Biomarkers in Autism of Aripiprazole and Risperidone Treatment (BAART): Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Biomarkers in Autism of Aripiprazole and Risperidone Treatment (BAART) (NCT01333072), a randomized phase 4 study. The primary outcome is Changes in the Irritability Subscale of the Larger ABC (Abberent Behavior Checklist) That Occur From Baseline to 10 Weeks.

2. Clinical Question

The registry frames the study as an evaluation of biomarkers as predictors of efficacy and safety in children receiving either treatment. The posted primary outcome, however, is the group-level score change on the ABC Irritability subscale; the record lists no posted secondary outcome results. The results therefore address the comparison of the recorded primary measure, rather than every possible aspect of treatment response or biomarker performance.

The reported mean difference is expressed as aripiprazole minus risperidone. Its positive value indicates that the posted mean for aripiprazole was higher than the posted mean for risperidone on the reported measure. Because lower ABC Irritability scores indicate improvement, the direction of a change score needs to be interpreted with the scale’s scoring convention in mind. The registry does not explain further how the direction of the reported change values is encoded.

3. Trial Design

BAART is listed as a phase 4, randomized, quadruple-masked, parallel-group trial. The registry names the Medical University of South Carolina as sponsor and reports enrollment of 80. The trial status is completed. Its recorded start is 2011-07, and its primary completion is 2015-08.

Participants were assigned to one of two drug groups: risperidone or aripiprazole. Both are described in the record as atypical antipsychotics. The masking designation is quadruple, but the registry record does not identify which four roles were masked or describe the practical masking procedures. Allocation is recorded as randomized; details of sequence generation, concealment, or stratification are not reported in the registry record.

The primary outcome was assessed over baseline to 10 weeks. The registry describes a multi-center trial across three academic centers in South Carolina. It reports a mean and standard deviation for each treatment group, with 24 participants listed for the risperidone outcome and 27 for the aripiprazole outcome. These outcome group sizes are not the same as the enrollment figure; the record does not explain the difference or state how missing outcome observations were handled.

The registry record does not report the original sample-size justification, the trial’s planned statistical model, or whether interim analyses were conducted. It also does not provide enough detail to determine whether the posted group summaries are based on an intention-to-treat population, a per-protocol population, or another analysis population. Those omissions limit reconstruction of the planned analysis from the registry alone.

4. Participant Flow and Accounting

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

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: Score on Aberrant Behavior Checklist (ABC) scale for Irritability

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 Changes in the Irritability Subscale of the Larger ABC (Abberent Behavior Checklist) That Occur From Baseline to 10 Weeks, measured over baseline to 10 weeks.

The primary endpoint is change in the Irritability subscale of the Aberrant Behavior Checklist (ABC-I) from baseline to 10 weeks. The ABC contains 58 items, and its Irritability subscale contains 15 items. A parent or caregiver completes the items under an investigator’s supervision. Each item is scored from 0, meaning no problem, to 3, meaning severe problem; the total ABC-I score ranges from 0 to 45. The registry states that a fall in scores indicates behavioral improvement.

Irritability is the named target of the primary measure in this trial. A continuous score retains gradations in reported severity rather than reducing outcomes to a simple event/no-event classification. The posted summaries therefore present group means and standard deviations, and the comparison is a difference between group means.

The outcome title specifies change from baseline to 10 weeks, but the registry does not explain the precise calculation or sign convention used for the change values. That detail matters when translating a positive or negative mean difference into a direction of score change. The registry also does not report the scale’s measurement properties or a threshold for interpreting a difference as clinically important. Statistical differences and clinical importance are not interchangeable, so the reported contrast should be read as a score-scale comparison.

7. Results Posted in the Registry

• Aripiprazole: mean 14.10 versus 12.70 on control; difference 1.40 (95% CI -0.38 to 3.18), standardized effect d = 0.43; z-test p = 0.124.

8. Statistical Interpretation

9. Statistical Methods Explained

The record describes a continuous endpoint summarized by a mean and standard deviation in each group. A mean is the arithmetic average of observed scores or changes, while a standard deviation describes how dispersed individual observations are around that average. The standard deviation is not the uncertainty in the mean itself: it describes variation among observations. General statistical explanation: when a continuous outcome is compared across two independent groups, a difference in means expresses the contrast on the original outcome scale, which is units on a scale for this endpoint.

The reported contrast is aripiprazole minus risperidone. Its estimate is 1.4. A positive estimate means the first named group’s posted mean is numerically higher than the second named group’s posted mean. This arithmetic direction does not by itself establish whether the outcome is better or worse, especially because this endpoint is a change score and the record does not spell out the sign convention. The scale description says a fall in ABC-I scores indicates improvement, but does not separately clarify how the change values are signed.

The confidence interval gives a range of effect values compatible with the data under the method’s assumptions. General statistical explanation: a confidence interval describes the uncertainty of an estimate; it is not a range containing a specified proportion of individual participants’ scores. The posted interval crosses zero, so the values compatible with it include no mean difference as well as differences in either direction. The registry does not identify the method used to calculate this interval or state whether any adjustment was applied, so those details are not inferred.

The reported unadjusted p-value is 0.124. General statistical explanation: a p-value measures how unusual data at least as inconsistent with a specified null hypothesis would be if that hypothesis and the model assumptions were true. It is not the probability that the null hypothesis is true, and it does not measure the probability that a result occurred “by chance.” The unadjusted label indicates that the posted value is not described as adjusted for multiple comparisons. Since no analysis details or multiplicity plan are reported, the precise testing framework cannot be reconstructed from the record.

The posted summaries include group standard deviations of 3.0 for risperidone and 3.5 for aripiprazole. The registry also gives a pooled standard deviation of 3.275 and a standardized mean difference of 0.427 in the comparison details. General statistical explanation: a standardized mean difference expresses a mean contrast relative to a measure of spread, so it has no original scale units. Its interpretation depends on the context and should not replace the original-scale estimate or its confidence interval. The registry does not state that the standardized measure was the primary inferential target.

The registry does not report whether the primary comparison used a t-test, a regression model, repeated-measures analysis, baseline adjustment, or another approach. Randomization is recorded, but randomization alone does not specify the analysis model. Nor does the record explain the treatment of missing values, departures from assigned treatment, or clustering by center. These are material analysis choices and should be regarded as unreported in the registry record rather than filled in by assumption.

General statistical explanation: unequal variances, skewed outcome distributions, and outliers can affect the performance of methods that rely on distributional assumptions. Over-dispersion refers broadly to variability greater than a model expects; when relevant, it can lead to uncertainty estimates that are too small if ignored. The posted group standard deviations offer a descriptive view of spread, but the registry does not report diagnostic checks or a model that would establish whether such concerns arose in this trial.

The posted comparison labels its p-value unadjusted and gives one primary outcome. The registry lists no secondary outcome analyses and no analysis count. General statistical explanation: when many outcomes, time points, or subgroup comparisons are tested, the chance of obtaining at least one small p-value can increase. A prespecified multiplicity strategy can address that issue. Here, the record does not report such a strategy or enough analysis detail to determine whether other testing occurred outside the posted results.

10. Confidence Intervals and Precision

The reported mean difference is 1.4 units, with a 95% confidence interval that extends from below zero to 3.184. Because zero lies within that interval, the posted data are compatible with no mean difference as well as with nonzero differences in either direction. The interval communicates uncertainty more directly than the point estimate alone; it does not show that every value in its range is equally plausible.

The interval is not narrow enough to settle the direction of the group contrast. At the same time, it does not establish that all potentially relevant differences are excluded. The registry record does not give a threshold for clinical importance, so statistical precision cannot be translated into a judgment about whether a difference of a particular size would matter clinically.

The unadjusted p-value is 0.124. It is consistent with the interval crossing zero, but it should not be used as a measure of effect size or as proof of equivalence. The reported means and standard deviations describe the observed outcome summaries; the confidence interval describes uncertainty in the estimated difference. These quantities answer related but distinct questions.

The registry does not identify the confidence-interval calculation method, the handling of missing outcomes, or the analysis population. Those missing details make it difficult to assess whether the interval reflects adjustment for baseline scores or other design features. Its interpretation should therefore remain tied to the posted comparison and its stated unadjusted result.

11. 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 expressed as participants per arm, not total events. They are based on the observed effect and assume a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. The effect to detect is set at the stated fraction of the observed effect. These are planning assumptions for the illustration; the trial’s original design assumptions are not reported in the registry record.

The 80% power column gives the number of participants per arm in an illustrative future design intended to have an 80% probability of detecting the specified effect under the stated assumptions, if that effect and the assumptions hold. The 90% power column describes the corresponding planning target for a 90% probability. These are not guarantees about the result of a future study, and power depends on the assumptions used.

For planning at 100% of the observed effect, the figures are 86 participants per arm for 80% power and 115 participants per arm for 90% power. At 75% of the observed effect, they are 153 and 205 participants per arm, respectively. At 50% of the observed effect, they are 344 and 460 participants per arm, respectively. Smaller effects require more participants to detect under the same planning assumptions, which is why the figures increase as the target fraction decreases.

The curve should be read as a visual representation of this relationship between the effect targeted and the number of participants required under the same assumptions. These sample-size values are planning illustrations, not observed power for the completed trial.

12. 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 = 1.400; pooled standard deviation σ = 3.275.
2. n = 2 × (1.96 + 0.842)² × 3.275² / 1.400² = 85.9, rounded up to 86 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 <- 1.4                      # observed difference in means
sigma <- 3.27482668181361                     # 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(1333072)
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 = 86 (80% power), 115 (90% power)
 75% of observed effect: n per arm = 153 (80% power), 205 (90% power)
 50% of observed effect: n per arm = 344 (80% power), 460 (90% power)
power.t.test: 86.9 
Simulated power at n = 86 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.

13. Multiplicity, Missing Data and Analysis Populations

The registry lists one primary outcome measure and no posted secondary analyses. The primary comparison’s p-value is described as unadjusted. The registry record does not report a multiplicity plan, such as how testing across multiple outcomes, time points, or subgroups would be controlled. It also does not report whether other analyses were conducted but not posted. With no secondary results listed, the extent of any broader testing cannot be determined from the record.

General statistical explanation: when several hypotheses are tested, the probability of at least one apparently unusual result can rise. Adjustment procedures can address this, but they may also affect precision and interpretation. Since the record does not describe a testing hierarchy or adjustment method, the posted p-value should be understood as unadjusted rather than treated as evidence that multiplicity was formally handled.

Missing-data procedures are also not reported in the registry record. The primary outcome lists 24 participants in the risperidone group and 27 in the aripiprazole group, while enrollment is recorded as 80. The record does not explain the difference, describe outcome completion, or say whether missing observations were imputed, excluded, or handled through a model. No assumptions about why observations may be absent can be made from these numbers alone.

The analysis population is not identified. The registry does not say whether the posted outcome comparison includes everyone randomized, only participants with observed outcome data, or a population defined by treatment received or adherence. It also does not describe handling of protocol deviations. These details affect the interpretation of a randomized comparison and cannot be reconstructed from the posted summary statistics.

The adverse-event at-risk counts are reported separately from the primary outcome group counts. The registry does not state whether they represent the same analysis population or provide a reconciliation. They should therefore be interpreted as the denominators reported for the adverse-event summary, not substituted for the primary endpoint denominators.

14. 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 0% reporting threshold.

Most frequent non-serious adverse event terms

The serious adverse-event summary reports zero affected participants among 30 at risk in the risperidone group and zero among 31 at risk in the aripiprazole group. Deaths are reported as zero in both groups. These are the serious-event counts recorded for the groups; the registry lists no serious-event terms among its top serious terms.

For other adverse events, the record reports 23 affected participants among 30 at risk in the risperidone group and 19 among 31 at risk in the aripiprazole group. These are participant counts, not event counts, and the record does not provide event descriptions in the posted safety summary. It does not state whether a participant could contribute to more than one event category.

The at-risk denominators differ from the group sizes listed for the primary endpoint. The registry does not explain that difference or provide a safety observation window, severity definitions beyond the serious-event classification, or details about collection and coding. These summaries describe what the record posts, but they do not establish the absence of unreported events or provide a complete comparison of safety profiles.

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

The posted results describe a randomized comparison of mean ABC Irritability change from baseline to 10 weeks. The reported difference is 1.4 units, and its 95% confidence interval crosses zero. The unadjusted p-value is 0.124. Together, these figures indicate uncertainty about the direction and size of the difference in this recorded outcome; they do not establish equivalence or prove that the groups have the same outcome.

The scale description gives useful context: the ABC-I has 15 items, each scored from 0 to 3, with a total score range from 0 to 45, and lower scores indicate improvement. However, the registry does not specify the sign convention for the reported change values. The numerical contrast can therefore be reported as posted, but a definitive interpretation of which group had a greater reduction cannot be made from the registry description alone.

The result concerns the primary endpoint as recorded. It does not answer whether biomarkers predict response, because no biomarker results or analyses are posted in the registry record. It also does not resolve all possible efficacy or safety questions: no secondary outcome results are listed, and the adverse-event summary lacks detailed event categories and observation-period information.

The registry reports that the trial is completed, but the posted outcome comparison has fewer observations in each group than the enrollment figure. Since missing-data handling and analysis populations are not described, the relationship between enrollment, randomization, and the analyzed outcomes remains unclear. The findings should be interpreted as the posted summaries rather than as a complete account of every participant or every analysis.

Statistical uncertainty is not a treatment recommendation. The registry comparison provides an estimate and an interval for a specific scale and time frame; it does not establish a universal effect, a clinically important threshold, or the result that would occur in another population or setting.

16. Why This Trial Matters Statistically

BAART illustrates why a clinical-trial result is more than a p-value. The mean difference of 1.4 is on the original outcome scale, while the confidence interval shows substantial uncertainty about its direction relative to zero. The unadjusted p-value of 0.124 is one part of the evidence, not a verdict about whether the treatments are equivalent or whether a difference is clinically meaningful.

The endpoint’s construction also matters. ABC-I is a caregiver-reported measure of irritability with a defined score range, and the registry states that lower scores indicate improvement. Because the reported endpoint is a change score but the sign convention is not explained, interpretation requires distinguishing the numerical contrast from an assumption about improvement. Clear outcome definitions are essential for translating a statistical estimate into a clinical description.

The record also demonstrates the importance of reporting analysis details. It does not state the analysis population, missing-data method, or statistical model. Without these, readers cannot fully evaluate how the group summaries relate to the randomized participants or how uncertainty was calculated. The outcome group sizes and safety denominators are reported separately, without an explanation of their relationship.

Finally, the planning illustration emphasizes how the size of a target effect affects the number of participants needed under specified assumptions. Smaller effects require larger per-arm samples in the illustration. Such calculations are useful for planning, but they are not evidence about the power or success of the completed trial. Together, these points make transparent reporting of estimates, precision, definitions, and analytic choices central to interpreting a clinical comparison.

17. 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 leaves several trial-specific questions unanswered. Although the trial is randomized and quadruple-masked, it does not describe the randomization sequence, allocation concealment, or which roles were masked. It identifies three academic centers in South Carolina but provides no center-level outcome summaries or analysis details addressing center effects.

The primary outcome is titled as change from baseline to 10 weeks, but the exact calculation and direction convention are not reported. This makes it difficult to translate the positive mean difference directly into a statement about relative improvement. The registry also provides no threshold for a clinically important ABC-I difference, so the clinical relevance of the estimated contrast cannot be determined from the record alone.

Outcome group sizes are reported as 24 for risperidone and 27 for aripiprazole, whereas registry enrollment is 80. The record does not explain the discrepancy, report flow through follow-up, or describe missing-data handling. It does not identify the analysis population or state whether the comparison was adjusted for baseline scores. These omissions limit assessment of the result’s susceptibility to attrition or analysis choices.

The posted comparison gives a confidence interval and an unadjusted p-value but does not name the statistical method. The registry does not report a multiplicity approach, protocol deviations, or interim analyses. No secondary analyses are posted. Safety data lack event descriptions in the summary, and the relationship between safety denominators and primary outcome groups is not explained. Each of these limitations constrains what can be concluded from the registry record without implying that any particular procedure or event occurred.

18. Related Tutorials

19. Related Statistical Calculators

20. Sources

• ClinicalTrials.gov: Biomarkers in Autism of Aripiprazole and Risperidone Treatment (BAART) — NCT01333072.
• Registry record last updated 2018-09-26; status last verified 2018-09.
• Results first posted 2018-09-26.
• 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.

21. Record Summary

BAART (NCT01333072) is recorded as a completed phase 4 randomized, quadruple-masked, parallel-group trial sponsored by the Medical University of South Carolina. The registry reports enrollment of 80, a start date of 2011-07, and primary completion in 2015-08. The population is children with autistic disorder, and the interventions are risperidone and aripiprazole.

The primary endpoint is change in ABC Irritability score from baseline to 10 weeks. The posted group means are 12.7 for risperidone and 14.1 for aripiprazole; the reported mean difference is 1.4, with an interval crossing zero and an unadjusted p-value of 0.124. The record lists no secondary outcome results. Its safety summary reports zero serious adverse events in both groups, alongside other adverse-event counts, but does not provide event detail or explain the relationship between safety and outcome denominators.

Important analysis details, including missing-data handling, analysis population, and the original sample-size justification, are not reported in the registry record.