Brexpiprazole for the Treatment of Patients With Agitation Associated With Dementia of the Alzheimer's Type: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Brexpiprazole for the Treatment of Patients With Agitation Associated With Dementia of the Alzheimer's Type (NCT03620981), a randomized phase 2, phase 3 study. The primary outcome is Change From Baseline in Cohen-Manfield Agitation Inventory(CMAI) Score at 10 Weeks After Dosing..

2. Clinical Question

The trial asked whether brexpiprazole, given once daily for 10 weeks, was superior to placebo for agitation associated with dementia of the Alzheimer's type. It evaluated two brexpiprazole doses, 1 mg and 2 mg, and also investigated safety and dose selection. The primary outcome was change from baseline in the Cohen-Mansfield Agitation Inventory (CMAI) score at 10 weeks after dosing.

The registry lists secondary assessments of Clinical Global Impression of Severity (CGI-S) change and Clinical Global Impression of Improvement (CGI-I) score at 10 weeks. These provide additional clinical assessments, but the primary question rests on the CMAI outcome. The registry record does not report a separate primary comparison between the two brexpiprazole doses; the reported comparisons are each dose against placebo.

3. Trial Design

The registry describes a Phase 2/3, multicenter, placebo-controlled, randomized, double-blind, parallel-group trial. It is listed as completed. The registry reports enrollment of 410 and gives a start date of 2018-08-20 and a primary completion date of 2023-04-21. The trial evaluated brexpiprazole 1 mg/day, brexpiprazole 2 mg/day, and placebo, each administered once daily for 10 weeks.

Random allocation is intended to make treatment groups comparable, on average, with respect to measured and unmeasured characteristics at the time of allocation. A placebo comparison can help distinguish changes occurring during the study from changes associated with receiving the active study drug, though randomization alone does not resolve questions about missing outcomes, adherence, or the analysis methods used. Parallel-group assignment means participants were assigned to a study group rather than being described as crossing between treatments.

The registry lists masking as quadruple. It does not identify in the information reported which specific roles were masked, so the exact masking arrangements are not reported in the registry record. The registry also does not report the original sample-size justification, intended allocation proportions, randomization procedures, or protocol-defined analysis plan. The listed enrollment should not be treated as a substitute for those design details.

The primary endpoint population was the full analysis set. That set included participants who received at least one dose and had a baseline CMAI total score and at least one score after treatment began. The outcome summaries list 108.0 participants for brexpiprazole 1 mg, 148.0 for brexpiprazole 2 mg, and 147.0 for placebo. The registry's adverse-event counts use their own at-risk group counts, which differ from the primary-outcome analysis counts. The registry record does not report why these analysis populations differ.

4. Participant Flow and Accounting

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

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: The full analysis set (FAS) comprised subjects who, after randomization, have received at least 1 dose of the IMP, and from whom CMAI total scores have been obtained at baseline and at least 1 time point after initiation of the treatment.

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 Cohen-Manfield Agitation Inventory(CMAI) Score at 10 Weeks After Dosing., measured over Baseline and 10 weeks first dose, on average 10 weeks.

The primary endpoint was change from baseline in the Cohen-Mansfield Agitation Inventory (CMAI) score at 10 weeks after dosing. The timeframe is described as baseline and 10 weeks after the first dose, on average 10 weeks. The CMAI assesses the frequency of agitated behaviors in elderly persons, including hitting, cursing, and restlessness. It consists of 29 items, each rated on a 1 to 7 scale, where 1 is the best rating and 7 is the worst. The possible total score ranges from 29 to 203. A decrease in score indicates improvement in symptoms.

The endpoint is reported as a least-squares mean change with a standard error for each group. The least-squares mean is a model-based estimate rather than simply a raw arithmetic average; the specific model details beyond the sponsor's named analysis are not reported in the registry record. The mean change describes the average direction and magnitude of score movement in the analysis group. It does not show how every participant's score changed or how outcomes varied across individuals.

A measure focused on change from baseline allows the comparison to reflect movement during the treatment period rather than only the final score. The primary analysis compares each dose group with placebo. Because a lower CMAI score is better on this scale, the negative group changes and negative between-group differences correspond to decreases in the measure. This interpretation is specific to the scale's stated direction and is not, by itself, a judgment about the importance of the size of a difference to patients.

7. Results Posted in the Registry

• Brexpiprazole 1 mg: mean -11.70 versus -8.00 on control; difference -3.70 (95% CI -6.80 to -0.60), standardized effect d = -0.30; z-test p = 0.019.
• Brexpiprazole 2 mg: mean -15.20 versus -8.00 on control; difference -7.20 (95% CI -10.08 to -4.32), standardized effect d = -0.57; 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 of Severity (CGI-S) Score at 10 Weeks After Dosing.

• 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:
  • Brexpiprazole 1 mg: mean -1.00 versus -0.60 on control; difference -0.40 (95% CI -0.66 to -0.14), standardized effect d = -0.37; z-test p = 0.003.
  • Brexpiprazole 2 mg: mean -1.40 versus -0.60 on control; difference -0.80 (95% CI -1.05 to -0.55), standardized effect d = -0.73; z-test p = <0.001.
• Interpretation:
  • The 95% confidence interval excludes no difference, so the mean is statistically lower for Brexpiprazole 1 mg than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.
  • The 95% confidence interval excludes no difference, so the mean is statistically lower for Brexpiprazole 2 mg than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.

Clinical Global Impression of Improvement (CGI-I) Score at 10 Weeks After Dosing.

• 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 deviations.
• Result:
  • Brexpiprazole 1 mg: mean 3.10 versus 3.50 on control; difference -0.40 (95% CI -0.73 to -0.07), standardized effect d = -0.29; z-test p = 0.019.
  • Brexpiprazole 2 mg: mean 2.60 versus 3.50 on control; difference -0.90 (95% CI -1.20 to -0.60), standardized effect d = -0.69; z-test p = <0.001.
• Interpretation:
  • The 95% confidence interval excludes no difference, so the mean is statistically lower for Brexpiprazole 1 mg than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.
  • The 95% confidence interval excludes no difference, so the mean is statistically lower for Brexpiprazole 2 mg than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.

The registry lists two secondary outcomes: change from baseline in CGI-S score and CGI-I score at 10 weeks after dosing. Across these additional measures, the overall direction of the posted comparisons is consistent with the primary CMAI finding: both brexpiprazole groups are reported as having differences from placebo in the direction labeled superior by the sponsor. The secondary outcomes therefore provide a broadly concordant pattern across different clinical assessments.

That consistency is relevant, but it does not make the secondary analyses equivalent to the primary endpoint. The CMAI was designated as primary in the outcome record, while the CGI-S and CGI-I outcomes are secondary. The measures assess different aspects of clinical status, use different score summaries, and are analyzed with methods that are not identical in every case. Their estimates should be interpreted according to the scale and method specified for each outcome, rather than combined as if they were repeated measurements of a single endpoint.

The registry reports sponsor analyses for these outcomes, but does not report whether multiplicity was controlled across secondary outcomes and their treatment comparisons. When many outcomes or comparisons are examined, the chance of at least one apparently notable result can increase. Without a stated adjustment strategy or a prespecified hierarchy, unadjusted secondary results carry less confirmatory weight than the primary analysis. They are supportive context and can help characterize the pattern of findings, but should not independently establish a treatment effect.

The registry record also does not report a clinically meaningful threshold for either CGI measure or the procedures used to handle missing secondary assessments. Thus the secondary findings can be described as directionally consistent with the primary result, while their clinical interpretation and statistical independence from the other reported tests remain uncertain.

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

• Change From Baseline in Clinical Global Impression of Severity (CGI-S) Score at 10 Weeks After Dosing.
• Clinical Global Impression of Improvement (CGI-I) Score at 10 Weeks After Dosing.

9. Statistical Interpretation

10. Statistical Methods Explained

The sponsor reports Mixed Models Analysis for the two primary comparisons, with the parameter described as LS Mean Difference (Final Values). The registry's group summaries, by contrast, present least-squares mean change from baseline and standard error. These are related kinds of model-based summaries but are not interchangeable descriptions: one is a group-level change summary, while the other is the reported between-group contrast. The registry does not provide the model specification, covariance structure, included covariates, or handling of incomplete repeated measurements, so those details are not reported in the registry record.

What a mean difference communicates

A mean difference is a contrast between group averages expressed in the outcome's units. For the primary measure, a negative difference favors a larger decrease in CMAI score in the brexpiprazole group, because the scale description says that a decrease indicates improvement. A difference is not a measure of how many individual participants benefited, and it does not show whether responses were consistent across participants. The registry reports the least-squares mean difference for each dose compared with placebo rather than a direct dose-versus-dose estimate.

What a standard error communicates

The primary group summaries include standard errors. In general, a standard error describes the uncertainty in an estimated mean, rather than the spread of individual participant scores. It tends to be smaller when an estimate is more precise, all else equal. It should not be read as if it were the range containing most individual outcomes. The posted primary summaries do not include standard deviations, and the registry does not report raw participant-level distributions.

What confidence intervals and p-values address

In general, a confidence interval gives a range of values compatible with the data and the analysis assumptions at a stated confidence level. A narrower interval signals greater precision than a wider interval, while an interval that crosses the null value can be compatible with no difference as well as differences in either direction. For a difference, the null value is zero. A p-value describes how unusual data at least as incompatible with the null as the observed data would be under a specified statistical model; it is not the probability that the null hypothesis is true. Neither measure establishes clinical importance.

The sponsor reports 95% confidence intervals and p-values for the primary comparisons. The sponsor's p-values use the model-based analysis. The separate unadjusted calculations use posted summary statistics, and the registry notes that the pooled standard deviations used in those calculations are approximations based on the reported standard errors. General statistical caution is therefore appropriate when treating a simple calculation as a reproduction of a model-based analysis.

Why over-dispersion can matter

General explanation: over-dispersion occurs when observed variability is greater than a statistical model assumes. If a model understates variability, its standard errors can be too small, which can make confidence intervals narrower and p-values smaller than warranted. If variability is modeled appropriately, uncertainty estimates better reflect the data. The primary CMAI outcome is continuous, and the registry does not report a specific over-dispersion assessment or a model feature addressing it. This general issue should not be taken as evidence that over-dispersion occurred in this trial.

How the full analysis set defines inclusion

The primary full analysis set included participants who received at least one dose and had a baseline CMAI total score and at least one post-treatment score. This definition is important because it describes who contributes to the stated primary outcome population. It is not the same as saying that every randomized participant contributed a 10-week score. The registry record does not report how many eligible participants lacked later assessments, how their missing outcomes were addressed, or whether any imputation method was used. The exact statistical handling of repeated observations is not reported beyond the named Mixed Models Analysis.

For the secondary CGI-I outcome, the sponsor reports a Cochran-Mantel-Haenszel analysis and a parameter described as Mean Difference (Final Values). The registry does not explain how that method was adapted to the reported score or provide its model details. The secondary CGI-S analyses are reported as Mixed Models Analysis. These names identify the methods reported but are not enough to reconstruct the full analysis. General principles about model assumptions apply, but trial-specific details not stated in the registry should not be inferred.

Finally, the posted recalculations are not adjusted for any unreported baseline factors, repeated-measure structure, or multiplicity. They are useful as checks against the posted group summaries, but they do not replace the sponsor's analysis. The registry identifies no discrepancy between the calculations and sponsor results. That supports reading them as broadly consistent, while retaining the distinction between an unadjusted reconstruction and the sponsor's analysis.

11. Confidence Intervals and Precision

Precision concerns how narrowly the data constrain an estimate. The sponsor reported 95% confidence intervals for the primary comparisons. For brexpiprazole 1 mg versus placebo, the estimate was -3.7, with a 95% confidence interval from -6.8 to -0.7. For brexpiprazole 2 mg versus placebo, the estimate was -7.2, with a 95% confidence interval from -10.0 to -4.3. Both intervals are below zero, consistent with the negative direction of the estimated differences.

The intervals also show that the point estimates are not exact quantities. For each comparison, a range of effect sizes remains compatible with the fitted analysis and its assumptions. The 1 mg interval spans a range that includes estimates closer to zero than the 2 mg interval, while the 2 mg interval also remains uncertain about the exact size of the difference. Confidence intervals do not state the probability that a particular treatment effect is true, and they do not define whether a difference matters clinically.

The posted group means are accompanied by standard errors, but the registry's simple re-calculation uses approximate standard deviations derived from those errors. That distinction matters because a calculation based on summary values cannot capture all details of the sponsor's model. The sponsor's reported estimates and intervals are therefore the relevant model-based precision results. Since the discrepancy indicator is false, the recalculations do not signal a material conflict with the sponsor analysis; they remain simplified estimates rather than a replacement for it.

Precision should also be considered alongside the analysis population and the registry's missing-data reporting. The full analysis set required at least one post-treatment CMAI measurement, but information about missingness and its handling is not reported in the registry record. Confidence intervals quantify uncertainty under the analysis used; they do not account for every possible source of bias or uncertainty omitted from the record.

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 figures are planning illustrations based on the observed effect, not a report of the original trial's design assumptions. For these continuous outcomes, the figures are participants per arm, not total events. The stated planning assumptions are a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. Each scenario uses an effect equal to the stated fraction of the observed effect. The original trial's own sample-size assumptions and justification are not reported in the registry record.

For each effect scenario, the 80% power column gives the approximate number of participants per arm that would be needed for an 80% chance of detecting an effect of that assumed size under the stated planning conditions, if the effect and model assumptions were correct. The 90% power column gives the corresponding per-arm planning figure for a 90% chance. These are conditional planning probabilities, not guarantees about a future study's result. The values are not observed power.

The scenarios use 100%, 75%, and 50% of the observed effect. As the assumed effect becomes smaller, more participants are required to distinguish it from random variation under the same planning assumptions. This is why the planning figures increase as the assumed fraction declines. The estimates are based on observed effects from this trial, so they may be optimistic or otherwise unrepresentative if the observed effects differ from the effects a future study should expect. They also do not incorporate design features or analysis choices that are not listed in the planning assumptions.

The display should therefore be read as a sensitivity exercise: it illustrates how changing the assumed effect size changes the planning requirement, rather than identifying a uniquely correct sample size. It does not show that the completed study had a particular probability of success, and it does not establish the adequacy of the original design.

13. How the Sample Size Is Derived

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

Formula

n is the number of participants in each arm, Δ is the difference in means to detect and σ is the common standard deviation.

Worked calculation for this trial

1. Observed difference Δ = -3.700; pooled standard deviation σ = 12.481.
2. n = 2 × (1.96 + 0.842)² × 12.481² / -3.700² = 178.6, rounded up to 179 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 <- -3.7                      # observed difference in means
sigma <- 12.4807639816326                     # 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(3620981)
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 = 179 (80% power), 240 (90% power)
 75% of observed effect: n per arm = 318 (80% power), 426 (90% power)
 50% of observed effect: n per arm = 715 (80% power), 957 (90% power)
power.t.test: 179.6 
Simulated power at n = 179 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.

14. Analyses Reported by the Sponsor

The sponsor's primary analyses are reported as Mixed Models Analysis, with least-squares mean differences for each brexpiprazole dose versus placebo. The reported estimate is -3.7 for brexpiprazole 1 mg and -7.2 for brexpiprazole 2 mg. Their reported 95% confidence intervals are -6.8 to -0.7 and -10.0 to -4.3, respectively. The registry labels both comparisons as superiority analyses.

Simple unadjusted recalculations from the posted group summaries yield mean differences of -3.7 and -7.2, with p-values of 0.019 and p<0.001. These align closely with the sponsor's estimates and support the same direction of comparison. The sponsor reports p=0.018 for 1 mg and p< 0.001 for 2 mg. Small differences between the recalculated and sponsor p-values are expected when summary-statistic calculations are compared with a model-based analysis; the recalculation does not reproduce every aspect of that model.

The registry does not mark a discrepancy. Accordingly, the sponsor's own analysis is the primary inferential result, and the recalculation is best understood as a consistency check on the posted summaries. It should not be used to override the sponsor's model-based estimates or to imply that the analysis methods were identical.

15. Multiplicity, Missing Data and Analysis Populations

The registry identifies a primary CMAI outcome and two secondary outcomes, with each outcome compared between two brexpiprazole doses and placebo. It does not report a multiplicity-control plan, such as a prespecified hierarchy or an adjustment across outcomes and comparisons. This is important because repeated statistical testing can increase the chance of findings that appear noteworthy by chance. Without a reported strategy, the secondary p-values should be interpreted as unadjusted evidence, and the record does not establish how the sponsor intended to control the overall error rate across tests.

The record also does not report whether interim analyses were planned or performed, whether early stopping rules were used, or how final analysis decisions were made. These items are not required to assume any particular procedure, but their absence limits the ability to assess the complete statistical plan. Details of the original sample-size justification and the trial's prespecified analysis plan are not reported in the registry record.

The primary analysis population is defined as the full analysis set: participants who received at least one dose and had a baseline CMAI total score plus at least one score after treatment began. This gives a clear minimum criterion for inclusion in the posted primary outcome summaries. However, the record does not report the number or pattern of missing observations at 10 weeks, reasons for missingness, or the method used to handle missing outcome data. It also does not report whether sensitivity analyses tested alternative assumptions about missing outcomes.

These omissions matter because the effect estimate and its uncertainty depend on who contributes data and how unavailable outcomes are treated. A model that uses repeated measurements may make assumptions about missingness, but the registry does not provide enough detail to determine those assumptions or assess their plausibility. No specific missing-data method should be inferred from the name Mixed Models Analysis alone.

Likewise, the registry's outcome population and adverse-event at-risk groups are not identical as reported. The reasons for that difference are not stated. The record permits description of the reported analyses and counts, but not a complete reconstruction of participant flow, all analysis populations, or the protocol's decisions about multiplicity and missing data.

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 2% reporting threshold.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: pneumonia aspiration (1.8%), agitation (1.3%), cardiac death (0.9%), hepatic mass (0.9%), femoral neck fracture (0.9%).

The adverse-event record lists serious adverse events for groups with 112 participants at risk in the brexpiprazole 1 mg group, 149 in the brexpiprazole 2 mg group, and 149 in the placebo group. The reported serious adverse event counts were 7, 9, and 7, respectively. These are counts, not rates calculated from the group sizes. They should be read using the stated at-risk counts rather than the primary-outcome analysis group sizes.

The registry lists pneumonia aspiration at 1.8% among the most frequent serious-event terms, agitation at 1.3%, and cardiac death, hepatic mass, femoral neck fracture, and breast cancer each at 0.9%. It reports deaths of 2 in the 1 mg group and 0 in each of the 2 mg and placebo groups. These are descriptive registry entries; the record does not establish that any listed event was caused by the assigned treatment.

The overall counts of other adverse events were 77, 113, and 93 for the 1 mg, 2 mg, and placebo groups, respectively, using the same group-specific at-risk counts. These totals do not describe severity, duration, or relatedness for each event. The registry record does not provide enough detail to compare event exposure time or to draw causal conclusions. Safety counts should therefore be interpreted as reported observations, not as proof of comparative safety.

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

The registry results show that average CMAI scores decreased from baseline in all three groups over the reported treatment period. The sponsor's primary comparisons estimate larger average decreases for each brexpiprazole dose than for placebo, with confidence intervals lying below zero. This is the statistical pattern reported for the primary outcome in the stated full analysis set.

That pattern does not mean every participant's symptoms improved. Group averages can coexist with a wide range of individual responses, and the registry does not provide participant-level outcome distributions in the record described. The results also do not by themselves establish the practical importance of the estimated differences. A clinically meaningful threshold for the CMAI comparison is not reported in the registry record.

The findings are specific to the recorded population, outcome definition, comparisons, and 10-week dosing period. They do not establish what would occur with longer treatment, in populations not represented by the study, or under different clinical circumstances. The registry record does not provide those results. Nor does a statistically supported comparison answer questions about individual care decisions; the results are a description of group-level trial data, not medical advice.

Statistical significance should not be equated with certainty, clinical importance, or causal proof beyond the trial's design and analysis assumptions. Randomization and placebo control support a comparative interpretation, but missing-data handling, multiplicity procedures, analysis details, and the original design rationale are not fully reported. The sponsor's model-based results are the primary posted inference, while the summary-based recalculations are a simplified consistency check.

The most defensible interpretation is therefore bounded: the posted primary analysis reports a negative between-group difference for each brexpiprazole dose compared with placebo, with the stated 95% confidence intervals and p-values. The magnitude and relevance of those differences, the robustness to unreported analysis choices, and the meaning for any individual remain separate questions that the registry results do not resolve.

18. Why This Trial Matters Statistically

This trial is statistically informative because it records a randomized, placebo-controlled comparison with a prespecified primary continuous outcome and reports both group summaries and sponsor analyses. The outcome is interpretable in a stated direction: a lower CMAI score represents improvement. That explicit scale direction makes it possible to understand why negative changes and negative comparisons correspond to larger average decreases in score.

The study also illustrates the distinction between an estimate and its precision. The sponsor provides effect estimates, 95% confidence intervals, and p-values for each primary comparison. Reading these together is more informative than relying on a p-value alone: the estimate gives the direction and scale of the contrast, while the interval conveys uncertainty about its size under the model.

There is also value in comparing model-based results with simple calculations from posted summaries. The recalculated primary estimates are close to the sponsor's estimates, offering a transparent consistency check. At the same time, the difference in analytical detail illustrates why re-calculation is not a substitute for the sponsor's method. Published summary statistics cannot reconstruct every feature of the model or the treatment of incomplete data.

The record highlights the limits of registry interpretation as well. It defines the primary analysis population, but does not report the full missing-data strategy, multiplicity plan, or original sample-size rationale. Those absences affect how completely a reader can evaluate robustness and confirmatory status. The sample-size illustrations show the sensitivity of planning figures to the assumed effect, but they do not recover the trial's original design process.

Finally, the secondary outcomes and adverse-event counts broaden the description without replacing the primary endpoint. Together, these elements make the record useful for statistical reading: estimates, uncertainty, analysis populations, and reporting gaps all matter to understanding what the trial's results can support.

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.

Several trial-specific limitations arise from details not reported in the registry record. The full analysis set requires at least one post-treatment CMAI score, but the record does not describe the extent or reasons for missing 10-week outcomes, the assumptions used to address them, or sensitivity analyses. Without those details, the robustness of the primary estimates to missing data cannot be assessed from the registry information.

The sponsor names Mixed Models Analysis for the primary comparisons, but the registry does not provide the model specification, covariates, covariance structure, or the exact use of repeated measurements. Consequently, the reported method cannot be independently reconstructed from the posted group summaries. The recalculations rely on approximate pooled variability based on reported standard errors and are not a substitute for a full reanalysis.

The registry does not report the multiplicity plan across the primary and secondary outcomes or across the two dose comparisons. It also does not report whether interim analyses were conducted. These gaps limit assessment of the planned inferential hierarchy and the extent to which secondary findings should be considered confirmatory.

The results concern a 10-week treatment regimen, and longer-term outcomes are not reported in the registry record. The record also provides no stated clinically meaningful threshold for the primary CMAI difference, making it difficult to judge the practical importance of the estimated magnitude using registry information alone. The outcome summaries are group-level statistics and do not describe the distribution of individual responses.

Safety entries are descriptive counts and selected serious-event terms. The registry information does not provide a causal attribution for these events or the additional context needed to make a comprehensive comparative safety assessment. Finally, the original design assumptions and sample-size justification are not reported, so the planning illustrations based on observed effects should not be mistaken for an account of why the trial enrolled its stated number of participants.

20. Related Tutorials

21. Related Statistical Calculators

22. Sources

• ClinicalTrials.gov: Brexpiprazole for the Treatment of Patients With Agitation Associated With Dementia of the Alzheimer's Type — NCT03620981.
• Registry record last updated 2024-11-14; status last verified 2024-04.
• Results first posted 2024-11-14.
• 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

The registry describes a completed Phase 2/3, multicenter, randomized, placebo-controlled, double-blind, parallel-group trial of brexpiprazole for agitation associated with dementia of the Alzheimer's type. Participants received brexpiprazole 1 mg/day, brexpiprazole 2 mg/day, or placebo once daily for 10 weeks.

The primary outcome was change from baseline in CMAI score at 10 weeks. The sponsor reported negative least-squares mean differences versus placebo for both brexpiprazole doses, with 95% confidence intervals below zero. Simple unadjusted recalculations from posted summaries were close to the sponsor's estimates. The registry also lists CGI-S and CGI-I secondary outcomes and descriptive adverse-event counts.

The full analysis set required at least one dose and baseline plus at least one post-treatment CMAI assessment. Details of missing-data handling, multiplicity control, the original sample-size rationale, and the full statistical model are not reported in the registry record. The results describe group-level differences over the recorded treatment period and do not establish clinical importance or individual outcomes.