Pediatric Aggression and Violence: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Pediatric Aggression and Violence (NCT01253343), a registered phase not stated study. The primary outcome is Salivary Hormone Correlation With Brief Rating of Aggression by Children and Adolescent (BRACHA) Score.

2. Clinical Question

3. Trial Design

This is a completed feasibility study of salivary hormones and pediatric aggression and violence, sponsored by Children's Hospital Medical Center, Cincinnati. The registry lists an enrollment of 17 and describes two groups: inpatient high aggression and inpatient low aggression. It does not state how participants were allocated, and it explicitly cautions that this is not a randomized comparison and that the arms are not a like-for-like control comparison. The group descriptions are not reported in the registry record.

The stated objective was to examine whether salivary hormone concentrations could improve prediction of aggression during psychiatric hospitalization beyond the BRACHA score. The brief summary also describes a broader intended approach combining BRACHA, the Predatory-Affective Aggression Scale, and salivary hormones to predict the severity and type of aggression. The registry record does not explain how the scales or hormone measurements would be combined, or identify a statistical model for prediction.

For the primary measure, saliva samples were collected three times over a 24-hour period on one of the initial three hospital days. The measure description says these samples were used to determine concentrations of cortisol, dehydroepiandrosterone (DHEA), and testosterone. The brief summary names cortisol, testosterone, and dehydroepiandrosterone sulfate (DHEAS); the record uses both DHEA and DHEAS terminology. The registry record does not clarify that terminology.

The record says the intended analysis population was 24 participants, based on a sample-size calculation attributed to Kelsey et al., but that samples were collected from 17 participants because of funding constraints. It does not report the number included in an analysis. The listed study dates are 2011-09 for start and 2012-11 for primary completion; the status is completed. No analysis results are 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: Inclusion criteria included boys aged 7 to 9 years, with BRACHA scores, admitted to a psychiatric inpatient unit. Exclusion criteria included presence of infection requiring antibiotics within 2 weeks, recent surgery within 8 weeks, bleeding gums within 8 weeks of admission, current detainment in juvenile detention, and steroid use.

Overall Study

Analysis population for the primary outcome, as defined in the registry: The number of participants for analysis was supposed to be 24 based on Kelsey's sample size calculation (Kelsey et al., Methods in Observational Epidemiology, 2nd Edition, Table 12-15). However, due to funding constraints we only collected samples from 17 participants.

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 Salivary Hormone Correlation With Brief Rating of Aggression by Children and Adolescent (BRACHA) Score, measured over Collected on one or two days.

The primary outcome is described as the correlation of salivary hormone concentrations with the BRACHA score. The measure description indicates that saliva was collected to determine peripheral concentrations of cortisol, dehydroepiandrosterone (DHEA), and testosterone, then compared with participants’ BRACHA scores. The listed unit is ul/dL, the summary measure is a mean, and dispersion is specified as standard deviation. The timeframe is described as collection on one or two days, while the detailed description says sampling occurred on one of the initial three hospital days. These descriptions are not reconciled in the registry record.

BRACHA is presented in the study rationale as a way to identify children and adolescents at higher risk for aggression and violence on inpatient units. Examining whether hormone levels vary in relation to BRACHA scores could therefore address whether those measurements might add information to risk prediction. The stated aim is predictive, but the registry record does not specify how improvement in prediction would be measured, what threshold or model would be used, or how future aggression would be defined for this outcome.

The listed endpoint does not itself establish that a hormone predicts later aggression or adds useful predictive information. A correlation describes association, not necessarily prediction or causation. No hormone means, standard deviations, correlation estimates, confidence intervals, or analysis results are reported in the registry record, so the observed relationship and its precision cannot be described.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The primary outcome is framed as a relationship between continuous measurements: salivary hormone concentrations and the BRACHA score. The registry lists the outcome summary as a mean with standard deviation, but it does not report the numerical summaries or describe the analysis used to estimate the relationship. It also lists zero analyses. The methods discussion below therefore explains relevant statistical ideas generally; it does not describe a confirmed analysis performed in this study.

What does a correlation address?

In general, correlation describes how two measurements vary together. It can indicate whether higher values of one measurement tend to occur alongside higher or lower values of another, and how closely the measurements follow a linear pattern when a linear correlation is used. Correlation alone does not show that one measurement causes the other to change. Nor does correlation by itself establish that a measurement improves prediction: that requires comparing predictive performance with and without the additional information, using a specified model and an appropriate assessment approach.

The outcome title refers to a correlation with BRACHA, but the registry record does not name the correlation statistic, state whether the analysis would be adjusted, or explain how repeated saliva samples would be handled. Those details matter because the choice of method depends on the data and the question. For example, an association between each participant’s average hormone level and BRACHA is a different analysis from an analysis that retains each sample as a separate observation.

Why do repeated samples matter?

General statistical principle: repeated measurements from one participant are usually related to one another. Treating them as independent observations can make an analysis appear to contain more independent information than it does. Methods for repeated measurements can account for within-participant dependence, while a summary such as a participant-level mean can reduce the data to one value per participant. The registry record says three saliva samples were collected over a 24-hour period, but it does not report how those samples were summarized or modeled.

The collection description also places sampling on one of the initial three hospital days, while the timeframe field says collection occurred on one or two days. These statements do not fully align. The record does not report whether timing was standardized across participants or how timing was incorporated into analysis. This is relevant in general because measurements taken at different times can vary for reasons unrelated to the clinical question.

Why are means and standard deviations used?

General statistical principle: a mean gives a measure of the center of a continuous variable, while a standard deviation describes how spread out individual measurements are around that mean. Together, they can help describe a distribution, but they do not show every feature of the data. If observations are highly skewed or contain influential extreme values, the mean and standard deviation may not adequately convey the distribution. The registry record names these summary measures but does not report values, distribution checks, or any alternative summaries.

Hormone concentrations are continuous measurements, and the endpoint lists ul/dL as the unit. The record does not state whether values were transformed, whether measurements below any assay limit occurred, or what laboratory methods were used. These details are not reported in the registry record; no assumptions about them can be made.

What does a confidence interval add?

General statistical principle: a confidence interval communicates the uncertainty around an estimate under the assumptions of the analysis. A narrow interval indicates greater precision than a wide interval, while an interval that includes a value representing no association may be compatible with a range of possible relationships. Intervals do not correct for bias, poor measurement, or a mismatched analysis. No confidence interval for the primary association is reported in the registry record, so its precision cannot be assessed.

What do the planning illustrations mean?

The record describes sample-size planning illustrations based on a fraction of the observed effect, using a two-sided significance level of 0.05, equal allocation between arms, and a normal approximation. In general, a planning calculation asks what sample size would be needed to detect an assumed effect under stated assumptions. It is not an estimate of the study’s actual power after data collection. The record states that the original intended analysis population was 24 and that samples were collected from 17 because of funding constraints; the planning illustrations should not be confused with observed evidence about the primary relationship.

The record does not give the effect estimate, the detailed calculation inputs, or the numerical planning outputs in the registry record for interpretation. It also does not report a protocol-defined statistical analysis plan, model specification, handling of repeated measures, missing-data approach, or interim analysis plan. These items are not reported in the registry record, so no claim can be made that a particular method was used.

10. Confidence Intervals for Each Arm

No two-arm comparison is shown for this trial, but each arm's mean can be given with a t-based 95% confidence interval computed from the posted mean, spread and number of participants.

R code and output

arm <- c("Low Aggression Group + Cortisol Sample1", "Low Aggression Group + Cortisol Sample2", "Low Aggression Group + Cortisol Sample3", "High Aggression Group + Cortisol Sample1", "High Aggression Group + Cortisol Sample2", "High Aggression Group + Cortisol Sample3")
n   <- c(8, 6, 6, 9, 9, 9)
m   <- c(0.8, 0.4, 0.3, 0.4, 0.8, 0.2)
sp  <- c(0.9, 0.1, 0.3, 0.2, 0.3, 0.1)
sd  <- sp      # posted standard deviations
for (i in seq_along(n)) {
  se <- sd[i] / sqrt(n[i]); h <- qt(0.975, n[i] - 1) * se
  cat(sprintf("CARM|%s|%d|%.3f|%.3f|%.3f|%.3f\n", arm[i], n[i], m[i], sd[i], m[i] - h, m[i] + h))
}

CARM  Low Aggression Group + Cortisol Sample1  8  0.800  0.900  0.048  1.552
CARM  Low Aggression Group + Cortisol Sample2  6  0.400  0.100  0.295  0.505
CARM  Low Aggression Group + Cortisol Sample3  6  0.300  0.300  -0.015  0.615
CARM  High Aggression Group + Cortisol Sample1  9  0.400  0.200  0.246  0.554
CARM  High Aggression Group + Cortisol Sample2  9  0.800  0.300  0.569  1.031
CARM  High Aggression Group + Cortisol Sample3  9  0.200  0.100  0.123  0.277

11. Multiplicity, Missing Data and Analysis Populations

The registry record lists two outcome measures but does not provide the details of the other measure in the record. It reports zero analyses, so no multiplicity strategy is available for assessment. Multiplicity arises in general when several outcomes, hormones, comparisons, or analyses are examined: the more opportunities there are to identify a result, the more important it is to specify which tests are primary and how interpretation will account for multiple testing. The registry record does not report a testing hierarchy, adjustment method, or prespecified set of comparisons.

The primary question involves several hormone measures and a BRACHA score, and the brief summary also names the Predatory-Affective Aggression Scale. The record does not state whether each hormone was tested separately, whether the scales were included in one model, or whether analyses were planned for aggression severity or type. Those analysis details are not reported in the registry record. It is therefore not possible to determine what multiplicity considerations would apply to the analyses actually conducted.

The record says samples were collected from 17 participants, compared with an intended analysis population of 24. It does not report how many participants had complete measurements, whether any samples were unusable, how missing measurements were handled, or whether all participants with available data were included in each analysis. These details are not reported in the registry record.

Nor does the record define the analysis population beyond the statement about the intended number. The criteria for inclusion in each outcome analysis, any exclusions after collection, and the treatment of incomplete BRACHA or hormone data are not reported in the registry record. Without these details and analysis results, readers cannot assess the effect of missing data or determine whether the analyzed participants differed from those whose data were unavailable.

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

This study addresses a practical research question: whether measurements of salivary hormones could contribute to understanding or predicting aggression among children and adolescents receiving psychiatric inpatient care. The record describes the hormones measured, the BRACHA score used as a comparison, and the intended use of multiple information sources to improve prediction. It also records that the study was completed and that samples were collected from 17 participants.

What the record does not establish is whether hormone concentrations were associated with BRACHA scores. No estimates or analysis results are reported in the registry record. It therefore cannot show the direction or magnitude of an association, how precisely it was estimated, or whether any finding would be compatible with no association. The record also does not report a comparison of predictive performance with and without hormone measurements, so it cannot establish that adding hormones improved prediction.

Even if an association had been reported, a correlation would not by itself show that hormone levels cause aggression or that changing a hormone level would change behavior. Predictive usefulness is also distinct from association: a measure may correlate with a score without improving predictions in a meaningful or reliable way. Establishing added predictive value generally requires a defined prediction model and an evaluation of how well it performs, neither of which is described in the registry record.

The two group labels do not imply a randomized treatment comparison. Allocation is not stated, and the registry record cautions that the groups are not a like-for-like control comparison. No treatment effect should be inferred from these group labels. The record’s zero entries for specified adverse-event categories describe the recorded fields but do not answer the primary association question.

Accordingly, the available information is best read as a description of the feasibility study’s aim, sampling plan, and enrollment relative to the intended analysis population. It is not evidence for or against the use of salivary hormones in clinical prediction, and it does not support medical decisions.

13. Why This Trial Matters Statistically

Feasibility studies can clarify whether research measurements can be collected in a clinical setting and can identify practical constraints on the intended study. This record describes saliva sampling in children and adolescents on psychiatric inpatient units, where the stated motivation is to improve prediction of aggression and violence. It also makes clear that funding constraints limited collection relative to the intended analysis population: the record cites an intended population of 24 and reports samples from 17 participants.

The statistical question is more demanding than asking whether hormone levels differ between two labeled groups. The stated primary outcome concerns hormone concentrations in relation to BRACHA scores, and the broader aim concerns improving prediction. Association, group separation, and predictive improvement are distinct questions. A clear analysis plan and reported estimates would be needed to distinguish among them. The registry record does not report those analyses or results.

The study also illustrates why design details matter when interpreting clinical research. The arms are not a randomized, like-for-like comparison, and allocation is not stated. The record describes repeated saliva sampling but does not explain how those measurements were incorporated into analysis. Both the group structure and repeated measurements affect what statistical conclusions an analysis can support.

For readers assessing the evidence, the absence of reported estimates is consequential: the intended relationship cannot be judged for direction, size, or precision. The study’s feasibility context may explain why the collected sample was smaller than intended, but it does not substitute for outcome reporting. Clear reporting of analysis populations, missing-data handling, methods, and estimates would be needed to understand what the collected data show and how much uncertainty remains.

14. 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 principal limitation for statistical interpretation is that the registry record reports no analysis results. Although the record lists two outcome measures, it reports zero analyses and gives no hormone summaries, association estimates, confidence intervals, or predictive-performance results. This prevents assessment of the primary question and its precision.

The record states that the intended analysis population was 24, based on a cited sample-size calculation, but that samples were collected from 17 because of funding constraints. It does not report the number of participants included in any analysis, the completeness of measurements, or the reasons for any missing data. The effect of incomplete data on the findings cannot therefore be assessed.

Study-design information is also limited. The record identifies high- and low-aggression inpatient groups but does not describe their membership or characteristics. Allocation is not stated, and the record cautions that the arms are not a like-for-like control comparison. The method used to define aggression group status is not detailed in the arm descriptions.

Several measurement and analysis details are unclear or absent. The summary names DHEAS, while the primary measure description names DHEA. The timeframe field says samples were collected on one or two days, while the detailed description specifies a 24-hour collection on one of the initial three hospital days. The record does not clarify these differences, describe the assay procedures, or explain how repeated samples were summarized or modeled. The statistical analysis plan, multiplicity approach, missing-data method, and analysis population are not reported in the registry record.

These limitations do not establish that the measurements or study procedures were inadequate; they identify what cannot be evaluated from the registry record. In particular, without reported analyses and methods, readers cannot determine whether the intended association was estimated or whether any estimate would be robust.

15. Related Tutorials

16. Related Statistical Calculators

17. Sources

• ClinicalTrials.gov: Pediatric Aggression and Violence — NCT01253343.
• Registry record last updated 2013-08-28; status last verified 2013-05.
• Results first posted 2013-08-28.
• 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.

18. Record Summary

The registry record describes a completed feasibility study of salivary hormones and pediatric aggression and violence, sponsored by Children's Hospital Medical Center, Cincinnati. The listed enrollment is 17. The study describes inpatient high-aggression and inpatient low-aggression groups, but allocation is not stated and the record cautions that the arms are not a like-for-like control comparison.

The primary outcome concerns the relationship between salivary hormone concentrations and BRACHA scores. Saliva sampling is described as three samples over a 24-hour period on one of the initial three hospital days. The record says the intended analysis population was 24, based on a cited calculation, but samples were collected from 17 because of funding constraints.

The record lists two outcome measures and zero analyses. It does not report results, estimates, or analysis methods sufficient to assess whether hormones were associated with BRACHA scores or improved prediction. The status is completed; the listed start and primary-completion dates are 2011-09 and 2012-11.