Propranolol for Challenging Behaviors in Autism: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Propranolol for Challenging Behaviors in Autism (NCT04047355), a randomized phase 2 study. The primary outcome is Aberrant Behavior Checklist--Community (ABC-C) Irritability Subscale Scores.

2. Clinical Question

This is a within-participant treatment question in the broad design sense: each participant served as their own control by receiving both propranolol and placebo. Randomizing the order can help balance treatment-order effects across the study, while masking is intended to reduce the influence of treatment knowledge on participants and study personnel. The registry record labels masking as quadruple, but does not identify which roles were masked.

The clinical question is focused on a selected group with severe and chronic behaviors, rather than on autism spectrum disorders generally. The registry rationale notes that challenging behaviors can affect morbidity and quality of life, and that alternative treatment options are of interest. Those statements explain the study question; they are not findings from this trial. The small pilot enrollment and the posted outcome data should therefore be read as evidence from this study group, not as a broad estimate applicable to every person with autism.

3. Trial Design

The registry record describes a randomized, quadruple-masked, phase 2 crossover study. Its stated enrollment was 8, and its status is completed. Participants were assigned to one of two sequences: propranolol first followed by placebo, or placebo first followed by propranolol. A washout period separated the phases. The duration of that washout is not reported in the registry record.

In a crossover design, each participant receives more than one treatment condition, so comparisons can use information from the same person under both conditions. This can be useful when outcomes vary substantially between people. It also makes treatment order, timing, and possible carryover important considerations. The registry record describes randomized order and a washout, but the length of washout and the analysis of sequence, period, or carryover effects are not reported in the registry record.

The primary outcome used each participant’s last baseline score and last scores from the placebo and propranolol phases. Titration schedules varied according to individual response and were not predetermined. On average, the placebo phase lasted 8 wks and the propranolol phase lasted 11 wks, including titration. This means that the period of observation was not described as a fixed duration for every participant.

Enrollment and the number contributing to the posted outcome summaries are not the same: the registry lists enrollment of 8, while each treatment-phase summary reports n=6. The reasons for this difference, and the exact analysis population, are not reported in the registry record. The number of participants listed for the open-label safety group is 5; how that group relates to the randomized crossover participants is not reported in the registry record.

The registry identifies Rutgers, The State University of New Jersey as the sponsor. It does not report the propranolol dose, a formal sample-size justification, or a detailed statistical analysis plan.

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: This was a crossover study. 6 subjects underwent both experimental phases (i.e., crossover design). 3 subjects underwent Propranolol first, then Placebo while the remaining 3 subjects underwent Placebo first, then Propranolol.

Overall Study

Analysis population for the primary outcome, as defined in the registry: Participants served as their own controls. Each participant received both Propranolol and Placebo. The treatment order was randomized.

5. Baseline Characteristics

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

Population: All participants underwent baseline phase. The primary measure during baseline and all phases thereafter was the Aberrant Behavior Checklist--Community Irritability Subscale.

6. Primary Endpoint

The registry lists the primary outcome as Aberrant Behavior Checklist--Community (ABC-C) Irritability Subscale Scores, measured over Titration schedule varied based on individual response; it was not predetermined. On average, Placebo was 8 wks and Propranolol was 11 wks, including titration. The last Baseline score and the last scores from Placebo and Propranolol phases were analyzed..

The primary endpoint was the Aberrant Behavior Checklist--Community (ABC-C) Irritability Subscale score. The ABC-C is a behavior checklist used to measure drug and other treatment effects in people with intellectual and developmental disabilities. It has five subscales, covering irritability, lethargy, inappropriate speech, hyperactivity, and stereotypy, and is based on 58 items describing behavioral problems.

The Irritability Subscale was the primary dependent measure in this study. Its score ranges from 0 to 45, with lower scores representing better outcome. That direction matters when reading the observed phase medians: a lower score indicates fewer or less severe behaviors captured by this measure, not a direct measure of every aspect of behavior or quality of life.

The endpoint is relevant to the study question because the registry describes severe challenging behaviors, including aggression and self-injury, as a concern for the intended population. A structured checklist offers a repeated way to quantify behavior across treatment phases. It can support comparison, but it does not by itself establish how participants functioned in every setting or how meaningful a score change was to each participant.

The analysis used the last baseline score and the last scores from each treatment phase. The timing of those final observations was not the same fixed duration for every participant: titration schedules varied based on individual response. The registry record does not report a clinically important difference threshold for this scale.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

Statistical interpretation starts with the outcome scale and the comparison being made. The ABC-C Irritability Subscale is a score from 0 to 45, and lower scores represent better outcome. The registry reports medians and inter-quartile ranges rather than means and standard deviations. A median identifies the middle of an ordered set of observations, while the inter-quartile range describes the spread of the middle half. These summaries are often useful when scores are skewed or when a few observations could strongly affect a mean. The registry does not report why these summaries were selected.

What does the rank-based test assess?

The reported test is Wilcoxon (Mann-Whitney), a rank-based method for comparing two groups. In general, rank-based methods order observations and compare their relative positions rather than relying directly on a normal model for the raw scores. The registry labels the analysis as a comparison of final values and reports a median difference of 3.00 and p=0.12. The method label does not, by itself, establish that the comparison estimates a difference between paired within-person changes.

That distinction is important because the trial is a crossover: each participant received both treatments. In general, analyses of crossover data can use the within-person pairing, which distinguishes each participant’s response under one condition from that same participant’s response under the other condition. A test treating phase observations as independent may not use that pairing. The registry record does not explain whether the analysis accounted for paired observations, treatment sequence, or period. It therefore does not provide enough detail to determine precisely how the reported test represents the crossover design.

How should the p-value be read?

In general, a p-value measures how unusual data at least as incompatible with a null hypothesis would be under that null hypothesis and the assumptions of the test. It does not give the probability that the null hypothesis is correct. Nor does it tell whether a difference is clinically important. The reported p=0.12 is one component of the result; it should be considered with the observed medians, the spread of scores, and the limited number of observations. A confidence interval for the treatment contrast is not reported in the registry record.

In general, confidence intervals give a range of values compatible with the data and the interval method, and show the precision of an estimate. They are useful because they communicate both the estimated contrast and uncertainty around it. The inter-quartile ranges reported for each phase describe score dispersion, not uncertainty around the treatment difference. They cannot be read as a confidence interval for that difference.

What does the effect size add?

The registry reports effect size r=-0.64. In general, an effect-size statistic aims to describe the magnitude of an observed contrast in a way that is not limited to the p-value. Interpretation depends on the statistic’s definition, its sign convention, the comparison order, and the analysis method. Those details are not fully stated in the registry record, so the sign should not be used independently to infer which phase had higher scores. The medians and the stated direction of the scale provide the direct descriptive comparison.

What does the post-hoc power value mean?

The registry reports a post-hoc power analysis of 0.39. In general, post-hoc power is calculated using the observed estimate and assumptions about the analysis. It does not provide an independent measure of how convincing the evidence is; it is closely tied to the observed result and the sample size. It is different from a prospective design calculation, and it should not be described as observed power. A prospective sample-size justification is not reported in the registry record.

Why can crossover details matter?

In general, crossover studies require attention to the possibility that treatment effects could persist into a later period, as well as differences between periods and sequences. A washout period is described in the registry, but its duration is not reported. The record also does not report whether the analysis tested or adjusted for carryover, period, or sequence. With phase durations that varied according to individual response, the timing and definition of the final observation are also relevant to interpretation.

The registry states that the primary analysis used the last baseline score and the last scores from placebo and propranolol phases. The phase durations averaged 8 wks for placebo and 11 wks for propranolol, including titration, but were not predetermined for individual participants. In general, such variation can mean that final measurements represent different exposure periods across participants. The registry record does not explain how these differing schedules were handled statistically.

Finally, the posted result is a summary of a small pilot dataset. In general, small samples tend to yield imprecise estimates, and individual observations can have a substantial influence on medians, ranks, and treatment contrasts. The reported phase counts are n=6 for each condition, whereas enrollment is 8. The reasons for the difference and the number of paired records contributing to the primary comparison are not reported in the registry record. Without those details and an interval estimate for the contrast, the size and precision of the treatment comparison remain difficult to assess.

10. Analyses Reported by the Sponsor

The registry reports the sponsor’s primary comparison as a Wilcoxon (Mann-Whitney) analysis of final values. It gives a median difference of 3.00 and p=0.12, and separately lists an effect size r=-0.64 and a post-hoc power analysis of 0.39. These are the posted analysis results; the registry record does not give a confidence interval for the median difference.

A separate recalculated estimate is not reported in the registry record, so there is no numerical recalculation to compare with the sponsor’s result. The record does not provide the participant-level paired data or enough analysis detail to reproduce the comparison. In particular, it does not explain whether the crossover pairing, treatment period, or sequence was included in the reported test. This matters because a crossover design creates related observations within participants, while the named Mann-Whitney method is generally used for independent groups.

The reported medians and inter-quartile ranges are descriptive summaries for each phase. They are not a substitute for a paired treatment contrast, and differences between the displayed summaries should not be treated as a newly calculated estimate. The posted sponsor analysis is therefore the available inferential result, with its method and limitations interpreted as stated in the registry record.

11. Multiplicity, Missing Data and Analysis Populations

The registry lists two outcome measures and six analyses. The primary endpoint is the ABC-C Irritability Subscale; the listed secondary outcome is the Clinical Global Impression Improvement (CGI-I) Scale. The record does not explain how the six analyses map to the outcomes, whether they include additional comparisons or analysis choices, or whether any adjustment for multiple testing was made.

In general, testing multiple outcomes or contrasts can increase the chance of obtaining a small p-value by chance if each is interpreted separately. Multiplicity procedures can address that issue, but the appropriate approach depends on the prespecified analysis plan and the family of hypotheses. Since an adjustment strategy and a full statistical plan are not reported in the registry record, the reported p-value should be understood as the value listed for its stated comparison, without assuming a multiplicity correction.

Missing-data handling is also not described. The registry reports enrollment of 8 and n=6 in each treatment-phase outcome summary. It does not report why fewer observations appear in the summaries, whether all participants completed both phases, how incomplete phase data were handled, or whether any imputation was used. It also does not report the analysis population definition, such as whether the primary comparison included all randomized participants or only participants with available phase scores.

Those omissions matter particularly in a crossover study, where a complete within-participant comparison depends on observations from both phases. If some participants contribute only one phase, the analysis may rely on a different set of information than a paired comparison. The registry record does not state how many participants contributed paired values to the primary analysis, so the extent of any incomplete pairing cannot be determined from the posted summaries.

The registry describes randomized treatment order, but does not report a detailed plan for sequence, period, or carryover effects. It also does not report interim analyses. These features should not be assumed either present or absent beyond what the registry record explicitly states.

12. 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 figures report 0 affected participants in the propranolol group among 6 at risk, 0 among 6 at risk in the placebo group, and 0 among 5 at risk in the open-label group. The registry also reports 0 deaths in each of these groups. No serious adverse event terms are listed.

These figures describe the counts recorded for the groups and the numbers at risk in those groups. They do not establish that no safety concerns occurred: the registry lists other, non-serious adverse events for 2 participants in the propranolol group and 0 in the placebo and open-label groups. The specific non-serious event terms are not included in the registry record reported for this analysis.

Safety counts in a small pilot study have limited ability to characterize uncommon events or provide precise comparisons between treatments. The record does not report a statistical comparison of adverse events, nor does it provide enough detail to interpret how the open-label group relates to the randomized treatment phases. The study description notes cardiac testing to monitor safety, but the outcomes of those tests are not reported in the registry record.

Thus, the posted serious adverse event counts are zero in each listed group, but they should be read as event counts observed in the reported groups, not as proof that serious events cannot occur.

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

The primary outcome is a behavioral score for which lower values represent better outcome. The posted medians are lower during the propranolol phase than during the placebo phase: 14.0 versus 18.5. The registry’s reported comparison gives a median difference of 3.00 and p=0.12. These registry entries describe the posted results; they do not establish that the observed difference is precise or broadly applicable.

The endpoint captures the irritability domain of a structured checklist, not every behavior, aspect of functioning, or quality-of-life outcome. Although the registry rationale refers to severe challenging behavior, the primary measure should not be interpreted as a complete assessment of aggression, self-injury, or overall well-being. A clinically important difference threshold is not reported in the registry record.

The result does not establish the effect in a larger population. The study enrolled 8, and the posted summaries report n=6 per treatment phase. The record does not report a prospective sample-size justification or a confidence interval for the treatment contrast. The uncertainty around the estimated contrast therefore cannot be described with a posted interval estimate.

Nor does the reported p-value establish that there is no difference. A p-value is not a direct probability that a treatment effect exists or does not exist. The score summaries and p-value should be interpreted together with the study’s small size, the analysis method, and the incomplete reporting of crossover-specific analysis details.

The results also do not answer every safety question. No serious adverse events are reported in the listed groups, while other adverse events are recorded for 2 participants in the propranolol group. The small groups and absent detail on specific event types limit the conclusions that can be drawn from these counts.

Overall, this is a pilot result that describes outcomes from the observed study phases. It identifies a lower observed median score during propranolol than during placebo, while leaving substantial uncertainty about the size, precision, and generalizability of the treatment contrast.

14. Why This Trial Matters Statistically

This trial is statistically instructive because it combines a randomized crossover design with a small pilot sample and a behavioral endpoint. In a crossover, participants receive both conditions, potentially allowing each person to serve as their own control. That design makes the handling of paired observations and treatment order important to the interpretation of the comparison.

The registry describes the within-participant structure, but names a Wilcoxon (Mann-Whitney) test for the final-value comparison. The record does not explain whether the analysis incorporated pairing, phase, or sequence. This gap highlights a general principle: readers need to assess whether the statistical method matches the design, not only whether the test name is familiar.

The endpoint also illustrates why scale direction and summary choice matter. Lower ABC-C Irritability Subscale scores represent better outcome, and the registry reports medians and inter-quartile ranges. Those summaries describe the observed score distributions, while the reported p-value addresses a model-based comparison. A confidence interval for the contrast would help communicate precision, but none is reported in the registry record.

Finally, the difference between enrollment of 8 and phase summaries of n=6 shows why participant flow and analysis populations are essential to interpreting small trials. Missing-data handling, paired sample counts, and multiplicity procedures are not reported. These omissions do not change the posted figures, but they constrain how fully the statistical evidence can be evaluated.

The study therefore offers a useful example of how design, analysis, endpoint definition, and reporting completeness jointly determine what can be concluded from a clinical trial result.

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.

The small study size is a central limitation. Enrollment was 8, and each treatment-phase outcome summary reports n=6. The registry record does not explain the difference, identify the number of complete paired observations, or describe the analysis population. Estimates based on such limited data can be sensitive to the observations included.

The primary comparison is reported with a Wilcoxon (Mann-Whitney) method, but the record does not explain how this analysis handled the crossover structure. It does not report whether treatment pairing, period, sequence, or carryover was considered. A washout period is described, but its duration is not reported. These details are important when interpreting phase comparisons in a crossover trial.

The outcome timing was not fixed uniformly: titration varied with individual response, and the average phase durations were 8 wks for placebo and 11 wks for propranolol, including titration. The registry record does not report how variation in duration was handled in the analysis or whether the final measurement times were comparable across participants.

Precision is difficult to assess because the registry does not report a confidence interval for the treatment contrast. The phase inter-quartile ranges describe score variation, but do not quantify uncertainty in the contrast. The reported p-value and post-hoc power do not replace an interval estimate. A prospective sample-size justification is not reported.

Finally, multiplicity adjustment, missing-data procedures, detailed analysis populations, and interim analyses are not reported in the registry record. The record lists a secondary outcome and six analyses but does not explain the testing strategy. These reporting gaps limit assessment of robustness and the extent to which the primary comparison was prespecified and protected from selective interpretation.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: Propranolol for Challenging Behaviors in Autism — NCT04047355.
• Registry record last updated 2025-06-04; status last verified 2025-05.
• Results first posted 2025-06-04.
• 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

NCT04047355 is a completed phase 2 randomized, quadruple-masked crossover study sponsored by Rutgers, The State University of New Jersey. It enrolled 8 participants and examined propranolol and placebo phases in adolescents and adults with autism spectrum disorders and severe challenging behaviors. Participants received both treatments in randomized order, with a washout period described.

The primary outcome was the ABC-C Irritability Subscale, for which lower scores represent better outcome. The reported medians were 14.0 for propranolol and 18.5 for placebo, with n=6 in each phase summary. The registry reports a median difference of 3.00 and p=0.12 from a Wilcoxon (Mann-Whitney) analysis.

Zero serious adverse events and zero deaths are reported in the propranolol, placebo, and open-label groups. Interpretation is limited by the small sample, absence of a confidence interval for the primary contrast, and unreported details on paired analysis, missing data, and multiplicity.