Study of Aripiprazole (Abilify) Versus Placebo in Children With Subsyndromal Bipolar Disorder: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Study of Aripiprazole (Abilify) Versus Placebo in Children With Subsyndromal Bipolar Disorder (NCT00194012), a randomized phase 3 study. The primary outcome is Young Mania Rating Scale (YMRS).

2. Clinical Question

3. Trial Design

The registry identifies this as a phase 3, randomized, quadruple-masked, parallel-group study comparing aripiprazole with placebo. The record lists an enrollment of 62 and a status of completed. The study start is recorded as 2004-08, and primary completion as 2012-05. The sponsor is University Hospitals Cleveland Medical Center.

Participants assigned to aripiprazole received pills at doses of 2mg, 5mg, 7mg, 10mg, 12mg or 15mg, depending on their response. Those assigned to placebo received pills and dosing made to look identical to aripiprazole. The registry does not report the allocation ratio, the duration of the randomized treatment period, or the detailed masking procedures. The original protocol’s sample-size justification and design assumptions are not reported in the registry record.

For the posted baseline YMRS measure, the aripiprazole arm has a reported sample size of 30 and the placebo arm 29. These are the arm counts attached to that measure; other outcomes and safety summaries have different posted counts. The registry record also lists open-label extension safety groups, which are distinct from the randomized-phase comparison described by the primary outcome summaries. The record does not report how participants entered those extension groups or provide randomized comparisons for them.

The primary outcome is a baseline score, while a posted secondary CDRS-R measure is recorded at 12 weeks. That distinction matters: a baseline score describes status at the start of the measured period and cannot, by itself, show a treatment-related change. The registry record does not report an analysis plan or explain how the listed outcome measures were selected for analysis.

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: After screening and baseline assessments study participants were assigned to aripiprazole or placebo. Treatment was initiated at a dose of 0.1mg/kg/d. The dose could be increased by approximately 0.05 mg/kg/d at each study visit if patient experienced residual symptomology with no intolerable side effect to maximum dose 15mg/d.

Randomized Phase

Reasons for not completing this period:

Open Label Extension (OLE)

Reasons for not completing this period:

5. Baseline Characteristics

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

6. Primary Endpoint

The registry lists the primary outcome as Young Mania Rating Scale (YMRS), measured over Baseline.

The Young Mania Rating Scale (YMRS) is a rating of manic symptom severity. The registry describes it as an 11-item scale based partly on the child’s subjective report of their clinical condition over the previous 48 hours and partly on observations made during a clinical interview. Items are rated for severity and draw on descriptions of core mania symptoms.

Four items—irritability, speech, thought content, and disruptive or aggressive behavior—are graded on a 0 to 8 scale; the remaining seven items are graded on a 0 to 4 scale. The four higher-range items receive twice the weight of the others, which the registry says is intended to compensate for poor cooperation from severely ill patients. The total score ranges from 0 to 60, with higher scores indicating more severe mania.

A rating scale such as the YMRS gives clinicians a structured way to represent symptom severity and makes scores available for statistical description. In this record, however, the primary YMRS timeframe is baseline. The posted means and standard deviations therefore describe the randomized groups at baseline. The registry record does not report a primary post-treatment YMRS result or a treatment-effect analysis for this measure.

7. Results Posted in the Registry

• Abilify Randomized Phase: mean 24.80 versus 21.60 on control; difference 3.20 (95% CI -0.09 to 6.49), standardized effect d = 0.49; z-test p = 0.057.

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.

Children's Depression Rating Scale-Revised (CDRS-R )

• 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:
  • Aripiprazole Randomized Phase: mean 33.70 versus 30.30 on control; difference 3.40 (95% CI -2.26 to 9.06), standardized effect d = 0.31; z-test p = 0.239.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Aripiprazole Randomized Phase and the control. The interval gives the range of effects that remain compatible with the data.

CDRS-R Children's Depression Rating Scale-Revised

• 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:
  • Aripiprazole Randomized Phase: mean 24.40 versus 25.90 on control; difference -1.50 (95% CI -5.70 to 2.70), standardized effect d = -0.19; z-test p = 0.484.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Aripiprazole Randomized Phase and the control. The interval gives the range of effects that remain compatible with the data.

Children's Global Assessment Scale (CGAS)

• 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:
  • Aripiprazole Randomized Phase: mean 56.10 versus 57.50 on control; difference -1.40 (95% CI -4.41 to 1.61), standardized effect d = -0.24; z-test p = 0.362.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Aripiprazole Randomized Phase and the control. The interval gives the range of effects that remain compatible with the data.

The secondary measures add descriptive context across symptom severity, depression ratings, and global functioning. The posted record includes CDRS-R at baseline and 12 weeks, and CGAS at baseline. The recalculated contrasts do not show a consistent pattern of clear separation: the direction of the CDRS-R mean difference changes between its baseline and 12-week summaries, and the CGAS baseline contrast points in the opposite direction from the baseline YMRS contrast. The confidence intervals for these comparisons include zero.

This pattern should be considered alongside the primary result, which is itself a baseline YMRS comparison rather than a post-treatment estimate. Neither the primary baseline measure nor the posted secondary summaries provide a reported analysis of treatment-related change. Therefore, the secondary outcomes do not establish whether the groups followed different courses during treatment or whether any observed differences were attributable to assignment.

The registry lists 12 secondary outcomes as posted, but the record does not report analyses for the secondary summaries described. The posted outcome descriptions include repeated labels, and the record does not explain the relationship among all listed outcomes, timepoints, or analyses. No multiplicity adjustment or hierarchy of secondary hypotheses is reported.

In general, examining multiple outcomes and timepoints creates more opportunities for an apparently notable result to arise by chance. Unadjusted secondary results consequently carry less weight than a clearly prespecified primary analysis with a reported analysis plan. They are useful for describing what was measured and identifying questions for further study, but they should not be treated as independent confirmation of a treatment effect. The absence of reported analyses and multiplicity procedures limits how strongly any secondary pattern can be interpreted.

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

• Children's Depression Rating Scale-Revised (CDRS-R )
• CDRS-R Children's Depression Rating Scale-Revised
• CDRS-R Children's Depression Rating Scale-Revised
• Children's Global Assessment Scale (CGAS)
• Children's Global Assessment Scale (CGAS)
• Children's Global Assessment Scale (CGAS)
• Clinical Global Impressions Scale (CGI-Severity)
• Clinical Global Impressions Scale (CGI-Severity)

9. Statistical Interpretation

10. Statistical Methods Explained

For continuous outcomes, the registry reports a mean and standard deviation for each arm. The mean is the arithmetic average of the observed scores in that group. The standard deviation describes how spread out individual scores are around their group mean; it is not the uncertainty in the mean itself. The posted YMRS summaries are means of a continuous measure, with standard deviations as the measure of spread.

What does a mean difference represent?

The recalculated mean difference subtracts the placebo mean from the aripiprazole mean. For baseline YMRS, that unadjusted difference is 3.2 points. A positive value means that the posted mean is higher in the aripiprazole arm; it does not imply a worse or better treatment outcome, because the measure is at baseline. The calculation uses only the posted group summaries and does not adjust for other participant characteristics or account for a model specified in a trial analysis plan.

The same general approach is used for the continuous secondary measures. These are simple comparisons of the posted group means at the stated timeframes. They are not substitutes for repeated-measures models, adjusted analyses, or analyses of change from baseline. The registry record does not report such analyses.

Why consider a confidence interval?

In general, a confidence interval represents statistical uncertainty around an estimated quantity under the assumptions of the method used to construct it. For a mean difference, an interval that includes zero is compatible with no difference as well as with differences in either direction. It should not be read as a list of values that are equally likely, and it does not show the range of individual participants’ scores. The intervals attached to the recalculated comparisons are based on posted summary statistics, rather than on a reported trial model.

The p-values in the comparisons are unadjusted. In general, a p-value measures how unusual data at least as discrepant as those observed would be under a specified null hypothesis and the assumptions of the test. It is not the probability that the null hypothesis is true, and it does not measure the size or clinical importance of a difference. Because the registry record does not report its own analysis for these outcomes, the recalculated p-values should be treated as descriptive analytical aids, not as sponsor-reported findings.

What do standard deviations and sample sizes contribute?

In general, the variability of scores and the number of observations both affect the precision of a comparison. More variability makes a group mean less precise, while a larger sample can make it more precise. The standard deviations show observed spread in each arm. The arm counts show how many observations are attached to the posted summary, but do not explain whether any participants lacked outcome data or why counts differ across outcomes. The registry record does not report that information.

For these recalculations, a pooled standard deviation is used to express the combined spread across the two arms, and a standardized difference is also computed. Such standardized quantities place a mean difference in relation to variability, but they do not provide an adjusted treatment effect. The recalculated quantities are not reported trial analyses. The registry record does not report covariate adjustment, the analysis population, or a model that would support a different interpretation.

What does the registry report about analysis methods?

The record lists no reported analyses for the posted primary and secondary outcomes. It also does not report a statistical analysis plan, a pre-specified significance threshold for the trial’s own outcome testing, an interim-analysis plan, or procedures for missing data. Their absence from the record does not establish that the trial had no such procedures; it means they are not reported in the registry record. The recalculations use the posted summary statistics and should not be confused with methods chosen or endorsed by the sponsor.

In general, these distinctions matter because a trial’s estimand, analysis population, handling of incomplete measurements, and adjustment strategy determine what a statistical result represents. A simple comparison of available means has a narrower interpretation than an analysis specified in advance and applied to a defined population. Here, the primary YMRS measurement is explicitly baseline, making its posted comparison descriptive regardless of the recalculation method.

11. Confidence Intervals and Precision

The primary YMRS mean difference is 3.2 points, with an unadjusted recalculated confidence interval whose lower limit is below zero and upper limit above zero. Thus, the interval includes zero and is compatible with no baseline mean difference as well as with differences in either direction under the calculation’s assumptions. Its width reflects uncertainty in the estimate; the estimate alone should not be treated as exact.

All three posted secondary comparisons also have recalculated confidence intervals that include zero. The intervals are broad relative to their point estimates, so the data do not pin down a single precise mean difference. This is a description of statistical precision, not a claim that the true difference is zero. Nor does an interval that crosses zero establish that the groups are equivalent.

These intervals were recalculated from summary statistics and are not sponsor-reported model-based intervals. They do not account for an analysis plan or adjustment strategy that is not reported in the registry record. As the primary measurement is at baseline, its interval concerns a baseline contrast rather than a treatment effect. The registry record does not report a post-treatment primary YMRS interval.

In general, greater precision would be reflected by a narrower interval, while an interval spanning a wide range of plausible differences signals less precision. Interpretation should consider the size and direction of the estimate, its uncertainty, the outcome’s timing, and the analysis context—not only whether the interval includes zero. The posted summaries do not establish the clinical importance of a difference.

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 the trial’s original design calculations. The listed figures are participants per arm for 80% power and participants per arm for 90% power. They are not total participants across both arms and are not event counts; the event-count convention applies to time-to-event outcomes, and these outcomes are continuous measures.

The planning assumptions are a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. For each illustration, the effect to detect is set at the stated fraction of the observed effect: 100%, 75%, or 50%. The registry record does not report the assumptions used to design the original trial, so these planning figures should not be interpreted as a reconstruction of its protocol.

In plain language, the 80% power column gives the illustrative number of participants per arm for a design with an 80% chance of detecting the specified effect under the stated assumptions. The 90% column describes a design with a higher chance of detecting that effect, if it is present and the assumptions hold. Smaller effects need more participants to detect reliably, which is why the figures increase when planning for a smaller fraction of the observed effect.

These sample-size illustrations are not observed power. They do not tell whether the trial was adequately designed, and they do not change the interpretation of the posted estimates or their uncertainty. The observed effects used as starting points come from the posted comparisons and may be imprecise.

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 = 3.200; pooled standard deviation σ = 6.466.
2. n = 2 × (1.96 + 0.842)² × 6.466² / 3.200² = 64.1, rounded up to 65 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 <- 3.2                      # observed difference in means
sigma <- 6.46561756727969                     # 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(194012)
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 = 65 (80% power), 86 (90% power)
 75% of observed effect: n per arm = 114 (80% power), 153 (90% power)
 50% of observed effect: n per arm = 257 (80% power), 344 (90% power)
power.t.test: 65.1 
Simulated power at n = 65 per arm: 0.799 (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. Multiplicity, Missing Data and Analysis Populations

The registry record lists 15 outcome measures and 12 posted secondary outcomes, but it does not report a multiplicity strategy. It does not describe a hierarchy among outcomes, a planned order of testing, or an adjustment of significance levels for multiple comparisons. Without that information, the unadjusted secondary p-values cannot be interpreted as if each were the sole planned test.

In general, multiplicity matters when multiple outcomes, timepoints, or comparisons are examined: the more opportunities there are to identify a small p-value, the greater the chance of a seemingly notable finding arising by chance. Adjustments or prespecified testing hierarchies can address this concern, but the registry record does not say whether any such approach was used. No procedure should be inferred from the listed outcomes alone.

The record also does not report how missing outcome measurements were handled, whether analyses used an intention-to-treat population, or which participants were included in each formal analysis. Posted arm counts vary by measure and differ from the enrollment field, but the record does not explain the reasons. Counts should not be used to infer a missing-data method, exclusions, or participant flow.

The primary endpoint is listed at baseline, and no primary analysis is reported. The registry record does not state whether the outcome was analyzed as a baseline covariate in another model, whether change scores were planned, or whether any analysis was prespecified. It also does not report interim analyses or a trial-specific statistical analysis plan. These omissions restrict assessment of selective reporting, analytic choices, and the relationship between the listed outcomes and the trial’s original hypotheses. They do not prove that procedures were absent; they indicate that they are not reported in the registry record.

15. Safety Results

The registry reports serious adverse events as affected participants divided by participants at risk.

Deaths

Other (non-serious) adverse events

Participants with at least one non-serious event at or above the registry's 5% reporting threshold.

Most frequent non-serious adverse event terms

For the randomized phase, the registry reports zero serious adverse events among 31 participants at risk in the aripiprazole group and zero among 31 at risk in the placebo group. It also reports zero deaths in both groups. In the open-label extension groups, the record reports zero serious adverse events among 22 participants at risk for aripiprazole and among 21 at risk for placebo, with zero deaths in each group. No serious adverse-event terms are listed.

These are event counts as recorded, not evidence that the risk of serious events is absent or that the treatments have equivalent safety. Zero reported events in these groups does not establish that uncommon events could be excluded. The open-label extension groups should not be interpreted as a randomized safety comparison; the registry record does not report their assignment process in detail.

The registry also lists other adverse events in the randomized and extension groups. Those counts are not serious-event counts, and the record does not provide event terms in the summary. The at-risk counts used for safety are not the same as all posted outcome counts, and the record does not explain the differences.

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

The most direct interpretation of the primary result is descriptive: the posted baseline YMRS mean was higher in the aripiprazole arm than in the placebo arm, with a recalculated mean difference of 3.2 points. The recalculated interval includes zero, and the unadjusted p-value is 0.057. Because the outcome is recorded at baseline, these results do not estimate an effect of treatment on mania symptoms. The registry record does not provide a primary post-treatment YMRS analysis.

The secondary summaries describe CDRS-R and CGAS measurements at their stated timeframes, but their recalculated intervals also include zero. The registry record does not report formal analyses for these outcomes or explain missing-data handling and multiplicity. They therefore offer limited evidence for conclusions about differences between treatment groups over time.

The record describes randomized assignment, masking, the interventions, outcome summaries, and adverse-event counts. It does not provide enough reported statistical detail to determine how the trial’s original investigators analyzed the outcomes or whether the posted summaries correspond to a prespecified analysis. Recalculations based on summary statistics can illuminate the scale and uncertainty of a contrast, but they cannot substitute for an unreported trial analysis.

These results do not show that aripiprazole works or does not work for children with the described symptoms and family history. They also do not demonstrate equivalence between aripiprazole and placebo. A baseline difference, absence of a reported treatment-effect analysis, and imprecise secondary comparisons leave the treatment question unresolved in the registry record. The findings should be understood as a limited account of the posted data, not as a basis for medical decisions.

17. Why This Trial Matters Statistically

This trial illustrates why the timing of an endpoint is as important as the name of the measure. YMRS is a symptom-severity rating, but the registry lists its primary timeframe as baseline. A numerical contrast between baseline means cannot answer whether the groups changed differently after assignment. Reading the score as a treatment outcome would give the statistic a meaning that the recorded timeframe does not support.

The record also demonstrates the difference between posted descriptive summaries and reported inferential analyses. Means, standard deviations, and arm counts permit simple recalculations, including a mean difference and an uncertainty interval. Those calculations can help readers assess the scale and precision of the summaries, but they do not reveal the sponsor’s analysis plan, adjustments, or missing-data decisions. The registry lists no reported analyses for the outcomes considered.

Finally, secondary outcomes and safety counts provide useful context but do not resolve the primary question. Multiple outcomes raise multiplicity concerns when no testing hierarchy is reported, and zero serious adverse events in the posted groups cannot establish absence of risk. The trial is therefore a useful statistical example of cautious interpretation: identify what was measured, when it was measured, what calculation was made, and what remains unreported before drawing conclusions.

18. Limitations

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

A key limitation of the record is that the primary YMRS outcome is reported at baseline, while no primary post-treatment result or treatment-effect analysis is listed. The posted primary summaries therefore cannot show change attributable to randomized assignment. The secondary CDRS-R summary at 12 weeks provides a later measurement, but the record does not report a formal longitudinal analysis or a corresponding primary YMRS result.

The registry does not report the statistical analysis plan, original sample-size justification, missing-data procedures, analysis populations, or multiplicity approach. The stated enrollment and the counts attached to individual outcome and safety summaries differ, and the record does not explain participant flow or why the counts vary. It is not possible to infer the reasons from the counts alone.

Dosing in the aripiprazole group varied according to response, so the record describes more than one dose rather than a single fixed regimen. The detailed titration schedule, treatment duration, and rules for dose changes are not reported. The registry also includes open-label extension safety groups, but their relationship to the randomized groups and their entry criteria are not described in the posted summary.

Finally, the statistical comparisons presented with the summaries are unadjusted recalculations rather than reported sponsor analyses. The registry record does not provide enough information to verify whether other analyses were conducted or how any such analyses were selected. These limitations constrain interpretation of both the effectiveness question and the safety summaries.

19. Related Tutorials

20. Related Statistical Calculators

21. Sources

• ClinicalTrials.gov: Study of Aripiprazole (Abilify) Versus Placebo in Children With Subsyndromal Bipolar Disorder — NCT00194012.
• Registry record last updated 2017-06-26; status last verified 2017-06.
• Results first posted 2014-05-05.
• Data retrieved from ClinicalTrials.gov on 2026-11-09.
• Group sizes, means, event counts and sponsor-reported p-values are reproduced as posted. Effect estimates, confidence intervals and p-values not attributed to the sponsor are independently calculated from the posted summary statistics.

22. Record Summary

The registry describes a completed, randomized, quadruple-masked, parallel-group phase 3 comparison of aripiprazole and placebo in children with subsyndromal bipolar symptoms and a family history of mood disorders. The primary measure is baseline YMRS, reported as a mean and standard deviation in each arm. The unadjusted recalculated baseline difference is 3.2 points, and its confidence interval includes zero.

Secondary summaries cover CDRS-R and CGAS measures, including a CDRS-R measurement at 12 weeks. The registry record lists no reported analyses for the outcomes considered, and does not report the trial’s analysis plan, missing-data methods, or multiplicity procedures. Serious adverse-event and death counts are zero in each listed group. The record supports descriptive interpretation of the posted summaries, but does not provide a reported primary treatment-effect analysis.