A Study to Investigate the Efficacy, Safety and Tolerability of Four Different Doses of BI 409306 Compared to Placebo Given for 12 Weeks in Patients With Schizophrenia on Stable Antipsychotic Treatment.: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for A Study to Investigate the Efficacy, Safety and Tolerability of Four Different Doses of BI 409306 Compared to Placebo Given for 12 Weeks in Patients With Schizophrenia on Stable Antipsychotic Treatment. (NCT02281773), a randomized phase 2 study. The primary outcome is Change From Baseline in the Composite Score of Measurement and Treatment Research to Improve Cognition in Schizophrenia (MATRICS) Consensus Cognitive Battery (MCCB) After 12 Weeks of Treatment.

2. Clinical Question

This is a question about the measured difference between randomized groups, not whether any one participant improved. The trial also assessed safety and tolerability, as well as secondary measures of everyday functioning, clinical severity, patient-rated improvement, and symptoms. Those outcomes provide additional context, but the primary endpoint is the prespecified focus of this interpretation.

The posted estimates are group-level results. They do not by themselves establish why scores changed, or whether a measured difference would be important to an individual participant. The estimates and their uncertainty should be considered together, particularly because the confidence intervals for the sponsor-reported primary comparisons include no difference.

3. Trial Design

This completed phase II study used randomized, double-blind, parallel groups and compared four once-daily BI 409306 dose levels with placebo. The planned treatment period was 12 weeks, and the recorded enrollment was 518. Participants had schizophrenia and were on stable antipsychotic treatment. The record lists Boehringer Ingelheim as the sponsor.

The study was described as a “learn and confirm” design with two stages. Stage 1 was intended to identify meaningful cognition endpoint or endpoints among CANTAB domains. The record says that none of the CANTAB outcome measures met the prespecified selection criteria at the planned time. As predefined for Stage 2, the MCCB composite score was therefore used as the primary endpoint.

For the primary endpoint, the reported arm summaries include 79 participants in the 10 milligram group, 76 in the 25 milligram group, 69 in the 50 milligram group, 76 in the 100 milligram group, and 150 in the placebo group. The primary outcome population was narrower than everyone randomized: the full analysis set required treatment exposure and the specified baseline and post-baseline primary endpoint data. Arm counts in the posted outcome summaries should therefore be read as analysis counts, not as a complete description of enrollment or allocation.

The registry record does not report the original sample-size justification, allocation ratio, or full details of the randomization procedure. It also does not report the trial’s complete statistical analysis plan. The record gives start and primary completion dates, but those dates do not add information about the design assumptions or analysis methods.

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: Actually, 518 subjects were entered/randomised however 2 subjects were not treated and hence the number of subjects that started equals 516. One belongs to BI 409306 - 25 milligram and another to Placebo group.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: The full analysis set (FAS) was to consist of all randomisation patients who were treated with at least one dose of study drug and had a baseline and at least one post baseline on treatment primary endpoint MCCB composite score.

5. Baseline Characteristics

The table reproduces the baseline characteristics posted in the registry, which describe the participants analysed at baseline and help show whether the randomized groups were comparable.

Population: The treated set (TS) was to consist of all patients who were randomised and treated with at least one dose of study drug.

6. Primary Endpoint

The registry lists the primary outcome as Change From Baseline in the Composite Score of Measurement and Treatment Research to Improve Cognition in Schizophrenia (MATRICS) Consensus Cognitive Battery (MCCB) After 12 Weeks of Treatment, measured over Baseline and Week 12.

The primary endpoint was change from baseline in the composite score of the MATRICS Consensus Cognitive Battery (MCCB) after 12 weeks of treatment. The MCCB comprises 10 tests assessing 7 cognitive domains: speed of processing, attention vigilance, working memory, verbal learning, visual learning, reasoning problem solving, and social cognition. The composite score was calculated by summing standardized domain scores for analysis.

The score ranges from -20 to 99, with higher scores indicating a better outcome. Measuring change from baseline focuses the analysis on how the score shifted during the treatment period, rather than on the week 12 score alone. Comparing those changes across randomized groups is intended to assess whether the groups differed on this cognitive measure over the study period.

The reported primary measure is a least-squares mean, with standard error as the measure of dispersion. The registry describes the analysis population as the full analysis set: randomized participants treated with at least one dose who had a baseline and at least one post-baseline on-treatment MCCB composite score. The endpoint was selected for Stage 2 under the study’s described two-stage approach after CANTAB outcomes were not selected at the planned time.

In general, a cognitive composite can summarize multiple domains, but it also means the result is about the composite as defined, not a separate conclusion for every test or domain. The registry record does not report a threshold for a clinically meaningful change.

7. Results Posted in the Registry

• BI 409306 - 10 Milligram: mean 1.20 versus 2.50 on control; difference -1.30 (95% CI -3.08 to 0.48), standardized effect d = -0.19; z-test p = 0.153.
• BI 409306 - 25 Milligram: mean 2.70 versus 2.50 on control; difference 0.20 (95% CI -1.63 to 2.03), standardized effect d = 0.03; z-test p = 0.830.
• BI 409306 - 50 Milligram: mean 2.80 versus 2.50 on control; difference 0.30 (95% CI -1.55 to 2.15), standardized effect d = 0.04; z-test p = 0.750.
• BI 409306 - 100 Milligram: mean 1.80 versus 2.50 on control; difference -0.70 (95% CI -2.52 to 1.12), standardized effect d = -0.10; z-test p = 0.450.

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 Everyday Functional Capacity as Measured by Schizophrenia Cognition Rating Scale (SCoRS) Global Ratings After 12 Weeks of Treatment

• 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:
  • BI 409306 - 10 Milligram: mean -2.20 versus -2.50 on control; difference 0.30 (95% CI -0.99 to 1.59), standardized effect d = 0.06; z-test p = 0.648.
  • BI 409306 - 25 Milligram: mean -3.10 versus -2.50 on control; difference -0.60 (95% CI -1.94 to 0.74), standardized effect d = -0.12; z-test p = 0.379.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between BI 409306 - 10 Milligram and the control. The interval gives the range of effects that remain compatible with the data.
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between BI 409306 - 25 Milligram and the control. The interval gives the range of effects that remain compatible with the data.

Change From Baseline in Clinical Global Impressions-Severity (CGI-S) Scale Score After 12 Weeks of Treatment

• 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:
  • BI 409306 - 10 Milligram: mean -0.10 versus -0.10 on control; difference 0.00 (95% CI -0.11 to 0.11), standardized effect d = 0.00; z-test p = 1.000.
  • BI 409306 - 25 Milligram: mean -0.10 versus -0.10 on control; difference 0.00 (95% CI -0.11 to 0.11), standardized effect d = 0.00; z-test p = 1.000.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between BI 409306 - 10 Milligram and the control. The interval gives the range of effects that remain compatible with the data.
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between BI 409306 - 25 Milligram and the control. The interval gives the range of effects that remain compatible with the data.

Patient Global Impressions-Improvement (PGI-I) Scale Score Measured After 12 Weeks of Treatment

• 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:
  • BI 409306 - 10 Milligram: mean 2.99 versus 3.04 on control; difference -0.05 (95% CI -0.35 to 0.25), standardized effect d = -0.04; z-test p = 0.742.
  • BI 409306 - 25 Milligram: mean 2.88 versus 3.04 on control; difference -0.15 (95% CI -0.46 to 0.15), standardized effect d = -0.14; z-test p = 0.319.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between BI 409306 - 10 Milligram and the control. The interval gives the range of effects that remain compatible with the data.
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between BI 409306 - 25 Milligram and the control. The interval gives the range of effects that remain compatible with the data.

The posted secondary outcomes cover everyday functional capacity, clinical severity, patient-rated improvement, and symptom measures. Across the summaries, the group estimates do not form a consistent pattern that clearly separates the active dose groups from placebo. This is broadly in keeping with the primary comparisons, which also do not show a clear separation in the sponsor-reported estimates.

Secondary measures can help describe whether a primary outcome pattern is accompanied by changes in related areas, but they do not replace the primary endpoint. They may measure different constructs, use different scales, and have different analysis populations. The registry lists 10 outcome measures and 12 analyses, which means the posted set involves multiple statistical comparisons.

In general, unadjusted secondary results carry less weight when many outcomes or comparisons are examined. Without a reported multiplicity strategy, a nominal finding can be more vulnerable to chance selection among several analyses. The registry record does not report a complete adjustment plan for the secondary outcomes. It is therefore more appropriate to treat them as supporting context than as independent confirmation of a treatment effect.

The secondary results should also be interpreted with their own uncertainty and analysis methods, rather than inferred from the primary result. The tables provide the available endpoint-level summaries; they should be read as distinct outcomes and not pooled into a single overall conclusion. The broader pattern is that the posted secondary evidence does not materially resolve the uncertainty in the primary comparison.

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

• Change From Baseline in Everyday Functional Capacity as Measured by Schizophrenia Cognition Rating Scale (SCoRS) Global Ratings After 12 Weeks of Treatment
• Change From Baseline in Clinical Global Impressions-Severity (CGI-S) Scale Score After 12 Weeks of Treatment
• Patient Global Impressions-Improvement (PGI-I) Scale Score Measured After 12 Weeks of Treatment
• Change From Baseline in PANSS Negative Symptom Factor Score After 12 Weeks of Treatment (for Subset of Patients Diagnosed With Negative Symptom)
• Change in Psychopathology Symptoms as Assessed by Positive and Negative Syndrome Scale (PANSS)

9. Statistical Interpretation

10. Statistical Methods Explained

The primary results are reported as least-squares means and sponsor comparisons from mixed models analysis. The separate recalculations shown alongside the results are unadjusted estimates based on posted arm summaries. These are different ways of presenting the data and should not be treated as interchangeable.

What does a mean difference represent?

A mean difference describes the contrast between two group summaries. Here, each dose group is compared with placebo for the primary cognitive outcome. A value below zero indicates that the dose-group estimate is lower than the placebo estimate under the comparison’s stated direction; a value above zero indicates a higher estimate. The interpretation of direction should be considered alongside the scale description, where higher MCCB scores indicate a better outcome.

The recalculated contrasts use the posted arm means, participant counts, and standard errors. The reported pooled standard deviations for these recalculations are marked as approximate. In general, a pooled standard deviation describes variability across groups on a common scale, but deriving it from summary standard errors is an approximation rather than a reanalysis of participant-level observations.

What is the role of the mixed models analysis?

The registry identifies mixed models analysis for the sponsor-reported primary comparisons and labels the parameter as mean difference in final values. A model-based analysis can account for the specified structure of the observed data and may use information differently from a simple contrast of arm summaries. The registry record does not report the full model specification, covariates, covariance structure, or missing-data assumptions, so those details cannot be assessed.

The primary measure is a least-squares mean with standard error. In general, a least-squares mean is a model-based group estimate, while a standard error describes uncertainty in an estimated mean. It is not the same as the standard deviation, which describes spread among individual observations. The registry gives both the outcome summaries and the sponsor’s group comparisons; the comparison is the more direct quantity for the randomized treatment question.

How should confidence intervals be read?

A confidence interval gives a range of values compatible with the observed data and the analysis method, under the assumptions used to construct it. In general, narrower intervals indicate greater precision, while wider intervals allow a broader range of effects. For the sponsor-reported primary comparisons, every 95% interval includes zero. This means the estimates do not rule out no difference; it does not establish that the groups are equivalent or that every possible effect is absent.

A p-value is another summary of how the data relate to a null comparison under a statistical model. It is not the probability that the null hypothesis is true, nor a measure of clinical importance. The posted primary p-values are paired with the sponsor’s model-based comparisons. The recalculated p-values are unadjusted and come from a different calculation, so they should not be used to replace the reported primary analysis.

Why do dose comparisons raise multiplicity questions?

The primary endpoint is compared across four dose groups against placebo, and the record also lists several secondary endpoints. In general, testing multiple comparisons increases the chance that at least one result may appear noteworthy by chance if no adjustment is made. The record describes a staged endpoint-selection process, but the multiplicity adjustment for the four primary dose comparisons and the secondary outcomes is not reported in the registry record.

The “learn and confirm” description is important context: the registry says that CANTAB outcome measures were not selected at the planned time under the stated criteria, and the MCCB composite score was chosen for Stage 2 as predefined. This does not, by itself, establish how all analyses were controlled for multiplicity or how exploratory findings were treated.

What is known about the analysis population?

The full analysis set required randomization, receipt of at least one dose, and baseline plus at least one post-baseline on-treatment primary endpoint measurement. This definition explains why primary endpoint analysis counts differ across groups and from overall enrollment. The registry record does not report the number excluded for each reason or a detailed account of missing observations.

In general, incomplete outcome data can affect an estimate if the availability of measurements is related to participant characteristics or outcomes. The posted summaries alone cannot show whether that occurred or how sensitive the results are to alternative assumptions. The defined analysis set and the model are reported, but the details needed for a full audit of missing-data handling are not.

11. Confidence Intervals and Precision

The sponsor-reported 95% confidence intervals for the four primary dose comparisons all include zero. Their widths show that the estimates remain compatible with a range of possible differences, not just the point estimate. In general, an interval that includes both negative and positive values leaves uncertainty about the direction of the group contrast as well as its magnitude.

The intervals are not proof of no effect. They indicate that, under the analysis used, the posted results do not distinguish the estimated contrasts from zero at the stated confidence level. Equivalence would require an appropriately designed analysis with a prespecified equivalence margin; such a margin is not reported in the registry record.

The recalculated comparisons use the reported arm summaries and approximate pooled standard deviations. Their intervals are also shown in the tables, but they are unadjusted calculations and are not the sponsor’s model-based confidence intervals. For primary interpretation, the sponsor’s mixed models results should be kept distinct from these summary-based recalculations.

In general, precision depends on the amount of information in the analysis and the variability of the outcome. The primary arm counts differ, and the estimates are not infinitely precise. The registry record does not provide a clinically meaningful difference threshold, so the intervals cannot be classified as excluding a clinically important effect. Readers can use the posted intervals to see the range of uncertainty while avoiding a binary reading based only on whether an interval crosses zero.

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 based on the observed effect, not a reconstruction of the trial’s original design. For these continuous outcomes, the table reports participants per arm, not a total event count. The planning assumptions are a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. The target effect is set to the stated fraction of the observed effect. The trial’s own original sample-size assumptions are not reported in the registry record.

The 80% power column gives the illustrated number of participants per arm for a design with an 80% chance of detecting the specified effect under those assumptions, if that effect and the assumptions were correct. The 90% power column gives the corresponding planning count for a higher chance of detection. These are design calculations, not evidence about what the completed trial could or could not detect.

Rows that plan for a smaller fraction of the observed effect generally require more participants, because smaller group differences are harder to distinguish from random variation. The table makes that relationship visible for each comparison. Large planning counts for a very small observed contrast should not be taken as a recommendation or as a claim about the original protocol.

This is not observed power. The illustrations use the observed effect as an input and should be read as hypothetical planning scenarios, not as a retrospective score of the trial’s success.

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 = -1.300; pooled standard deviation σ = 6.758.
2. n = 2 × (1.96 + 0.842)² × 6.758² / -1.300² = 424.2, rounded up to 425 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 <- -1.3                      # observed difference in means
sigma <- 6.75818290099846                     # 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(2281773)
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 = 425 (80% power), 568 (90% power)
 75% of observed effect: n per arm = 755 (80% power), 1,010 (90% power)
 50% of observed effect: n per arm = 1,697 (80% power), 2,272 (90% power)
power.t.test: 425.2 
Simulated power at n = 425 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-reported primary analysis uses mixed models and reports mean differences in final values with 95% confidence intervals and p-values. The recalculations in the tables use unadjusted contrasts derived from posted arm summaries. Their estimates are close in broad pattern, but they are not the same analysis: the recalculations do not reproduce the sponsor’s model.

For example, the sponsor reports estimates of -1.2, 0.3, 0.3, and -0.6 for the four dose comparisons, while the summary-based recalculations are -1.3, 0.2, 0.3, and -0.7. These values should be read as outputs of distinct methods, rather than as competing versions of a single calculation. The registry record does not flag a discrepancy.

The sponsor-reported 95% confidence intervals all include zero. That is the central precision point for the primary result. The recalculated intervals are useful for illustrating what can be obtained from the posted summaries, but the sponsor’s analysis is the trial’s reported model-based analysis. The record does not provide enough model detail to independently reproduce that analysis from participant-level data.

15. Multiplicity, Missing Data and Analysis Populations

The study compared four BI 409306 dose groups with placebo on the primary cognitive endpoint and included multiple secondary outcomes. The registry describes a two-stage “learn and confirm” approach for cognition: Stage 1 was intended to identify meaningful CANTAB endpoint or endpoints, and the MCCB composite score was selected for Stage 2 after none of the CANTAB outcomes met the prespecified selection criteria at the planned time.

This description clarifies why the MCCB became the primary endpoint, but it does not fully describe the statistical control of multiple comparisons. Any adjustment for the four primary dose comparisons, or for the secondary outcomes, is not reported in the registry record. The record also does not report whether there were interim analyses beyond the described endpoint-selection stage.

The full analysis set is defined as randomized participants who received at least one dose and had a baseline and at least one post-baseline on-treatment MCCB score. The record does not give a breakdown of missing primary endpoint data or exclusions by reason. It does not report how missing observations were handled in the model, whether sensitivity analyses were conducted, or whether other analysis populations were used for the primary inference.

In general, multiplicity affects how confidently isolated findings can be interpreted across many tests, while missing-data assumptions affect which participants contribute information and how their unobserved results are treated. These concerns do not invalidate the posted estimates, but they limit what can be concluded from the registry summary alone. Further protocol and analysis-plan detail would be needed to evaluate them fully.

16. 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 serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: schizophrenia (2.9%), suicidal ideation (1.7%), acute myocardial infarction (0.6%), ischaemic cardiomyopathy (0.6%), empyema (0.6%).

The safety summaries distinguish serious adverse events from other adverse events and give the number at risk in each group. The serious-event table reports 0 affected participants in each of the four BI 409306 groups, with 87, 85, 85, and 86 participants at risk, respectively. The placebo group reports 10 affected participants among 173 at risk. These are counts within the posted safety groups, not comparisons adjusted for exposure or other factors.

The record lists serious-event terms including schizophrenia at 2.9%, suicidal ideation at 1.7%, and acute myocardial infarction, ischaemic cardiomyopathy, empyema, and psychotic disorder at 0.6% each. These listed percentages are registry entries; they should not be used to infer comparative risk without the full event definitions and analysis context.

For other adverse events, the posted affected counts are 5, 9, 9, and 12 across the BI 409306 groups, and 12 in the placebo group. Death counts are not reported in the registry record. A zero serious-event count in an arm does not establish that the risk is zero, and the summaries alone do not establish a causal relationship between an event and study treatment.

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

The primary results describe the measured difference in MCCB outcome between each BI 409306 dose group and placebo over the study period. The sponsor-reported estimates vary in direction, and every reported 95% confidence interval includes zero. The evidence in these posted comparisons therefore remains uncertain about the size and direction of a difference from placebo.

This does not prove that all groups had identical outcomes or that no possible effect exists. A confidence interval that includes zero can contain both no difference and nonzero differences. The registry record also does not report a threshold for a clinically meaningful MCCB change, so the estimates cannot be translated into a judgment about clinical importance.

The outcome is a composite cognitive score, not a direct summary of every aspect of schizophrenia or everyday life. The study also measured secondary outcomes, but multiple outcomes and comparisons make isolated secondary findings less decisive, especially when an adjustment strategy is not reported. Safety summaries are descriptive and do not establish that an event was caused by the study drug.

In general, randomized comparisons estimate group-level contrasts under the design and analysis used. They do not predict an individual participant’s outcome. The record supports a careful description of the estimates and their uncertainty, but does not support a broader claim that treatment is effective or ineffective, nor a claim about outcomes beyond the measured population, treatment period, and endpoints.

18. Why This Trial Matters Statistically

This trial is statistically informative because it makes the endpoint-selection process visible. The record describes a “learn and confirm” structure, explains that the planned CANTAB selection criteria were not met at the planned time, and identifies the MCCB composite as the predefined Stage 2 primary endpoint. That context matters when judging which outcome should anchor interpretation.

It also illustrates why the point estimate alone is not enough. The posted dose-group estimates are not identical, but their sponsor-reported confidence intervals all include zero. The interval conveys uncertainty that a single estimate cannot. The accompanying summary-based recalculations illustrate how estimates derived from arm summaries can differ slightly from the sponsor’s model-based results.

The trial also raises common analysis questions: how multiple dose comparisons are handled, how the analysis population is defined, and how missing observations enter a model. The registry provides some relevant information, including the full analysis set definition, but does not report all details needed to assess multiplicity or missing-data handling.

Finally, the planning illustrations show how sample-size requirements depend on the effect chosen for detection. They are hypothetical calculations based on observed effects and stated assumptions, rather than a record of the original design rationale. Keeping planned power, observed results, model-based inference, and descriptive safety counts distinct is essential to a sound reading of this trial.

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 the original sample-size justification or the trial’s complete statistical analysis plan. It also does not give the full mixed-model specification, including covariates, covariance structure, or detailed missing-data assumptions. These omissions limit independent assessment of the sponsor’s primary analysis.

The primary full analysis set requires treatment exposure and baseline plus at least one post-baseline on-treatment MCCB score. Counts in the primary endpoint summaries are therefore not the same as overall enrollment, and the registry does not provide a participant flow account explaining missingness or exclusion by reason. The posted summaries alone cannot show whether missing measurements could have influenced the estimates.

There are four primary dose comparisons and multiple secondary outcomes. The record describes how the MCCB endpoint was chosen for Stage 2, but does not report the adjustment strategy for all comparisons. This makes isolated secondary findings harder to interpret as confirmatory evidence.

The recalculated comparisons rely on summary statistics, with approximate pooled standard deviations for the continuous outcomes. They are unadjusted and do not reproduce the sponsor’s mixed models analysis. The primary score is a composite, and the registry does not report a threshold for a clinically meaningful change. Finally, the safety summaries are counts and listed event terms; death counts are not reported in the registry record, and these summaries do not establish causality.

20. Related Tutorials

21. Related Statistical Calculators

22. Sources

• ClinicalTrials.gov: A Study to Investigate the Efficacy, Safety and Tolerability of Four Different Doses of BI 409306 Compared to Placebo Given for 12 Weeks in Patients With Schizophrenia on Stable Antipsychotic Treatment. — NCT02281773.
• Registry record last updated 2017-10-19; status last verified 2017-09.
• Results first posted 2017-10-19.
• 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 II study randomized participants with schizophrenia on stable antipsychotic treatment to four once-daily BI 409306 dose groups or placebo for 12 weeks. The primary endpoint was change from baseline in the MCCB composite score, a cognitive measure for which higher scores indicate a better outcome.

The sponsor-reported mixed models comparisons show estimates in both directions, with all four 95% confidence intervals including zero. Summary-based recalculations are also displayed but use a different, unadjusted method. The record describes a staged process leading to MCCB selection, while leaving the full multiplicity and missing-data strategies unreported.

Secondary outcomes and safety summaries provide additional context, but do not resolve the uncertainty in the primary comparisons. The sample-size illustrations are hypothetical planning calculations rather than the original design rationale.