identify the precise source of the discrepancy. It would be inappropriate to attribute it to a particular model, analysis population, missing-data method, or adjustment without those details.

14. Multiplicity, Missing Data and Analysis Populations

The registry provides one primary outcome measure, MSLT change from baseline, and an overall sponsor analysis involving the five groups. The record also indicates two outcome measures and two analyses, but no secondary results are posted. The registry does not report a complete hierarchy of hypotheses, a prespecified testing sequence, or a multiplicity-adjustment procedure for the sponsor analysis.

The recalculated results compare four active arms separately with placebo. Each p-value is unadjusted. General statistical practice treats multiple tests as a multiplicity concern because examining several contrasts can increase the chance of at least one small p-value even when no true difference exists. The unadjusted p-values do not account for that issue. The registry does not say whether these particular contrasts were prespecified or how they relate to the sponsor’s overall analysis.

Information about missing data and analysis populations is limited. The primary outcome description notes that one placebo patient did not have an MSLT but completed other requirements, and that one patient in the armodafinil 200 mg/day group had an MSLT but discontinued study drug before reaching endpoint. The description of the latter case is incomplete. The record does not fully specify whether those observations were included in the endpoint summaries, how missing endpoint values were handled, or whether the primary analysis used an intention-to-treat, modified intention-to-treat, or per-protocol population.

The registry also does not report sensitivity analyses addressing missing outcomes, details of protocol deviations, or how the analyzed groups were defined. These omissions matter because different analysis populations or missing-data assumptions can affect estimates. They prevent a full appraisal of whether the posted summary statistics and the sponsor’s ANCOVA use the same participants and endpoint handling. No inference about the impact of missingness or multiplicity should be made without those details.

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 5% reporting threshold.

Most frequent non-serious adverse event terms

The serious adverse event record lists zero serious adverse events reported for each treatment group. The corresponding numbers of participants at risk are 29 for PROVIGIL 200 mg/day, 28 for armodafinil 250 mg/day, 26 for armodafinil 200 mg/day, 25 for armodafinil 150 mg/day, and 26 for placebo. These are group-specific safety figures and should be read separately from the primary endpoint group sizes.

A zero in the serious adverse event count means that no serious adverse events were reported in the listed group during the recorded safety reporting period. It does not establish that the risk of a serious event is absent, nor does it by itself allow precise comparison of risks across groups. The registry record does not give a serious-event term list, and it does not report deaths in these fields. It also lists other adverse events, but those counts do not replace a full safety analysis.

Safety interpretation is limited by the registry’s reporting detail. The record does not provide event timing, severity categories beyond the serious-event classification, exposure duration, or a narrative assessment of relatedness. These figures describe the posted serious adverse event counts only and should not be used alone to make a broader safety conclusion.

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

The posted results show positive mean MSLT changes from baseline in every listed arm. The largest reported group means are 3.7 minutes in both armodafinil 250 mg/day and armodafinil 200 mg/day; placebo has a mean change of 1.1 minutes. The unadjusted recalculated differences versus placebo are positive for all active arms. The confidence intervals for armodafinil 250 mg/day and armodafinil 200 mg/day exclude zero, while the intervals for PROVIGIL 200 mg/day and armodafinil 150 mg/day include zero.

These statements describe the posted summaries and recalculations. They do not establish that a treatment works or does not work, and they do not establish equivalence among treatments. The sponsor’s analysis is the reported ANCOVA with p = 0.124, classified as non-inferiority or equivalence; the information needed to evaluate that analysis fully, including its margin and effect estimates, is not reported in the registry record. Because the recalculated comparisons are unadjusted and differ from the sponsor’s analysis, the sponsor’s analysis takes precedence.

The measured endpoint is an objective sleep-latency test, not a direct measure of every aspect of participants’ functioning or experience. The registry does not report a threshold for interpreting how large a change is meaningful. The posted group averages also do not describe each participant’s response, and the standard deviations show that measurements varied around those averages.

The study purpose refers to plasma concentration-time profiles as well as overnight efficacy, but plasma results are not posted in the registry record. The record also lacks full detail about missing-data handling, analysis populations, multiplicity, and the original sample-size rationale. The conclusions that can be drawn are therefore limited to the reported outcome summaries, the sponsor’s stated overall analysis, and the clearly labeled unadjusted recalculations.

17. Why This Trial Matters Statistically

This trial illustrates why a clinical-trial result should not be reduced to a single p-value. The registry gives group means and standard deviations for a continuous endpoint, an overall sponsor ANCOVA result, and recalculated pairwise comparisons with confidence intervals. Those analyses answer related but distinct questions. Reading them together shows both the estimated magnitude of the posted group differences and the limits of what can be inferred from the record.

It also demonstrates the role of precision. The estimated differences are not equally precise: their confidence intervals have different widths, and some include zero while others do not. Even when an interval excludes zero, the estimate is not a measure of clinical importance. A meaningful-change threshold is not reported, and the sponsor’s adjusted estimates are unavailable. Statistical evidence, precision, and clinical relevance are separate considerations.

The discrepancy between the sponsor’s overall ANCOVA and simple unadjusted recalculations is another important lesson. Reconstructing comparisons from summary statistics can be useful for understanding the posted arm results, but it does not substitute for a prespecified adjusted analysis. The sponsor’s own result takes precedence, while missing model details limit independent appraisal.

Finally, planning calculations based on observed effects should be kept separate from a trial’s original design and from observed power. The registry does not report the original sample-size assumptions. The illustrations show how hypothetical planning needs change when the effect targeted is smaller, but they do not show what the original investigators planned or establish the probability that the completed trial detected an effect.

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 of interpretation arise from the registry record. Although the study is listed as randomized and double-blind, the record does not describe the randomization procedure, masking implementation, or allocation concealment. The original sample-size justification, interim analysis plans, and detailed statistical plan are also not reported. The listed enrollment does not by itself clarify which participants contributed to each endpoint analysis.

The primary endpoint description is incomplete regarding a participant in the armodafinil 200 mg/day group who discontinued study drug before reaching endpoint. It also notes a placebo participant who completed other study requirements but did not have an MSLT. The registry does not fully describe the analysis population, missing-data handling, or the effect of these cases on the endpoint summaries. The available outcome group sizes should therefore not be assumed to represent every randomized participant or a single consistently defined analysis population.

The sponsor’s ANCOVA result lacks adjusted estimates, confidence limits, model details, and a reported non-inferiority or equivalence margin. The recalculations are unadjusted and are not a replacement for that analysis. The record does not explain the discrepancy between them. The handling of multiple comparisons is also not reported, and no secondary outcome results are posted.

The primary outcome is a particular objective measure of sleepiness; the registry does not establish a threshold for clinical importance. The study purpose includes plasma concentration-time profiles, but those results are not reported. The safety record gives serious adverse event counts but not enough detail to assess timing, exposure, or broader safety patterns. These limitations constrain interpretation to the outcomes and information actually recorded.

19. Related Tutorials

20. Related Statistical Calculators

21. Sources

• ClinicalTrials.gov: Study of the Efficacy and Multiple-Dose Plasma Concentration-Time Profiles of Armodafinil and PROVIGIL — NCT00236080.
• Registry record last updated 2013-07-19; status last verified 2013-07.
• Results first posted 2010-02-03.
• 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

NCT00236080 is a completed Phase 3 randomized, double-blind, parallel-group study sponsored by Cephalon in patients with chronic Shift Work Sleep Disorder. It compared placebo with PROVIGIL 200 mg/day and armodafinil at 150 mg/day, 200 mg/day, and 250 mg/day. The primary outcome was mean change from baseline to endpoint in MSLT sleep latency, measured in minutes.

The registry reports positive mean changes across all arms and an overall sponsor ANCOVA result of p = 0.124, classified as non-inferiority or equivalence. Unadjusted recalculations compare each active arm with placebo, but the discrepancy means the sponsor’s own analysis takes precedence. The registry does not report the sponsor’s adjusted estimates or the margin needed to evaluate the analysis fully. Plasma concentration-time results and secondary results are not posted, and endpoint missing-data and analysis-population details are incomplete.