Lurasidone Pediatric Autism Study: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Lurasidone Pediatric Autism Study (NCT01911442), a randomized phase 3 study. The primary outcome is Change in Aberrant Behavior Checklist (ABC) Irritability Subscale Score at Week 6.

2. Clinical Question

This is a question about differences between randomized groups on a measured symptom scale over a defined follow-up period. It is not, by itself, a question about every aspect of autistic disorder, every possible benefit or harm, or longer-term outcomes. The primary endpoint focuses specifically on irritability, as captured by the ABC Irritability Subscale. A group average describes the outcome across participants in that group; it does not show that every participant had the same experience.

The registry record describes the study as designed to evaluate efficacy and safety of two fixed doses compared with placebo. It does not report an explicit superiority margin or a threshold defining a clinically important difference. Whether an observed difference would be important to patients or caregivers therefore cannot be determined from statistical estimates alone. The record gives the score's range and direction, but a minimally important difference is not reported in the registry record.

3. Trial Design

This was a Phase 3, randomized, parallel-group, quadruple-masked study. The official title describes it as a 6-week, fixed-dose, multicenter, double-blind, placebo-controlled study. The registry record identifies an enrollment of 150 and the study status as completed. The study start is listed as 2013-08 and primary completion as 2014-11. The sponsor was Sumitomo Pharma America, Inc.

Participants were assigned to once-daily lurasidone 20 mg, lurasidone 60 mg, or placebo. The reported primary endpoint analysis included 48 subjects in the 20 mg group, 51 in the 60 mg group, and 49 in the placebo group. The endpoint population is defined as randomized subjects who received at least one dose of study medication and had at least one post-baseline assessment in any efficacy variable. This definition is more specific than enrollment: it describes who qualified for the posted efficacy population, rather than everyone who may have been enrolled.

Randomization is intended to make treatment groups comparable on average, including on factors that are measured and factors that are not measured. Masking is intended to reduce the influence of treatment knowledge on care, assessment, and reporting. The record says quadruple masking and describes the study as double-blind in the official title; it does not specify which roles were included in the masking description. Allocation concealment details are not reported in the registry record.

The primary outcome was change in ABC Irritability Subscale score from baseline to 6 weeks. The sponsor's posted primary analyses used Mixed Models Analysis and reported mean differences in final values. The record also lists secondary outcomes covering other behavior scales, caregiver strain, and responder proportions. The protocol's randomization details, planned analysis model specifications, and original sample-size justification 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.

Before assignment: Participants were screened and washed out up to 21 days.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: Intent to treat (ITT) population includes all randomized subjects who receive at least one dose of study medication and have at least one post-baseline assessment in any efficacy variable.

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 disposition table for the overall study used all subjects randomized. The baseline characteristics summary used safety population. 1 subject in placebo group did not take any study medication, and was removed from the safety population.

6. Primary Endpoint

The registry lists the primary outcome as Change in Aberrant Behavior Checklist (ABC) Irritability Subscale Score at Week 6, measured over Baseline to 6 Weeks.

The primary endpoint is change from baseline to Week 6 in the Aberrant Behavior Checklist Irritability Subscale score. The subscale sums 15 items, each rated from 0, Not at all, through 3, Severe in degree. Its total ranges from 0 to 45, and higher values represent greater severity of illness. Thus, the measure turns ratings of irritability-related behavior into a numeric outcome that can be compared across randomized groups.

Change from baseline is used to describe how the measured score differs at follow-up from each participant's starting score. Comparing those changes between groups can help distinguish change associated with assignment to a study group from change observed under placebo during the same follow-up. The endpoint is clinically focused on irritability associated with autistic disorder, not a broad measure of all symptoms or functioning.

The posted arm results are least-squares means with standard errors. The registry's sponsor analyses separately report Mixed Models Analysis mean differences in final values. These are not worded identically as change scores, so sign conventions should be read with care when comparing them. The posted score definition establishes that higher scores indicate greater severity, but the record does not give a threshold for interpreting the size of a change as clinically important. A measured average difference also cannot describe the range of individual responses.

7. Results Posted in the Registry

• Lurasidone 20 mg Once Daily: mean 8.80 versus 7.50 on control; difference 1.30 (95% CI -2.89 to 5.49), standardized effect d = 0.12; z-test p = 0.543.
• Lurasidone 60 mg Once Daily: mean 9.40 versus 7.50 on control; difference 1.90 (95% CI -2.19 to 5.99), standardized effect d = 0.18; z-test p = 0.363.

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) at Week 6

• 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:
  • Lurasidone 20 mg Once Daily: mean -1.10 versus -0.70 on control; difference -0.40 (95% CI -0.87 to 0.07), standardized effect d = -0.34; z-test p = 0.096.
  • Lurasidone 60 mg Once Daily: mean -1.00 versus -0.70 on control; difference -0.30 (95% CI -0.76 to 0.16), standardized effect d = -0.26; z-test p = 0.199.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Lurasidone 20 mg Once Daily 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 Lurasidone 60 mg Once Daily and the control. The interval gives the range of effects that remain compatible with the data.

Change From Baseline in Aberrant Behavior Checklist (ABC) Hyperactivity Subscale Score at Week 6

• 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:
  • Lurasidone 20 mg Once Daily: mean -9.70 versus -7.10 on control; difference -2.60 (95% CI -6.79 to 1.59), standardized effect d = -0.25; z-test p = 0.223.
  • Lurasidone 60 mg Once Daily: mean -6.60 versus -7.10 on control; difference 0.50 (95% CI -3.59 to 4.59), standardized effect d = 0.05; z-test p = 0.811.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Lurasidone 20 mg Once Daily 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 Lurasidone 60 mg Once Daily and the control. The interval gives the range of effects that remain compatible with the data.

Change From Baseline in Children's Yale-Brown Obsessive Compulsive Scales (CY-BOCS) Modified for Pervasive Developmental Disorders (PDDs)

• 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:
  • Lurasidone 20 mg Once Daily: mean -1.00 versus -1.20 on control; difference 0.20 (95% CI -1.12 to 1.52), standardized effect d = 0.06; z-test p = 0.766.
  • Lurasidone 60 mg Once Daily: mean -1.00 versus -1.20 on control; difference 0.20 (95% CI -1.09 to 1.49), standardized effect d = 0.06; z-test p = 0.761.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Lurasidone 20 mg Once Daily 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 Lurasidone 60 mg Once Daily and the control. The interval gives the range of effects that remain compatible with the data.

The secondary outcomes cover severity ratings, hyperactivity, obsessive-compulsive symptoms, caregiver strain, and responder definitions. Detailed numeric results are posted for changes in CGI-S, ABC Hyperactivity, and CY-BOCS. The other listed outcomes are caregiver strain, a CGI-I improvement proportion, and a proportion with at least 25% reduction on the ABC Irritability Subscale. Their numeric results are not reported in the registry record.

Across the detailed continuous secondary measures, the summary-statistic re-calculations do not show narrow separation from placebo: the confidence intervals for the posted comparisons include zero. This broad pattern does not independently confirm a clear difference on the primary outcome, whose own comparisons also have intervals including zero. It is a contextual comparison across outcome domains, not a combined or formal test of overall treatment effect.

Each secondary endpoint measures something different, and results across them should not be treated as interchangeable. The secondary measures may help describe the range of outcomes assessed, but they do not replace the prespecified primary endpoint. The record lists multiple outcomes and analyses, while a multiplicity adjustment strategy is not reported in the registry record. When many endpoints or comparisons are examined, the chance of at least one apparently unusual result can increase. Unadjusted secondary p-values, if available, therefore carry less confirmatory weight than a prespecified primary analysis with a documented multiplicity plan.

The absence of detailed numeric results for some listed secondary outcomes limits conclusions about those domains. It should not be interpreted as evidence of no change or no difference. Taken together, the posted secondary information offers partial context, but it does not settle the primary question or support broad claims beyond the measures and follow-up recorded.

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

• Change From Baseline in Clinical Global Impression-Severity (CGI-S) at Week 6
• Change From Baseline in Aberrant Behavior Checklist (ABC) Hyperactivity Subscale Score at Week 6
• Change From Baseline in Children's Yale-Brown Obsessive Compulsive Scales (CY-BOCS) Modified for Pervasive Developmental Disorders (PDDs)
• Change From Baseline in the Caregiver Strain Questionnaire (CGSQ)
• Proportion of Subjects Who Have CGI-I Score of 1 (Very Much Improved) or 2 (Much Improved) at Week 6
• Proportion of Subjects Who Have at Least 25% Reduction From Baseline to Week 6 in the ABC Irritability Subscale Score.

9. Statistical Interpretation

10. Statistical Methods Explained

The results combine sponsor-reported analyses with simple re-calculations based on posted summary statistics. Those approaches answer related but not identical questions, and their outputs should not be conflated. The sponsor reports Mixed Models Analysis for the primary comparison. The re-calculated estimates use the posted group summaries to form unadjusted comparisons. They provide a useful numerical cross-check, but they cannot reproduce all details of a model-based analysis from summary values alone.

What is a mean difference?

A mean difference compares the average outcome in one group with the average in another. Its sign depends on which group is subtracted from which. Reversing the order reverses the sign without changing the underlying separation between the group averages. This matters in this record: the re-calculated primary differences are expressed as positive values, while the sponsor's reported final-value differences are negative. The estimates are close in magnitude, and the sponsor's reported comparison should be read according to its own contrast convention and analysis description.

The arm values for the primary endpoint are least-squares means, not raw individual observations. A least-squares mean is a model-based estimate of a group mean. The registry record also gives a standard error for each arm estimate. A standard error describes the estimated uncertainty in a mean; it is not the spread of individual participant outcomes. The registry record does not provide the full model specification or the covariates used by the sponsor's Mixed Models Analysis.

What does a confidence interval add?

A confidence interval gives a range of values compatible with the data and the statistical method at a stated confidence level. In repeated use of a method under its assumptions, intervals constructed at that level would cover the underlying parameter at the stated rate. For an individual trial, the interval is best read as a measure of estimation precision, not as a probability statement that the true effect lies inside the particular interval. An interval spanning zero leaves both no difference and differences in either direction compatible with the data.

The re-calculated intervals are based on approximate pooled standard deviations inferred from the reported standard errors and arm sizes. That is why they are described as simple, unadjusted re-calculations rather than as the sponsor's model-based intervals. The re-calculation uses the reported summary statistics and does not have access to participant-level outcome patterns, correlations, missingness, or model terms. The trial's reported primary intervals come from the sponsor's Mixed Models Analysis and are the appropriate reference for what the registry reports as the sponsor's own analysis.

How should p-values be read?

A p-value measures how unusual data at least as incompatible with a null hypothesis would be if that null hypothesis and the analysis assumptions held. It does not give the probability that the null is true, the probability that a finding will replicate, or the size of a treatment effect. The re-calculated primary p-values are unadjusted and arise from the summary-statistic comparison. The sponsor reports its own p-values from Mixed Models Analysis. Differences between the two are expected when analyses use different methods, even if the point estimates are similar.

Neither a p-value above a chosen threshold nor an interval including zero proves that there is no effect. Failure to detect a difference can reflect a small effect, substantial variability, limited precision, or a combination of these. Conversely, statistical evidence of a difference would not by itself establish that its magnitude matters in practice. Interpretation requires the estimate, interval, outcome meaning, and design context together.

Why can over-dispersion matter?

General statistical explanation: over-dispersion occurs when observed variability is greater than a model assumes. In count or event outcomes, that can make standard errors too small if the extra variation is ignored, causing intervals to appear narrower and tests to appear more decisive than they should. The primary outcome is continuous, so over-dispersion is not the central issue for its reported mean comparisons. For other types of outcomes, the chosen model should match the outcome distribution and its variability. The registry record does not provide enough detail to assess every model assumption used for its secondary analyses.

These methods do not turn the posted results into a causal conclusion beyond the randomized comparison and its assumptions. Randomization supports a comparison by assigned group, but interpretation still depends on follow-up, outcome measurement, analysis choices, and missing data. The registry record defines the primary efficacy population, but does not report the detailed handling of missing outcome values, sensitivity analyses, or a full statistical analysis plan. Those omissions limit independent assessment of the model and robustness of the estimates.

11. Confidence Intervals and Precision

The primary re-calculated confidence interval for the 20 mg comparison extends from -2.886 to 5.486; the interval for the 60 mg comparison extends from -2.19 to 5.99. Both are 95% intervals and both include zero. The sponsor's corresponding 95% intervals are -5.6 to 3.0 and -6.1 to 2.2. The sponsor intervals likewise cover zero. Across these analyses, the estimates are not precise enough to rule out a range of differences in either direction.

Precision is about how tightly an analysis estimates a group contrast, not whether a result is important or whether a treatment has an effect. A narrower interval generally means less statistical uncertainty about the estimated contrast, while a wider interval leaves more possible values compatible with the data and method. The primary intervals are wide relative to their point estimates, so the results do not locate the group difference tightly. The posted data do not establish a clinically important threshold against which to judge the width.

The re-calculated intervals and sponsor intervals are not identical. The re-calculations use approximate pooled variability inferred from the reported arm standard errors and group sizes; the sponsor uses Mixed Models Analysis. Such methods can produce different standard errors and intervals. The re-calculations are therefore not a replacement for the sponsor's analysis, and their precision should be interpreted as approximate.

Confidence intervals also do not summarize individual variation. The arm standard errors describe uncertainty in estimated means; they do not show the full distribution of participant scores. Individual outcomes may vary substantially around a group average, and the registry record does not report the distribution needed to describe that variation fully. In general, a nonsignificant comparison with a broad interval is inconclusive about the size of a possible difference, rather than evidence of exact equality.

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 planning figures are illustrative calculations for participants per arm, not total event counts. They use the assumptions stated for these calculations: a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. The detectable effect is set as a stated fraction of the observed effect. The original trial's own sample-size assumptions and justification are not reported in the registry record.

The 80% power column represents the planned number of participants per arm for a hypothetical study designed to have an 80% chance of detecting the specified effect under those assumptions, if that effect were the true effect and the model assumptions held. The 90% power column expresses the analogous planning target for a 90% chance. These are prospective planning illustrations, not observed power; observed power is not a useful substitute for reading the estimated effect and its confidence interval.

Rows that plan for 100% of the observed effect use the full effect estimated in this trial as the target for a hypothetical future study. Rows for 75% and 50% of the observed effect use smaller targets. Smaller effects are harder to distinguish from random variation, so they require more participants per arm to reach the same planned power under otherwise unchanged assumptions. The figures can therefore increase sharply as the target effect becomes more modest.

These calculations are sensitive to the observed effect, which is uncertain. A small trial estimate may overstate or understate the effect that would be seen in another study. The planning numbers should not be read as a recommendation, as the sponsor's original target, or as evidence about whether the completed study had adequate power. They are illustrations of how a specified effect size and statistical assumptions translate into a planning sample size.

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 = 10.518.
2. n = 2 × (1.96 + 0.842)² × 10.518² / 1.300² = 1,027.6, rounded up to 1,028 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 <- 10.5181851546438                     # 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(1911442)
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 = 1,028 (80% power), 1,376 (90% power)
 75% of observed effect: n per arm = 1,827 (80% power), 2,446 (90% power)
 50% of observed effect: n per arm = 4,111 (80% power), 5,503 (90% power)
power.t.test: 1028.6 
Simulated power at n = 1,028 per arm: 0.802 (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 reports the primary comparisons using Mixed Models Analysis. For lurasidone 20 mg versus placebo, the reported mean difference in final values is -1.3, with a 95% confidence interval from -5.6 to 3.0 and p = 0.546. For lurasidone 60 mg versus placebo, the reported mean difference is -1.9, with a 95% confidence interval from -6.1 to 2.2 and p = 0.359. Both reported intervals include zero.

The simple re-calculations from the arm summary statistics give differences of 1.3 and 1.9, with 95% intervals of -2.886 to 5.486 and -2.19 to 5.99 and unadjusted p-values of 0.543 and 0.363. The signs are opposite because the comparisons are expressed in different directions; the magnitudes are close. The re-calculated intervals and p-values are also close to, but not identical to, the sponsor's, as expected when an approximate summary-statistic calculation is compared with a model-based analysis.

The registry record does not flag a discrepancy between the posted and re-calculated primary results. The sponsor's Mixed Models Analysis is the reported trial analysis; the re-calculation is a transparent, unadjusted check from summaries and should not be substituted for the sponsor's model. The record does not give enough detail about the model specification to reproduce it independently.

15. Multiplicity, Missing Data and Analysis Populations

Multiplicity arises when a study tests more than one outcome, dose comparison, time point, or subgroup. If many hypotheses are tested, the chance of at least one result appearing statistically unusual by chance can rise. Common approaches include prioritizing a primary endpoint, controlling an overall error rate, or treating secondary outcomes as exploratory. Which approach was planned for this study is not reported in the registry record. The record lists one primary endpoint and multiple secondary outcomes, but it does not describe a hierarchy, adjustment procedure, or confirmatory testing sequence.

The sponsor's primary analyses are reported separately for the two dose comparisons. The registry record does not say whether the statistical plan adjusted for making both comparisons with placebo. The re-calculated p-values are explicitly unadjusted. They should not be interpreted as multiplicity-controlled evidence. Secondary results require additional caution because the record includes several outcomes and does not specify a correction strategy.

The primary efficacy population includes randomized subjects who received at least one dose of study medication and had at least one post-baseline assessment in any efficacy variable. This definition excludes randomized participants who did not meet those conditions from the stated population. The registry record does not report the number excluded from this definition, the amount or pattern of missing outcome data, or the method used to handle missing values in the primary analysis. It also does not report sensitivity analyses that would show how results change under different missing-data assumptions.

Mixed Models Analysis can use available repeated measurements under assumptions about the missing-data process, but the model details and assumptions used by the sponsor are not reported in the registry record. No conclusion about whether missingness was harmless or whether alternative approaches would change the estimate can be drawn from the posted information alone. The record also does not provide a full statistical analysis plan, interim-analysis details, or the original sample-size justification. These are unknowns, not evidence that a particular procedure was or was not used.

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 4% reporting threshold.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: torus fracture (2.0%), upper limb fracture (2.0%), irritability (2.0%), appendicitis (2.0%).

The registry record lists serious adverse events in the safety groups as follows: 3 affected participants among 49 at risk in the lurasidone 20 mg group, 2 among 51 in the 60 mg group, and 0 among 49 in the placebo group. These are counts of participants affected, not a measure of the total burden, duration, or severity of every adverse event. The record also lists common serious-event terms, each at 2.0%: Torus Fracture, Upper Limb Fracture, Irritability, and Appendicitis. The posted terms should not be assigned to a particular arm unless the record does so.

The registry record does not report death counts. The figures do not by themselves establish that a treatment caused an event or that groups differ in safety; event counts are descriptive and the study follow-up was 6 weeks. Assessment of safety also depends on event definitions, exposure time, event details, and the completeness of reporting. Further detail on those matters is not reported in the registry record. These figures are not a substitute for a full safety assessment.

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

The posted primary data describe a randomized comparison of two fixed lurasidone doses with placebo on change in ABC Irritability Subscale score over 6 weeks. The sponsor's estimates and the summary-statistic re-calculations are imprecise: the reported primary confidence intervals include zero. The available results do not provide a precise estimate of the direction or size of a group difference.

This does not establish that lurasidone and placebo are equivalent, nor does it prove that there is no difference. Equivalence or non-inferiority requires a defined margin and an analysis designed for that question; such a margin is not reported in the registry record. A result that does not demonstrate a difference may leave clinically relevant values compatible with the data, particularly when intervals are broad.

The endpoint is a scale score, not a direct measure of every feature that matters to children, adolescents, or caregivers. The listed secondary outcomes broaden the measurement domains, but detailed results are not reported for every listed outcome. No numerical evidence should be generalized beyond the studied groups, dose regimens, outcomes, and follow-up period without further evidence.

Group averages also conceal individual variation. An average change does not show how many participants improved, remained similar, or worsened unless the relevant individual-level or responder data are reported. Although the registry lists responder outcomes among the secondary measures, their numerical results are not reported in the registry record. The available safety counts are likewise descriptive and do not establish causation.

The most defensible reading is therefore limited: the registry reports no clearly separated primary group comparison in these analyses, and the uncertainty intervals leave a range of possible differences. The record does not provide enough detail to determine whether different missing-data assumptions, model specifications, or multiplicity procedures would materially alter interpretation.

18. Why This Trial Matters Statistically

This trial is statistically informative as an example of the difference between a point estimate and the precision around it. The primary point estimates are modest relative to their confidence intervals, and the intervals cover zero. Reporting only a point estimate or a p-value would conceal how much uncertainty remains about the comparison. Reading the estimates with their intervals makes that uncertainty explicit.

The study also illustrates why analysis details matter. The sponsor used Mixed Models Analysis, while the summary-statistic re-calculations use approximate variability and an unadjusted comparison. The resulting values are close but not identical, and signs depend on contrast direction. The results cannot be understood accurately without knowing whether a value refers to a change, a final value, or which group is subtracted from which.

Multiple doses and multiple outcomes create questions about multiplicity, while the registry's limited description of the analysis plan leaves those questions unresolved. The record also defines a specific efficacy population, but does not detail missing-data handling. These are not merely technical footnotes: they affect the strength and scope of inferences that can be made from posted estimates.

Finally, the planning illustrations show that sample-size requirements depend strongly on the effect a future study aims to detect. Smaller target effects require more participants per arm under the stated assumptions. Because planning based on an observed effect inherits uncertainty from the trial estimate, such figures are best understood as illustrations rather than as a definitive prescription or retrospective judgment of trial adequacy.

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 limitations constrain interpretation of this particular record. The primary follow-up is 6 weeks, so the posted endpoint does not describe longer-term outcomes. The population is children and adolescents with irritability associated with autistic disorder in a community setting; findings should not automatically be extended to other populations or clinical settings.

The efficacy population requires at least one dose and at least one post-baseline efficacy assessment. The registry record does not report how many randomized participants did not meet that definition, how much outcome data were missing, or how missingness was handled. Without those details, the potential influence of exclusions and missing outcomes cannot be assessed.

The sponsor's primary analysis is identified as Mixed Models Analysis, but the model specification, covariates, and assumptions are not reported in the registry record. The simple re-calculations are approximate, unadjusted comparisons and do not reproduce participant-level modeling. The statistical plan, multiplicity procedure, interim analyses, and sample-size justification are also not reported in the registry record.

Although several secondary outcomes are listed, detailed numeric results are not reported for all of them. The available summary measures do not show participant-level distributions, and a clinically important threshold for the primary scale is not reported. Safety information is limited to posted serious-event counts and selected terms; deaths are not reported in the registry record. These limits restrict assessment of clinical importance, robustness, individual response patterns, and the full safety profile.

20. Related Tutorials

21. Related Statistical Calculators

22. Sources

• ClinicalTrials.gov: Lurasidone Pediatric Autism Study — NCT01911442.
• Registry record last updated 2016-02-25; status last verified 2016-01.
• Results first posted 2016-02-25.
• 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

NCT01911442 is a completed Phase 3 randomized, parallel, quadruple-masked, placebo-controlled study of once-daily lurasidone 20 mg and 60 mg in children and adolescents with irritability associated with autistic disorder. The study evaluated the treatments over 6 weeks, with change in ABC Irritability Subscale score as the primary endpoint.

The sponsor reported Mixed Models Analysis comparisons for both doses against placebo. Their 95% confidence intervals include zero. Summary-statistic re-calculations are similar in magnitude but approximate and unadjusted. The registry record also lists secondary outcomes and serious adverse event counts, but detailed results are not reported for every listed secondary outcome. Information on multiplicity, missing-data handling, the original sample-size justification, and deaths is not reported in the registry record.