Trial to Investigate the Safety and Efficacy of Cannabidiol Oral Solution (GWP42003-P; CBD-OS) in Children and Adolescents With Autism Spectrum Disorder: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Trial to Investigate the Safety and Efficacy of Cannabidiol Oral Solution (GWP42003-P; CBD-OS) in Children and Adolescents With Autism Spectrum Disorder (NCT04745026), a randomized phase 2 study. The primary outcome is Change From Baseline in Aberrant Behavior Checklist (ABC) Subscale Total Scores.

2. Clinical Question

The study was designed to evaluate whether GWP42003-P, compared with placebo, changes symptom severity as assessed by caregivers using the Aberrant Behavior Checklist (ABC). Its primary outcome concerns change from baseline in ABC subscale total scores, rather than a single overall score. The five subscales cover irritability, social withdrawal, stereotypic behavior, hyperactivity/noncompliance, and inappropriate speech. The registry describes assessment at Week 4, Week 8, and Week 12.

For interpretation, the central comparison would be the difference in change between randomized groups for each relevant subscale and assessment time. A change within one group alone would not establish how the groups compare; the placebo group provides the concurrent reference. The registry says lower scores indicate a better clinical outcome. However, it does not report group-level changes, between-group estimates, confidence intervals, or primary-outcome significance tests in the record available for this account. Accordingly, the clinical question can be described, but its answer cannot be estimated from the reported primary outcome information.

3. Trial Design

This was a Phase 2, randomized, double-blind, placebo-controlled, parallel-group trial. The registry identifies quadruple masking. These design features indicate that participants were allocated to an active treatment group or a placebo group, with masking intended to limit the influence of treatment knowledge on participants and other parties involved in the study. The registry record does not specify the identities of all parties covered by quadruple masking.

The registry lists an enrollment of 103. This is the recorded enrollment, not an outcome-specific analysis count. For the safety data, the numbers at risk are listed separately by group: 49 for GWP42003-P and 54 for placebo. Those safety denominators should not automatically be treated as the analysis populations for the primary outcome. The primary outcome description says scores were assessed in participants with available data in the Full Analysis Set, but the record does not provide a full analysis-set count for each outcome and time point.

Participants were stratified by age group, antipsychotic use, and region before randomization. The age groups are 6 to 11 years old and 12 to 17 years old; antipsychotic use is described as on versus off; and region is North America versus Rest of the World. Stratification is intended to balance these characteristics across randomized groups, but the record does not report the resulting distributions or describe the allocation sequence.

The active-arm label identifies GWP42003-P at 10 mg/kg/day, and its description begins with 5 mg/kg/day for 1 week followed by 10 mg/kg/day. The remainder of that dosing description is not available in the registry record. The placebo description specifies matching placebo for 12 weeks. The trial status is completed, with a start date of 2021-05-31 and a primary completion date of 2023-12-21. Further details about treatment adherence, discontinuation, or follow-up after treatment are not 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: Eligible participants were randomized 7 to 14 days after the screening visit (Visit 1), once all required assessments were completed and laboratory results were reviewed. If required, screening assessments were permitted to be split over 2 visits; however, both visits were conducted within 7- to 14-day window prior to randomization (Visit 2).

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: Aberrant behavior checklist subscale total scores were assessed in participants with available data in the Full Analysis Set.

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: Demographic and baseline characteristics were assessed in the Safety Analysis Set.

6. Primary Endpoint

The registry lists the primary outcome as Change From Baseline in Aberrant Behavior Checklist (ABC) Subscale Total Scores, measured over Baseline up to Week 12.

The primary endpoint is change from baseline in Aberrant Behavior Checklist subscale total scores from baseline up to Week 12. The ABC is a caregiver-assessed checklist intended to assess the presence and severity of problem behaviors commonly observed in people with intellectual and developmental disability. Its five subscales cover irritability, social withdrawal, stereotypic behavior, hyperactivity/noncompliance, and inappropriate speech.

Individual items are rated from 0, meaning never a problem, to 3, meaning severe. Each subscale has its own possible score range: 0-45 for irritability, 0-48 for social withdrawal and hyperactivity/noncompliance, 0-21 for stereotypic behavior, and 0-12 for inappropriate speech. The registry states that higher scores indicate a worse clinical outcome and that lower change scores indicate a better outcome.

Using subscales allows behavior domains to be considered separately rather than combining them into one measure that could conceal different patterns across domains. Repeated assessments at Week 4, Week 8, and Week 12 can describe how scores change over the treatment period. These features make the measure relevant to the trial’s stated question about symptom severity. They also mean that interpretation depends on knowing which subscale and time point is being considered. The registry record does not report the numerical group results for these assessments, so the magnitude and precision of any treatment-placebo comparison cannot be described.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

Statistical methods help translate trial observations into estimates of group differences and the uncertainty around those estimates. The registry identifies this as a continuous outcome summarized using means and standard deviations. Those summaries describe the distribution of scores within a group, but they do not by themselves give the treatment-placebo contrast or its precision. The registry record does not report the primary outcome means, standard deviations, or a specific model used to compare groups.

What does a mean change describe?

A mean change is the average difference between participants’ scores at follow-up and their baseline scores within an analysis group. For this endpoint, the measure is a change in ABC subscale scores. The direction matters: the registry states that lower scores indicate better clinical outcome, so a lower change is described as more favorable on the scale. A group’s mean change alone is not the randomized treatment effect. The relevant comparison is the difference between the groups’ changes, interpreted with its uncertainty and the analysis plan in mind.

Standard deviation describes how spread out individual observations are around a group mean. It is not the same as a confidence interval, which describes uncertainty around an estimated quantity such as a group mean or a between-group contrast. A confidence interval can help readers assess both the range of values compatible with the data under the model and the precision of the estimate. The registry record does not report confidence intervals for the primary outcome.

Why does randomization matter?

Random allocation is intended to make treatment groups comparable, on average, with respect to measured and unmeasured characteristics at the start of a trial. This supports a comparison between groups, but it does not guarantee identical groups in a particular trial. Stratification by age, antipsychotic use, and region is reported in this study. The registry record does not give the baseline distributions for those characteristics or describe whether the primary analysis adjusted for stratification factors.

In a parallel-group design, participants remain assigned to their randomized group rather than crossing between treatments as part of the design. Placebo provides a concurrent comparison under the trial’s masking arrangements. Double-blind and quadruple-masked descriptions indicate efforts to limit knowledge of assigned treatment, although the registry record does not explain exactly which people were masked or how masking was maintained.

How should repeated measurements be understood?

ABC subscale scores were scheduled from baseline through Week 12, with changes described at Week 4, Week 8, and Week 12. Repeated measurements can provide information about the course of scores over time. They also create related observations: a participant’s later scores are usually not statistically independent of that participant’s earlier scores. General methods for repeated measurements account for this within-participant relationship, but the registry record does not state which method was used.

Multiple subscales and time points also create multiple possible comparisons. A prespecified primary comparison or a multiplicity strategy can help define how to interpret results across those comparisons. The record does not report a hierarchy among the subscales or time points, nor does it report an adjustment strategy. This absence limits the ability to determine which comparisons were intended to carry the main confirmatory interpretation.

What do p-values and confidence intervals contribute?

A p-value is a measure of how incompatible observed data are with a specified null model, under the assumptions of the analysis. It is not the probability that the null hypothesis is true, nor does it measure the size or clinical importance of an effect. A confidence interval adds information about the range and precision of an estimate. Both need to be interpreted alongside the effect estimate, analysis population, and statistical plan. No primary outcome p-values or confidence intervals are reported in the registry record.

The registry record does not state the significance threshold used for the primary outcome analysis. The sample-size assumptions describe a planning illustration based on a two-sided significance level of 0.05, equal allocation, and a normal approximation. That planning description is not evidence that the trial’s own analysis used those same assumptions, and it does not report the study’s original design justification.

What does over-dispersion mean in general?

Over-dispersion occurs when observed variability is greater than a statistical model assumes. It matters because a model that understates variability can produce uncertainty estimates that are too narrow and tests that are too optimistic. This is a general concept; the primary endpoint is described as a mean of a continuous measure, and the registry record does not report an over-dispersion assessment or identify it as an analysis issue in this trial.

How should planning illustrations be read?

A sample-size calculation is a planning exercise: it estimates how many participants per arm might be needed under stated assumptions to detect a specified effect with a chosen power and significance level. The record describes its illustration as based on a fraction of the observed effect, using equal allocation and a normal approximation. Such an illustration is not observed power and is not proof that the trial was designed around those assumptions. The registry record does not provide the primary outcome effect estimate needed to interpret numerical illustrations based on the observed effect, and it does not report the trial’s original sample-size justification.

10. Multiplicity, Missing Data and Analysis Populations

The primary outcome is not a single score at a single assessment. It covers five ABC subscales and describes change through Week 12, including assessments at Week 4, Week 8, and Week 12. Considering multiple domains and times can be informative, but it also raises a multiplicity question: if many comparisons are tested, some may appear notable by chance even when there is no consistent underlying difference. This is a general statistical issue, not a finding about the trial.

The registry lists 14 outcome measures and 8 analyses. It does not report a statistical analysis plan, a hierarchy of primary comparisons, or a method for controlling multiplicity. The record also does not state whether any interim analyses were conducted. These details are not reported in the registry record, so it is not possible to establish from this information which comparisons were designated as primary for inference or how results across outcomes were intended to be interpreted.

For missing data, the primary outcome description says scores were assessed among participants with available data in the Full Analysis Set. The record does not define that set, state how participants entered or left it, report missingness by group or time point, or describe an imputation or other missing-data method. It also does not report whether the primary analysis used all available observations or required complete observations across assessment times. Missingness can affect interpretation if the likelihood of missing data is related to participants’ scores or treatment experience; that is a general concern, not a documented pattern in this record.

The available safety denominators are reported separately from the primary outcome population. They should not be used as substitutes for outcome-specific analysis counts. Without detailed analysis populations, missing-data handling, and multiplicity procedures, the precision and robustness of the primary comparison cannot be assessed.

11. Safety Results

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

Deaths

Other (non-serious) adverse events

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

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: suicidal ideation (2.0%), covid-19 (2.0%).

The registry reports serious adverse events affecting 2 participants in the GWP42003-P group and 1 participant in the placebo group. The corresponding numbers at risk are 49 and 54. These figures are counts, not estimates of a causal treatment effect, and they do not indicate whether the event counts differ beyond what could reasonably occur by chance. The registry record does not report a comparative statistical analysis or confidence intervals for these safety figures.

Suicidal ideation and COVID-19 are listed as serious-event terms, each with a reported value of 2.0%. The record does not identify the group allocation for these terms in the available information, so they should not be attributed to a particular treatment group. The registry reports zero deaths in each group.

Other adverse events are listed as affecting 14 participants in the GWP42003-P group and 20 in the placebo group. These descriptive counts need context, including event definitions, timing, exposure, severity, and handling of repeated events. That context is not fully reported in the registry record. A count of participants affected is different from a count of all event occurrences, and the record does not provide enough detail to resolve every event-level question. These figures describe what is listed; they do not establish whether either treatment is safer.

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

The registry record establishes the trial’s research question and design: a randomized, masked, placebo-controlled Phase 2 comparison of GWP42003-P with placebo in children and adolescents with autism spectrum disorder. It identifies caregiver-assessed ABC subscale change as the primary outcome and reports selected adverse-event counts. Those details describe what the trial set out to examine and some of the recorded safety information.

They do not provide the primary outcome results. No numerical changes, between-group differences, confidence intervals, or p-values for ABC scores are reported in the registry record. Consequently, the direction, magnitude, and precision of the primary comparison cannot be determined from this record. The absence of those estimates is not evidence that the groups were the same, and it is not evidence that one group had better scores.

The available adverse-event counts also have limits. They are descriptive, and the record does not provide comparative uncertainty estimates or sufficient event-level context to interpret them as evidence of a treatment-related difference. A small difference in counts, or a larger difference, is not by itself a measure of causality or clinical importance.

More broadly, a trial result is interpreted in light of its analysis population, missing-data methods, outcome hierarchy, multiplicity handling, and prespecified statistical plan. Several of these details are not reported in the registry record. The endpoint’s scale direction helps explain what a score change would mean, but does not supply the missing result. The registry’s status of completed likewise indicates trial status, not the direction or strength of its findings.

The appropriate statistical reading is therefore limited: the record describes a trial intended to compare GWP42003-P with placebo on ABC subscale changes and lists selected safety figures, but it does not report the primary outcome estimates needed to judge the central efficacy comparison. No treatment recommendation or conclusion about clinical benefit follows from these details alone.

13. Why This Trial Matters Statistically

This trial matters statistically because its primary question concerns multiple related behavioral domains measured repeatedly over time. A result for irritability, for example, would not automatically describe social withdrawal or the other ABC subscales. Likewise, a score at an earlier assessment does not necessarily represent the same pattern as a score at Week 12. Domain-specific and time-specific results require clear specification and careful interpretation.

The randomized, placebo-controlled design offers a structure for comparing changes between groups rather than relying only on change within the active group. Masking can also reduce certain biases in reported outcomes, which is relevant when the primary measure is caregiver-assessed. These design features support the intended comparison, but they cannot replace reporting the estimates and their uncertainty.

The registry record illustrates why outcome reporting must include more than the endpoint name. Readers need group-level results, between-group contrasts, precision estimates, analysis populations, and details of how multiple comparisons and missing observations were handled. Here, the record describes the primary measure and selected safety counts but does not report the numerical primary outcome results or key analysis details. That makes the limits of interpretation as important as the study’s design.

Finally, planning calculations and observed trial results answer different questions. Sample-size illustrations depend on assumptions about the effect and analysis, whereas outcome estimates describe what was observed. A planning illustration based on an observed effect should not be mistaken for observed power or for evidence that the design’s original assumptions were met. Distinguishing design, analysis, and results is essential for sound statistical reading of a clinical trial.

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.

Several limitations arise from what is and is not reported in the registry record. Most importantly, the record does not report numerical results for the primary ABC outcome. It gives the endpoint definition and time frame but not the group means, changes, between-group estimates, or uncertainty measures. This prevents assessment of the primary comparison’s size, direction, and precision.

The active treatment description is incomplete: it states that dosing begins at 5 mg/kg/day for 1 week and then gives 10 mg/kg/day, but the remainder of the schedule is not reported in the registry record. The placebo description specifies 12 weeks. The record does not supply further treatment exposure details that would clarify adherence, interruptions, or duration by participant.

The registry identifies stratification factors but does not report the resulting group characteristics or the statistical treatment of those factors. It also does not define the Full Analysis Set, provide primary-outcome analysis counts, or describe missing-data methods. This leaves uncertainty about how the available-data population relates to all randomized participants and how incomplete follow-up was handled.

The primary outcome includes multiple subscales and assessment times, but the record does not report a comparison hierarchy or multiplicity approach. It lists 14 outcome measures and 8 analyses without providing the associated analysis plan. The record also does not report whether interim analyses took place, the trial’s original sample-size justification, or detailed statistical models for the primary outcome.

Safety counts are limited to reported figures and do not include the full event-level context or comparative uncertainty estimates needed for a detailed statistical assessment. These limitations concern the information reported in the registry record; they do not establish what analyses may exist elsewhere.

15. Related Tutorials

16. Related Statistical Calculators

17. Sources

• ClinicalTrials.gov: Trial to Investigate the Safety and Efficacy of Cannabidiol Oral Solution (GWP42003-P; CBD-OS) in Children and Adolescents With Autism Spectrum Disorder — NCT04745026.
• Registry record last updated 2025-03-30; status last verified 2025-03.
• Results first posted 2025-03-30.
• 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 identifies a completed Phase 2, randomized, quadruple-masked, parallel-group trial of GWP42003-P versus placebo in children and adolescents with autism spectrum disorder. The recorded enrollment is 103. Stratification factors were age group, antipsychotic use, and region. The study’s primary outcome is change from baseline in caregiver-assessed ABC subscale total scores through Week 12, with lower scores indicating a better clinical outcome.

The registry record does not report numerical primary outcome estimates or the statistical analysis details needed to evaluate the main comparison. It lists serious adverse events affecting 2 participants in the GWP42003-P group and 1 in the placebo group, with zero deaths in each group. These figures are descriptive and do not establish a comparative treatment effect. Details about multiplicity, missing-data handling, and the trial’s original sample-size justification are not reported in the registry record.