A Study to Evaluate the Efficacy and Safety of SEP-363856 in Acutely Psychotic Adults With Schizophrenia: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for A Study to Evaluate the Efficacy and Safety of SEP-363856 in Acutely Psychotic Adults With Schizophrenia (NCT02969382), a randomized phase 2 study. The primary outcome is Change From Baseline in Positive and Negative Syndrome Scale (PANSS) Total Score at Week 4.

2. Clinical Question

The primary endpoint is a between-group comparison of change in PANSS total score over the study period. PANSS measures symptoms using 30 items, which contribute to positive, negative, and general psychopathology subscales and a total score. The registry describes higher total scores as indicating more severe outcomes. Accordingly, the posted negative changes represent reductions in score from baseline. The key statistical question is whether the estimated average change differed between the randomized groups, and how uncertain that estimate is.

The registry record identifies the trial as randomized, parallel-group, placebo-controlled, and quadruple-masked; its official title also describes it as double-blind. The record does not specify which people or roles were included in the masking designation. The tables present the reported results alongside straightforward summaries of the posted group estimates. Those summaries help describe the size and precision of the observed contrast, but they do not establish what would happen in other populations or settings.

3. Trial Design

This was a phase 2, randomized, parallel-group study of SEP-363856 versus placebo in acutely psychotic adults with schizophrenia. The official title describes a four-week, multicenter, flexibly dosed study. The registry lists 245 participants enrolled: 120 in the SEP-363856 group and 125 in the placebo group. Participants assigned to SEP-363856 received a 50 mg or 75 mg capsule once daily; the comparator group received a placebo capsule once daily.

The registry records quadruple masking, while the official title calls the study double-blind. The record does not identify the specific parties masked or explain the relationship between those descriptions. It lists randomization and a parallel-group model, but details such as the randomization procedure, allocation concealment, site count, eligibility criteria, and dose-adjustment rules are not reported in the registry record.

The primary outcome was change from baseline in PANSS total score at Week 4. The posted primary analysis is described as a mixed models analysis and reports a mean difference, confidence interval, and p-value. The registry also lists secondary outcomes covering clinical severity, PANSS subscales and response, negative symptoms, depression symptoms, and adverse events. The record lists 12 outcome measures and 11 analyses, although detailed numerical comparisons in the posted results represented are available for the primary endpoint and three secondary endpoints.

The study status is completed. The record gives a start date of 2016-12-05 and a primary completion date of 2018-07-31. The sponsor is Otsuka Pharmaceutical Development & Commercialization, Inc. The original design’s sample-size rationale, planned analysis populations, missing-data procedures, interim analyses, and multiplicity strategy are not reported in the registry record.

4. Participant Flow and Accounting

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

Overall Study

Reasons for not completing this period:

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 Change From Baseline in Positive and Negative Syndrome Scale (PANSS) Total Score at Week 4, measured over Baseline, Week 4.

The primary endpoint is change from baseline in PANSS total score at Week 4. PANSS is a structured measure composed of 30 items, each scored on an anchored Likert scale from 1 - 7. The registry states that values of 2 and above indicate the presence of progressively more severe symptoms. Item scores contribute to positive, negative, and general psychopathology subscales, as well as a total score. The listed range for the total score is 30-210, and higher total scores mean a more severe outcome.

Change from baseline summarizes how the score at the follow-up visit differs from the participant’s starting score. The registry reports group least-squares means for this change and a between-group mean difference. The least-squares mean is an estimate from a statistical model; it is not necessarily the same thing as a simple arithmetic average. The registry does not report enough model detail to explain precisely how its estimates were adjusted or which observations contributed to the model.

Using a symptom scale as the primary endpoint allows the study to compare average measured symptom change across randomized groups over the stated follow-up. A negative change on this scale means the measured score was lower than at baseline; because the registry says higher PANSS scores indicate more severe outcomes, lower scores correspond to lower measured severity on this scale. This endpoint is a score-based measure, not a complete account of each participant’s experience, functioning, or longer-term course. The four-week result also addresses the recorded time point, not outcomes beyond it.

7. Results Posted in the Registry

• SEP-363856: mean -17.20 versus -9.70 on control; difference -7.50 (95% CI -12.03 to -2.97), standardized effect d = -0.41; z-test p = 0.001.

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.

Change From Baseline in Clinical Global Impression - Severity (CGI-S) Score at Week 4

• Endpoint type: Continuous: a mean value per participant, compared as a difference in means.
• Test: Unadjusted two-sample z-test for the difference in means, using the posted standard errors.
• Result:
  • SEP-363856: mean -1.00 versus -0.50 on control; difference -0.50 (95% CI -0.75 to -0.25), standardized effect d = -0.50; z-test p = <0.001.
• Interpretation:
  • The 95% confidence interval excludes no difference, so the mean is statistically lower for SEP-363856 than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.

Change From Baseline in Positive and Negative Syndrome Scale (PANSS) Positive Subscale Score at Week 4

• Endpoint type: Continuous: a mean value per participant, compared as a difference in means.
• Test: Unadjusted two-sample z-test for the difference in means, using the posted standard errors.
• Result:
  • SEP-363856: mean -5.50 versus -3.90 on control; difference -1.60 (95% CI -3.04 to -0.16), standardized effect d = -0.28; z-test p = 0.030.
• Interpretation:
  • The 95% confidence interval excludes no difference, so the mean is statistically lower for SEP-363856 than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.

Change From Baseline in Positive and Negative Syndrome Scale (PANSS) Negative Subscale Score at Week 4

• Endpoint type: Continuous: a mean value per participant, compared as a difference in means.
• Test: Unadjusted two-sample z-test for the difference in means, using the posted standard errors.
• Result:
  • SEP-363856: mean -3.10 versus -1.60 on control; difference -1.50 (95% CI -2.65 to -0.35), standardized effect d = -0.33; z-test p = 0.011.
• Interpretation:
  • The 95% confidence interval excludes no difference, so the mean is statistically lower for SEP-363856 than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.

The detailed secondary results cover change in CGI-S and PANSS positive and negative subscale scores at Week 4. For each of these measures, the posted group estimates indicate a larger average reduction in the SEP-363856 group than in the placebo group. The sponsor’s reported analyses give between-group differences and p-values for these comparisons. These outcomes are related to the primary result because they assess clinical severity or components of the symptom domains captured by PANSS.

The overall direction across these posted secondary comparisons is therefore consistent with the primary endpoint’s direction. That consistency can provide context, but it does not make the secondary results equivalent to the primary result. Secondary measures address different aspects of symptoms, and they are not independent replications of the primary outcome. The registry lists a wider set of secondary outcomes, including negative symptoms, depression symptoms, PANSS response, and adverse events; detailed numerical results for all of those outcomes are not reported in the registry record represented.

The secondary p-values are unadjusted in the posted results. When several outcomes are tested, the chance of observing at least one small p-value by chance can rise. Without a reported testing hierarchy or multiplicity adjustment, individual secondary p-values carry less confirmatory weight than a prespecified primary analysis. The registry record does not report how multiplicity was handled in the original statistical plan, so no adjustment should be assumed.

These secondary outcomes are best read as supportive descriptive evidence, alongside their estimates and uncertainty, rather than as a separate set of definitive conclusions. Their interpretation also depends on the analysis details, missing-data handling, and measurement properties, which are not fully described in the registry record.

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

• Change From Baseline in Clinical Global Impression - Severity (CGI-S) Score at Week 4
• Change From Baseline in Positive and Negative Syndrome Scale (PANSS) Positive Subscale Score at Week 4
• Change From Baseline in Positive and Negative Syndrome Scale (PANSS) Negative Subscale Score at Week 4
• Change From Baseline in Positive and Negative Syndrome Scale (PANSS) General Psychopathology Subscale Score at Week 4
• Change From Baseline in Brief Negative Symptom Scale (BNSS) Total Score at Week 4
• Change From Baseline in Montgomery-Asberg Depression Rating Scale (MADRS) Total Score at Week 4
• Positive and Negative Syndrome Scale (PANSS) Response at Week 4, Defined as a 20% or Greater Improvement From Baseline in PANSS Total Score
• The Incidence of Overall AEs, Serious AEs (SAEs) and AEs (or SAEs) Leading to Discontinuation

9. Statistical Interpretation

10. Statistical Methods Explained

What does the mixed models analysis report?

The registry describes the primary analysis as a mixed models analysis and reports a mean difference (net). A model-based mean difference compares the estimated average change between groups, while a mixed model is a general class of methods that can account for structured data, such as repeated measurements. That is general statistical background; the registry record does not report the exact model, its included terms, covariance structure, or estimation details. It is therefore not possible from the posted record to reconstruct the sponsor’s full analysis.

The reported primary contrast is -7.5, with a 95% confidence interval from -11.9 to -3.0 and p = 0.001. The group estimates are least-squares means, which are model-based estimates rather than necessarily raw averages. The record does not describe whether or how the model adjusted for baseline score, site, or other variables. Do not assume those details from the phrase “mixed models analysis.”

What is a confidence interval?

General explanation: a confidence interval is a range of estimates compatible with the data and the statistical method under its assumptions. It communicates precision as well as direction. A narrower interval indicates less uncertainty in the estimate than a wider interval on the same scale. A 95% confidence interval does not mean there is a 95% probability that the particular interval contains the true effect; it describes a long-run property of the interval-producing procedure under the model assumptions.

For the primary result, the posted sponsor interval extends from -11.9 to -3.0. Its location below zero is consistent with a negative estimated difference on the change-score scale. Since the scale’s higher values indicate more severe outcomes, this direction corresponds to a larger average score reduction in the SEP-363856 group. The interval does not describe the full range of individual participant responses.

How should the p-value be read?

General explanation: a p-value summarizes how unusual results at least as incompatible with a specified null hypothesis would be under that hypothesis and the analysis assumptions. It is not the probability that the null hypothesis is true, nor does it measure the size or clinical importance of an effect. The primary reported p-value is 0.001. The registry does not report enough information to assess every modeling assumption or the full analysis plan.

The secondary results include p-values for separate outcomes. An unadjusted p-value evaluates an individual comparison without accounting for the fact that several outcomes may have been examined. When many tests are considered, the chance of at least one apparently unusual result can increase. General statistical practice may address that issue through a prespecified hierarchy or adjustment, but the strategy for this trial is not reported in the registry record. The secondary p-values should therefore be interpreted as individual, unadjusted results.

What is the role of the re-calculation?

The re-calculation uses the posted group least-squares means, standard errors, and group sizes to make a simple unadjusted comparison. The posted values for the primary endpoint are -9.7 for placebo and -17.2 for SEP-363856, with standard errors 1.61 and 1.66. The re-calculated contrast is -7.5, with a 95% confidence interval from -12.03 to -2.968 and unadjusted p = 0.001. This is not a reconstruction of the mixed model. The pooled standard deviation is marked as approximate because it is inferred from the standard errors and sample sizes rather than reported directly.

General explanation: a standard error describes the estimated uncertainty of a sample estimate, whereas a standard deviation describes variation among observations. They are related but not interchangeable. Translating a standard error into an approximate standard deviation requires assumptions about the estimate and sample size. That approximation is useful for an illustrative standardized effect and planning calculation, but it should not be confused with a model-based treatment estimate.

Why do power and sample-size illustrations depend on assumptions?

The sample-size table uses an illustrative calculation based on the observed effect, a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. It considers the stated fractions of the observed effect. These are planning scenarios, not the trial’s reported original design calculations. The registry record does not report the trial’s own sample-size justification or assumptions.

General explanation: power is the probability that a specified statistical procedure would reject a null hypothesis when a particular effect is present, under the assumptions used in the calculation. A planning illustration asks what sample size would be needed for a chosen power under its assumed effect and model. It does not tell us the probability that this completed trial’s finding is true. The listed participant counts are not observed power. Smaller effects are harder to distinguish from random variation and therefore require more participants under otherwise similar assumptions.

What cannot be recovered from the registry description?

The record does not describe the analysis population, missing-data method, prespecified covariates, treatment of protocol deviations, interim analyses, or multiplicity procedure. It also does not give enough information to determine whether the reported model included every randomized participant or how incomplete Week 4 observations were handled. These details matter because analytical choices can affect estimates and uncertainty. The results can be summarized as posted, but the absent details should not be filled in by assuming a conventional approach.

11. Confidence Intervals and Precision

The sponsor-reported primary estimate is a mean difference of -7.5, with a 95% confidence interval from -11.9 to -3.0. This interval conveys uncertainty around the estimated group contrast. Its values are all negative, so the direction of the reported contrast is consistent across the interval with a greater average reduction in PANSS total score in the SEP-363856 group, given the outcome’s scoring direction. It does not imply that every participant had a reduction or that individual responses fall within this interval.

The simple re-calculation gives a 95% confidence interval from -12.03 to -2.968. This is an unadjusted approximation from posted summary statistics, not the sponsor’s mixed models analysis. Its endpoints differ from the sponsor-reported interval, and the sponsor’s own analysis is the primary reported result. The approximate standard deviation used for re-calculation is inferred from standard errors and group sizes, rather than directly reported.

General explanation: precision concerns how tightly an estimate is determined by the data and method. Confidence intervals help readers see both the estimated direction and the range of values reasonably compatible with the analysis. Statistical precision is not the same as clinical importance. The registry does not provide a threshold for a clinically important PANSS difference, so the interval cannot establish whether particular values are meaningful to patients.

The posted group sizes were 125 for placebo and 120 for SEP-363856. Sample size contributes to uncertainty, but interval width also depends on variability, the model, and the analysis. The registry record does not provide individual-level data or sufficient model details to independently verify the sponsor interval. The displayed precision should therefore be read as the precision reported for this analysis, not as a guarantee about estimates in future studies or other populations.

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 table and curve are planning illustrations derived from the observed effect, not the trial’s own original design calculation. The entries are participants per arm, not a total across both arms and not a total number of events. This is a continuous-outcome comparison, so the figures describe participant counts for each arm.

The calculation assumes a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. It considers detecting 100% of the observed effect, 75% of the observed effect, or 50% of the observed effect. The registry record does not report the original sample-size assumptions or the justification used when planning this trial.

The 80% power column gives the illustrative participants per arm for a design with an 80% chance of detecting the assumed effect under the stated calculation assumptions. The 90% power column gives the corresponding number for a 90% chance. These are conditional planning targets, not guarantees; actual operating characteristics depend on the assumptions and the analysis. This is not observed power.

The table shows why the assumed effect matters. Planning for 75% or 50% of the observed effect requires more participants per arm than planning for the full observed effect. Smaller effects are more difficult to distinguish from random variation, so they generally require larger samples when other assumptions are held constant. The illustrative values should not be treated as evidence that the completed trial was adequately or inadequately sized, because its original design assumptions are not reported in the registry record.

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

Worked calculation for this trial

1. Observed difference Δ = -7.500; pooled standard deviation σ = 18.091.
2. n = 2 × (1.96 + 0.842)² × 18.091² / -7.500² = 91.3, rounded up to 92 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 <- -7.5                      # observed difference in means
sigma <- 18.0907085716816                     # 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(2969382)
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 = 92 (80% power), 123 (90% power)
 75% of observed effect: n per arm = 163 (80% power), 218 (90% power)
 50% of observed effect: n per arm = 366 (80% power), 490 (90% power)
power.t.test: 92.3 
Simulated power at n = 92 per arm: 0.798 (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.

14. Analyses Reported by the Sponsor

The sponsor’s reported primary analysis is a mixed models analysis with a mean difference (net) of -7.5, a 95% confidence interval from -11.9 to -3.0, and p = 0.001. This is the trial’s posted analysis and should be the main reference for interpreting the primary endpoint.

The separate re-calculation uses the posted group estimates and standard errors to produce an unadjusted mean difference of -7.5, a 95% confidence interval from -12.03 to -2.968, and an unadjusted p-value of 0.001. It is a simple summary, not a substitute for the sponsor’s model. Its standard deviation is approximate, and it does not reproduce the model’s unreported specifications.

The two estimates have the same reported contrast and similar interval and p-value, while their confidence interval endpoints differ. This page does not identify a discrepancy between the sponsor’s result and the re-calculation. The distinction remains important: the sponsor’s analysis is the official reported analysis, whereas the re-calculation is an explanatory check based on summary statistics. The registry does not provide enough information to reproduce the full mixed models analysis or evaluate all of its assumptions.

15. Multiplicity, Missing Data and Analysis Populations

The registry lists a primary endpoint and multiple secondary outcomes, and it reports analyses for the primary endpoint and several secondary measures. Testing multiple outcomes raises a multiplicity question: if each comparison is assessed separately, the overall chance of at least one apparently positive result may be greater than the chance associated with any single test. General statistical approaches include prespecified testing sequences or adjustments to significance thresholds, but the registry record does not report the strategy used for this trial.

The secondary p-values should therefore be treated as unadjusted individual results. Their pattern may be useful context, but the record does not establish that they were protected from multiplicity or that they were all confirmatory. The primary analysis is distinct because it is designated as the primary outcome in the registry, but details of the original analysis plan are not reported in the registry record.

Missing-data handling is also not described. The record does not say how participants without a Week 4 score were handled, whether any imputation was used, or how the mixed models analysis treated incomplete follow-up. The numbers assigned to the two groups are reported, but the available record does not establish how many participants contributed data to each outcome analysis or whether all randomized participants were included.

The analysis population is not reported in the registry record. It is not possible to determine from the posted details whether the analysis followed an intention-to-treat principle, a per-protocol definition, or another population. The record also does not specify exclusions after randomization, protocol deviations, interim analyses, or sensitivity analyses. These omissions do not change the numerical result as posted, but they limit independent assessment of how robust it is to analytical choices.

General explanation: missing outcomes can affect estimates when the likelihood of missingness is related to outcomes or treatment, and different assumptions can lead to different analyses. A mixed model can use available repeated observations under particular assumptions, but the method label alone does not tell us which assumptions or implementation were used. The registry record does not provide sufficient detail to judge that issue.

16. 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: schizophrenia (2.4%), cardiovascular insufficiency (0.8%), suicide attempt (0.8%).

The registry lists serious adverse events among 3 of 125 participants at risk in the placebo group and 2 of 120 participants at risk in the SEP-363856 group. It lists other adverse events among 47 of 125 participants at risk in the placebo group and 35 of 120 participants at risk in the SEP-363856 group. These are counts of participants with events as posted; they should not be read as proof that assignment caused or prevented an event.

The record lists one death in the SEP-363856 group and none in the placebo group. It does not report the cause of death or whether it was considered related to study treatment. The listed serious-event terms include Schizophrenia at 2.4%, Cardiovascular insufficiency at 0.8%, and Suicide attempt at 0.8%. The record does not provide enough context to attribute those terms to a particular group or infer causality.

Safety counts are descriptive and do not, by themselves, establish comparative safety. The study lasted four weeks, and the record does not provide the timing, severity, relatedness, or detailed clinical course for each event in this summary. Absence of an event in a group’s count is not evidence that the event cannot occur.

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

The primary result describes an average difference in change in PANSS total score between randomized groups at Week 4. The sponsor-reported estimate is -7.5, with a 95% confidence interval from -11.9 to -3.0. Because higher PANSS total scores indicate more severe outcomes, the negative changes and contrast are consistent with a larger average score reduction in the SEP-363856 group than in the placebo group in this study.

This does not mean that every participant experienced the same change. A mean difference summarizes a group-level comparison and does not show the distribution of individual responses. It also does not establish how participants would fare beyond the recorded four-week period. The trial’s listed population is acutely psychotic adults with schizophrenia; results should not automatically be extended to populations or circumstances not described in the registry.

The p-value and confidence interval address statistical uncertainty under the reported analysis. They do not establish clinical importance. The registry record does not provide a threshold for a clinically meaningful change in PANSS total score, and statistical significance alone cannot answer that question. The record also does not provide enough analysis detail to independently evaluate the mixed model, missing-data assumptions, or analysis population.

Secondary measures show a broadly similar direction among the detailed results, but they are not substitutes for the primary result. They are individual secondary comparisons, and multiplicity handling is not reported. Safety counts are likewise descriptive; they do not determine whether an event was caused by study treatment or establish a full balance of benefits and harms.

Overall, the posted data support a description of the observed Week 4 group contrast and its reported precision. They do not, on their own, settle questions about longer-term outcomes, clinical importance, generalizability, or treatment decisions.

18. Why This Trial Matters Statistically

This trial is statistically informative because it reports a randomized, placebo-controlled comparison with a defined primary symptom-scale endpoint and a model-based estimate of the between-group difference. Randomization is intended, in general, to make treatment groups comparable on average, while a concurrent placebo group helps distinguish changes occurring over time from differences between assigned groups. The registry identifies the study as randomized, though details of the randomization process are not reported.

The posted primary result is accompanied by a confidence interval and p-value, allowing readers to assess direction and statistical precision rather than relying only on a point estimate. The simple re-calculation offers a transparent comparison based on posted summary statistics, while the sponsor’s mixed models analysis remains the trial’s reported analysis. Seeing both makes clear that summary-statistic checks and model-based analyses answer related, but not identical, questions.

The study also highlights the importance of separating primary from secondary outcomes. The secondary comparisons provide context about related symptom measures, but their unadjusted p-values should be interpreted cautiously when multiple outcomes are examined. The registry’s safety counts likewise illustrate why descriptive event totals should not be treated as causal comparisons.

Finally, the sample-size illustrations show how planning depends on assumed effect size and power, while the trial’s original sample-size rationale is not reported. Distinguishing those planning scenarios from the completed trial’s observed results helps avoid treating an illustrative calculation as evidence about the design quality or the probability that the result is true.

19. Limitations

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

The registry record does not report several details needed for a full appraisal of the analysis. These include the original sample-size justification, exact mixed-model specification, analysis population, missing-data method, interim analyses, and multiplicity strategy. The record also does not identify which people or roles were masked, despite listing quadruple masking and describing the study as double-blind in its official title.

The primary endpoint is a continuous score at Week 4. The registry describes the scale and its direction, but does not provide a threshold for clinical importance or individual-level response distributions for the primary endpoint. The result is therefore most directly interpreted as a group-average score contrast, not as a count of participants who experienced a particular degree of benefit.

The posted secondary results are not a complete numerical account of every secondary outcome listed in the registry. The broader list includes measures and event outcomes for which detailed comparisons are not reported in the registry record represented. The available secondary p-values are unadjusted, and the record does not explain how multiple testing was addressed.

The re-calculation is limited by reliance on summary statistics and an approximate standard deviation inferred from standard errors and group sizes. It is not a replication of the sponsor’s mixed models analysis. The sponsor-reported result should remain the reference for the primary analysis.

The safety summary gives event counts and selected serious-event terms, but does not give sufficient detail about event timing, treatment relatedness, severity, or causes. The four-week follow-up also limits what can be said about longer-term outcomes. These limitations define the scope of interpretation; they do not alter the values reported in the posted results.

20. Related Tutorials

21. Related Statistical Calculators

22. Sources

• ClinicalTrials.gov: A Study to Evaluate the Efficacy and Safety of SEP-363856 in Acutely Psychotic Adults With Schizophrenia — NCT02969382.
• Registry record last updated 2024-07-05; status last verified 2024-06.
• Results first posted 2021-09-22.
• 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.

23. Record Summary

This completed phase 2 trial randomized 245 acutely psychotic adults with schizophrenia to SEP-363856 or placebo. The SEP-363856 group received a flexible 50 mg or 75 mg capsule once daily; the placebo group received a placebo capsule once daily. The primary outcome was change from baseline in PANSS total score at Week 4.

The sponsor-reported mixed models analysis gave a mean difference of -7.5, with a 95% confidence interval from -11.9 to -3.0 and p = 0.001. The direction corresponds to a larger average reduction in PANSS total score in the SEP-363856 group. The registry also posts secondary symptom outcomes and descriptive adverse-event counts, but details of the analysis plan, missing-data handling, and multiplicity strategy are not reported in the registry record.