Side Effect Study of Antipsychotic Medicines to Treat Childhood Bipolar Disorder: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Side Effect Study of Antipsychotic Medicines to Treat Childhood Bipolar Disorder (NCT00746252), a randomized na study. The primary outcome is Weight Gain.

2. Clinical Question

The study record does not establish a broader clinical question about all children with bipolar disorder or all antipsychotic medicines. Its posted results concern these named arms and the recorded weight-gain outcome. Because the registry lists a very small enrollment and incomplete summary information, the results offer limited precision for comparing the arms. The posted record also does not report detailed eligibility criteria, participant characteristics, or a fuller description of how treatment was delivered. Those omissions limit assessment of who the findings represent and of how the comparison was conducted.

3. Trial Design

The registry describes a randomized, parallel-group study with no masking. The two listed interventions were risperidone and aripiprazole, with one arm for each medicine. The study is marked terminated. The listed enrollment is 5. Its start is recorded as 2008-06 and its primary completion as 2010-10. The phase is recorded as NA, which does not provide a phase designation.

The primary outcome was weight gain, treated as a continuous measure and reported as a mean in pounds. The stated outcome period was from baseline to last observation up to 12 weeks. The timing description says measurements were done biweekly from baseline up until 12 weeks; the outcome description adds that the last observation was carried to 12 weeks. The registry lists one outcome measure and one analysis. The reported comparison method is a two-sided t-test.

Random allocation means assignment was intended to be determined by chance, but the registry record does not report the sequence-generation method, allocation concealment, or the procedure used to implement assignment. No masking was used, according to the record. Further details about study visits, concomitant care, adherence, withdrawals, or the reasons for termination are not reported in the registry record. The record also does not state a detailed statistical analysis plan or explain how the study's enrollment was chosen.

The listed enrollment and arm-level outcome counts are not interchangeable descriptions: the registry gives an enrollment of 5, while the weight-gain summaries list 1.0 for risperidone and 2.0 for aripiprazole. The record does not explain this difference or identify the analysis population in a way that resolves it. These details matter because a reader cannot assume that every enrolled participant contributed to the posted outcome summaries.

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: Two participants enrolled and completed baseline assessments but were withdrawn from the study at baseline due to the following reasons: 1. Abnormal glucose tolerance test requiring further assessment with endocrinology. 2. Unable to complete lab work due to increased distress.

Overall Study

Reasons for not completing this period:

5. Baseline Characteristics

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

6. Primary Endpoint

The registry lists the primary outcome as Weight Gain, measured over These measurements are done biweekly from baseline up until 12 weeks.

The primary endpoint is weight gain from baseline to the last observation available up to 12 weeks. It is expressed in pounds and summarized as a mean for each treatment arm. The registry says measurements were taken biweekly from baseline up until 12 weeks, and that the last observation was carried to 12 weeks. This defines the reported endpoint as an arm-level summary of change over the stated period, rather than a measurement at every scheduled visit.

Weight is a continuous outcome: values can take a range of amounts rather than representing only whether an event occurred. A mean provides a compact description of the average outcome in an arm, while a measure of dispersion such as a standard deviation would describe how spread out participant outcomes are around that mean. In the posted results, the aripiprazole arm has a standard deviation of 1.48 pounds. A standard deviation for risperidone is not reported in the registry record.

Weight was selected in this study because the registry's stated purpose concerns weight gain and related side effects during treatment with antipsychotic medicine. The outcome description specifies a baseline-to-last-observation measure, with the final observation carried to 12 weeks. That rule can make follow-up observations represent a common stated endpoint, but it does not show how much information was available for each participant or whether observations were missing. Details on those matters are not reported in the registry record.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

What does the reported t-test assess?

A t-test is a method for comparing means. In this record, the stated method is a two-sided t-test, meaning the test considers departures in either direction rather than testing only whether one named arm has a larger mean. Its null model represents no difference in the arm means. The reported p-value, .495, is calculated under that model and the test assumptions. It does not tell the reader the probability that the null model is correct, nor does it quantify the size of the arm contrast.

For a conventional independent-groups t-test, the comparison relies on the arm observations being independent and on the summaries adequately representing the outcome distributions. The exact version of the test and its implementation are not reported in the registry record. The posted record does not supply individual outcomes, the risperidone standard deviation, or a confidence interval. Consequently, the reported p-value cannot be independently reconstructed from the posted arm summaries. A recalculated unadjusted comparison is not available from those summaries.

Why do the mean and spread both matter?

The mean describes the average weight gain recorded in an arm. A standard deviation describes how dispersed participant measurements are around that mean. These quantities answer different questions: the mean describes the center of the observed values, while the spread describes their variability. The registry reports a mean of 6.65 pounds and a standard deviation of 1.48 for aripiprazole, but only a mean of 8.5 pounds for risperidone. The missing risperidone spread is not reported in the registry record.

General statistical interpretation requires caution when outcome summaries are based on very limited observations. A mean can be strongly affected by any one participant's result when an arm contains very little information. A standard deviation is also difficult to estimate reliably from a sparse group. These constraints affect both descriptive interpretation and the reliability of a test comparing means. The registry's p-value is the reported analysis, but the available arm summaries do not permit a reader to evaluate its calculation fully.

What does a confidence interval add?

A confidence interval gives a range of values that are compatible with the data and the method used, under the assumptions of the analysis. It communicates the uncertainty of an estimate, such as a mean or a difference between means. An interval around a treatment contrast is often more informative than a p-value alone because it shows both the estimated direction and the range of plausible magnitudes. No confidence interval for the arm comparison is reported in the registry record. Nor is a numerical contrast estimate posted. The uncertainty of the comparison therefore cannot be read directly from a reported interval.

A non-significant-looking p-value, considered without a confidence interval or a threshold specified in the record, should not be treated as proof of equivalence. Equivalence requires a defined margin and an analysis designed to assess whether the plausible difference lies within that margin. Such a margin and analysis are not reported in the registry record. Likewise, a non-significant test does not establish that any observed difference is unimportant.

How should the endpoint rule be understood?

The registry describes measurements as biweekly from baseline up until 12 weeks and defines weight gain as the change from baseline to last observation up to 12 weeks. It also says the last observation was carried to 12 weeks. In general, carrying a last observation forward fills a later endpoint using the most recent available measurement. This makes an endpoint value available when the scheduled final observation is missing, but it relies on an assumption that the carried value is an acceptable representation of the later outcome.

That assumption may be weak if weight would have continued to change, or if the reasons for a missing later measurement are related to a participant's outcome. The record does not report missingness by arm, reasons for missing measurements, or sensitivity analyses using other assumptions. It is therefore not possible to assess how the endpoint rule affected the arm means or the comparison.

What do planning assumptions mean?

Planning calculations based on an observed effect are illustrations of how a future study might be sized under specified assumptions; they are not a measure of power actually achieved by this study. The stated assumptions for such illustrations are a two-sided significance level of 0.05, equal allocation between arms, a normal approximation, and an effect to detect defined as a stated fraction of the observed effect. Participant counts for those planning illustrations are not reported in the registry record. No sample-size justification or design calculation for the original study is reported in the registry record.

General sample-size planning depends on the outcome's variability, the effect considered meaningful, the statistical method, and the desired operating characteristics. An effect derived from a sparse observed sample may be unstable, so planning based on it should be treated as illustrative rather than as a reliable prediction. Any planning counts would describe hypothetical future designs, not observed evidence and not observed power for this terminated study.

10. Analyses Reported by the Sponsor

The registry reports the study's own comparison as a two-sided t-test with p = .495. This is the sponsor-reported analysis available in the posted record. The record does not include a numerical test statistic, degrees of freedom, confidence interval, or between-arm estimate. A recalculated unadjusted comparison is not available because the posted continuous-outcome summaries do not contain the information needed to reconstruct the comparison, including the risperidone standard deviation.

The reported arm means are 8.5 pounds for risperidone and 6.65 pounds for aripiprazole. The aripiprazole standard deviation is 1.48 pounds; the corresponding risperidone spread is not reported in the registry record. These summaries do not supply a complete set of inputs for independently checking a conventional comparison of means. The registry record does not flag a discrepancy between a sponsor analysis and a recalculation, but there is no recalculated comparison available to place alongside the reported test.

The p-value should therefore be described as the posted analysis result, not as a value independently confirmed from the arm summaries. Its interpretation remains limited by the sparse outcome counts and lack of an interval estimate. It does not establish equality, clinical importance, or the probability that a particular treatment effect is true.

11. Multiplicity, Missing Data and Analysis Populations

The registry record lists one primary outcome measure and one reported analysis, with weight gain as the primary endpoint. It lists no secondary posted outcomes. These details describe what is posted; they do not establish that no other outcomes or analyses were considered during the study. Whether there were additional planned or exploratory analyses, or whether any adjustment for multiple testing was intended, is not reported in the registry record.

Multiplicity arises when several outcomes, time points, subgroups, or statistical tests are examined. As a general matter, looking across many analyses can increase the chance of finding an apparently notable result by chance. A prespecified analysis plan and an appropriate multiplicity strategy can help readers distinguish confirmatory questions from exploratory ones. The registry does not report enough information to assess whether such issues arose in this study. The listed single analysis should not be taken as evidence that no other analyses occurred.

Missing-data information is also limited. The primary endpoint description says that the last observation was carried to 12 weeks, but the record does not describe the number or timing of missing measurements, the reasons for missingness, or how missing baseline or follow-up data were handled beyond that endpoint rule. It does not report sensitivity analyses testing alternative assumptions. In general, the consequences of missing data depend on why data are missing and how missing outcomes differ from observed ones; those features cannot be determined from the posted record.

The analysis population is not reported in the registry record. The record lists enrollment as 5, while the outcome summaries list 1.0 for risperidone and 2.0 for aripiprazole. It does not explain whether these summaries include all participants assigned to the arms, only participants with available endpoint data, or another defined group. No intention-to-treat or per-protocol population is identified. Without that information, the relationship between randomization, follow-up, and the analyzed endpoint is unclear.

12. 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 0% reporting threshold.

Most frequent non-serious adverse event terms

The serious adverse event record lists 0 serious events affecting 1 participant at risk in the risperidone group, and 0 serious events affecting 2 participants at risk in the aripiprazole group. No serious-event terms are listed. These are the posted serious adverse event figures for the registry record. The record does not provide further detail about how serious events were defined, collected, or reviewed.

The separate other-adverse-event entries list 1 affected participant among 1 at risk for risperidone and 2 affected participants among 2 at risk for aripiprazole. These entries are not the serious adverse event counts. They should not be combined with the serious-event figures or treated as evidence about event severity without additional information. The registry does not identify the event terms in the posted summary.

Counts from such limited groups provide little basis for comparing safety between arms. A zero serious-event count in these summaries does not establish that serious events cannot occur, and the figures do not characterize risks beyond the participants and events represented in the record. Death counts are not reported in the registry record. The adverse-event entries should be read as limited recorded counts, not as a comprehensive assessment of comparative safety.

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

The record shows that mean weight gain was reported as 8.5 pounds in the risperidone arm and 6.65 pounds in the aripiprazole arm over the endpoint defined as baseline to last observation up to 12 weeks. A two-sided t-test is reported with p = .495. The aripiprazole arm's standard deviation is 1.48 pounds, while a standard deviation for risperidone is not reported in the registry record.

These results do not establish that the treatments have the same effect, nor do they establish that one treatment produces a clinically important difference. The registry does not give a confidence interval or a numerical between-arm estimate, so the precision and plausible range of the comparison cannot be directly assessed. The p-value alone cannot answer how large a difference may be compatible with the data.

The small outcome summaries and the unexplained difference between listed enrollment and outcome counts further limit interpretation. The record does not identify the analysis population, explain why the study was terminated, or provide detailed missing-data information. Last observation carried to 12 weeks is the stated endpoint procedure, but the impact of that approach cannot be evaluated from the posted details.

The findings relate only to the recorded groups, outcome definition, and study conduct described in the registry. They do not establish results for other populations, treatment settings, or follow-up periods. The study's stated purpose was to learn about weight gain and related side effects when children with mood disorders receive antipsychotic medicine; the posted outcome offers a limited description of weight gain for the two listed arms. It is not a complete account of treatment effects or safety.

As a general principle, absence of a precise comparison is not evidence that a difference is absent. Stronger interpretation would require fuller outcome information, a clearly defined analysis population, and uncertainty estimates. Those elements are not reported in the registry record.

14. Why This Trial Matters Statistically

This trial matters statistically as an example of how a randomized design and a reported hypothesis test can still leave substantial uncertainty when the outcome summaries are sparse. Randomization is an important design feature, but it cannot by itself make a very limited set of observations precise. A p-value may be posted even when the accompanying information needed to understand the estimate and its uncertainty is incomplete.

The record also illustrates why arm means should be read alongside measures of spread and a comparison estimate. One arm has a reported standard deviation and the other does not. Without a numerical contrast and confidence interval, readers cannot readily judge the magnitude or precision of the comparison. This is particularly important for continuous outcomes, where both the center and variability of observed values contribute to interpretation.

The endpoint definition raises a second general issue: the rule used to fill an endpoint when later observations are unavailable can affect what a mean represents. Last observation carried to 12 weeks is stated, but the missingness pattern and sensitivity analyses are not reported. Statistical interpretation depends not only on the test named in the record, but also on data collection, endpoint construction, and the population included in analysis.

Finally, the enrollment count, outcome counts, and terminated status emphasize the value of complete registry reporting. The record does not explain the relationship between enrollment and analysis, the reason for termination, or the original sample-size rationale. These omissions prevent firm conclusions about design adequacy and generalizability. The case demonstrates why readers should distinguish a posted result from a precise or comprehensive estimate of a treatment comparison.

15. 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 specific to this study constrain interpretation. The record lists enrollment as 5, while the arm-level outcome summaries list 1.0 for risperidone and 2.0 for aripiprazole. The reason for the difference is not reported in the registry record. The analysis population is also not identified, so it is unclear how the participants represented by the outcome summaries relate to enrollment and random assignment.

The continuous-outcome information is incomplete for reconstructing or evaluating the comparison. Means are posted for both arms, but the risperidone standard deviation is not reported. The registry reports a two-sided t-test and p = .495, but provides no confidence interval or numerical estimate of the between-arm contrast. A recalculated comparison is therefore unavailable from the posted summaries.

The endpoint uses last observation carried to 12 weeks. The record does not describe missingness patterns, reasons for missing observations, or analyses under alternative missing-data assumptions. Because the study is marked terminated, the lack of information about the reason for termination and its relationship to follow-up is another limitation. Those details are not reported in the registry record.

Further limitations include the absence of a reported sample-size justification, a detailed statistical analysis plan, and a clear description of allocation procedures. The record does not state whether additional outcomes or analyses were considered, or whether any multiplicity adjustment was planned. Participant characteristics and eligibility details are also not reported in the registry record. These gaps make it difficult to assess the study's representativeness, the robustness of the analysis, and the extent to which the posted results address its intended question.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: Side Effect Study of Antipsychotic Medicines to Treat Childhood Bipolar Disorder — NCT00746252.
• Registry record last updated 2020-01-06; status last verified 2020-01.
• Results first posted 2018-03-21.
• 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.

19. Record Summary

The registry record describes a randomized, unmasked, parallel study of risperidone and aripiprazole in children with bipolar disorder. It is marked terminated and lists enrollment as 5, with start recorded as 2008-06 and primary completion as 2010-10. The primary outcome is weight gain in pounds from baseline to last observation up to 12 weeks, with last observation carried to 12 weeks.

The reported arm means are 8.5 pounds for risperidone and 6.65 pounds for aripiprazole. The aripiprazole standard deviation is 1.48 pounds; the risperidone standard deviation is not reported in the registry record. The reported two-sided t-test has p = .495. No confidence interval or numerical between-arm estimate is posted, and the outcome summaries do not support a recalculated comparison.

The registry does not report the analysis population, the reason for termination, or enough missing-data detail to assess the endpoint rule. These limitations make the posted results a sparse description rather than a precise estimate of the comparative outcome.