This page provides an independent statistical analysis and educational interpretation of publicly reported results. ClinicalTrials.gov provides the official trial registry record. Numerical trial results on this page are restricted to the results posted in the ClinicalTrials.gov record; calculations added by Clinical Biostats are labelled as such.
1. Trial at a Glance
This page covers the registry results for Pilot Study of Sleep Therapy and Biomarkers in Children With Autism Spectrum Disorders (NCT02487082), a randomized phase 2 study. The primary outcome is Aberrant Behavior Checklist (ABC) Lethargy/Social Withdrawal Subscale.
| Feature | Registry value |
|---|---|
| ClinicalTrials.gov identifier | NCT02487082 |
| Phase | Phase 2 |
| Condition | Autism Spectrum Disorder |
| Allocation | Randomized |
| Design model | Parallel |
| Masking | Quadruple |
| Enrollment | 12 |
| Arms in the primary outcome | 2 |
| Lead sponsor | Stony Brook University |
| Results posted | Yes |
2. Clinical Question
Population
The study concerned children with autism spectrum disorders, a population described in the registry record as more likely than other children to have sleep problems. The primary endpoint results are reported for participants who completed the trial. The record does not provide further demographic detail for that endpoint population, such as ages or other participant characteristics.
Intervention
The intervention was melatonin and donepezil. The Group A description lists melatonin 3mg and donepezil 1.25-5mg each night, with dosage varying by age and weight. The description gives a treatment period of 10 weeks across Phase 1 and Phase 2. The registry record does not provide enough detail to resolve all dosing and timing information for every participant.
Comparator
The comparator was placebo. Both groups had a placebo run-in phase. Group B's description lists placebo in Phase 1, followed by melatonin and donepezil or placebo in Phase 2 for 5 weeks. The record describes Group A as receiving active medication in Phase 1 and Phase 2. It does not provide a separate posted estimate for each treatment period.
Primary question
The primary outcome was the Aberrant Behavior Checklist Lethargy/Social Withdrawal Subscale, rated by a parent or caregiver. The registry record gives endpoint medians of 6.0 for Group A and 5.5 for Group B among trial completers. These are descriptive summaries; no statistical comparison or uncertainty interval for the group contrast is reported in the registry record. The endpoint was collected after a washout period.
3. Trial Design
This was a Phase 2, randomized, quadruple-masked, parallel study sponsored by Stony Brook University. The registry record lists an enrollment of 12 and the condition as autism spectrum disorder. It records a start date of 2015-06, a primary completion date of 2020-12-31, and a terminated status. The reason for termination and the date the study was terminated are not reported in the registry record.
The study description concerns the combined use of melatonin and donepezil to assess sleep and behavior. The intervention listings are melatonin and donepezil, and placebo. The listed arm descriptions include a placebo run-in and subsequent phases. Group A is described as receiving active medication during Phase 1 and Phase 2, for 10 weeks. Group B is described as receiving placebo during Phase 1 and active medication or placebo during Phase 2, for 5 weeks. The descriptions also mention a washout period without active or placebo medication. The registry record does not clearly explain how the Phase 2 alternatives in Group B were assigned or how the timing differs across participants.
Randomization indicates that assignment was planned to be made by chance, while quadruple masking indicates that the record classified the study as masked at four levels. Which people or roles were masked is not reported in the registry record. A parallel model means the trial was registered with separate groups, rather than describing a single group that all received treatments in sequence. The registry record does not report a fuller statistical analysis plan or the procedures used to preserve masking.
Although the record lists enrollment of 12, the primary endpoint population is described as participants who completed the trial. Endpoint summaries are posted for Group A and Group B, but the endpoint group sizes are 2.0 and 6.0, respectively. The record does not explain why the endpoint population differs from the enrollment figure. It also does not report reasons for non-completion.
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.
Run-In Phase (5 Weeks)
| Stage | Group A | Group B |
|---|---|---|
| STARTED | 4 | 8 |
| COMPLETED | 3 | 6 |
| NOT COMPLETED | 1 | 2 |
Reasons for not completing this period:
| Reason | Group A | Group B |
|---|---|---|
| Withdrawal by Subject | 1 | 2 |
Phase 1 (5 Weeks)
| Stage | Group A | Group B |
|---|---|---|
| STARTED | 3 | 6 |
| COMPLETED | 2 | 6 |
| NOT COMPLETED | 1 | 0 |
Reasons for not completing this period:
| Reason | Group A | Group B |
|---|---|---|
| Withdrawal by Subject | 1 | 0 |
Phase 2 (5 Weeks)
| Stage | Group A | Group B |
|---|---|---|
| STARTED | 2 | 6 |
| COMPLETED | 2 | 6 |
| NOT COMPLETED | 0 | 0 |
Wash-Out Phase (2-12 Weeks)
| Stage | Group A | Group B |
|---|---|---|
| STARTED | 2 | 6 |
| COMPLETED | 2 | 6 |
| NOT COMPLETED | 0 | 0 |
Analysis population for the primary outcome, as defined in the registry: Participants who completed the trial.
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.
| Characteristic | Unit | Group A | Group B | Total |
|---|---|---|---|---|
| Participants analysed | 4 | 8 | 12 | |
| Age, Continuous | years | 11 (7 to 15) | 9.5 (7 to 16) | 9.5 (7 to 16) |
| Sex: Female, Male: Female | Participants | 0 | 1 | 1 |
| Sex: Female, Male: Male | Participants | 4 | 7 | 11 |
| Race and Ethnicity Not Collected | Participants | — | — | 0 |
| Region of Enrollment: United States | participants | 4 | 8 | 12 |
| ABC Lethargy/Social Withdrawal | units on a scale | 8 (3 to 23) | 10 (5 to 28) | 9.5 (3 to 28) |
| ABC Stereotypic Behavior | units on a scale | 7.5 (3 to 12) | 7.5 (1 to 15) | 7.5 (1 to 15) |
| Sleep Duration (hours) | hours | 8.2 (7 to 9) | 7 (4.5 to 8) | 7.1 (4.5 to 9) |
6. Primary Endpoint
The registry lists the primary outcome as Aberrant Behavior Checklist (ABC) Lethargy/Social Withdrawal Subscale, measured over Subscale rating was made by parent/caregiver every 5 weeks for a total of 5 times. Due to small N, only the endpoint of the study is reported here. Study endpoint (Visit 5) was after at least 18 wks, including 2 wks of washout, up to a maximum of 27 wks..
| Element | Registry entry |
|---|---|
| Result type | Median |
| Measure of spread | Full Range |
| Unit | score on a scale |
| Endpoint classification | not reduced to a single comparison |
The primary endpoint was the Aberrant Behavior Checklist (ABC) Lethargy/Social Withdrawal Subscale. A parent or caregiver rated the subscale at repeated visits; the registry record says ratings were made every 5 weeks for a total of 5 times. Because of the small N, only the study endpoint is reported. Visit 5 occurred after at least 18 wks, including 2 wks of washout, up to a maximum of 27 wks.
The subscale includes 16 items about behaviors such as being listless or inactive, preferring solitary activities, and showing few social reactions. Each item is rated on a 4 point scale, with 0 meaning not at all and 3 meaning severe. Possible scores range from 0 to 48, with higher scores indicating more lethargy and/or social withdrawal. The registry description notes that ABC subscales were derived by factor analysis and that the checklist has been used in research with children with autism.
Using this endpoint gives a caregiver-rated behavioral measure relevant to the study's stated focus on sleep and behavior. The posted group summaries are medians and full ranges, not means or model-adjusted estimates. The endpoint description identifies the population as participants who completed the trial. The registry record does not report a baseline-to-endpoint change estimate or an analysis that adjusts for baseline scores.
7. Why No Re-analysis Is Shown
8. Statistical Interpretation
The endpoint summaries are medians of 6.0 for Group A and 5.5 for Group B. Their corresponding full ranges are 0.0 to 12.0 and 1.0 to 33.0. These are descriptive statistics for the participants who completed the trial, with endpoint group sizes of 2.0 and 6.0. The values describe the observed scores; by themselves, they do not establish a reliable difference between groups.
The full ranges show that scores varied within both groups, and the Group B range extends from 1.0 to 33.0. A median is the middle of the ordered observations and is less influenced by extreme values than a mean. With very small endpoint groups, however, a median can represent only a limited amount of information about the wider population. The ranges also do not indicate how scores were distributed between their endpoints or how individual participants changed from baseline.
No inferential analysis is posted for the primary endpoint. In particular, the registry record does not report a p-value, confidence interval, adjusted effect estimate, or model-based comparison. Without those results, the precision of a group contrast cannot be quantified from the registry endpoint summaries. The absence of a reported statistical test should not be read as evidence that the groups were equivalent, nor should the close medians be treated as proof of equivalence.
The primary outcome is reported at the final visit rather than across the repeated ratings. As a result, the posted endpoint summaries do not describe the trajectory of scores over time. They also do not show whether any changes observed during treatment persisted, diminished, or differed after washout. The registry record does not provide the repeated measurements or an analysis of them.
The trial is listed as terminated, and the registry records enrollment of 12. The reported endpoint population is smaller than that enrollment figure, but the record does not give reasons for this or describe how missing endpoint data were handled. These limitations make the posted summaries especially uncertain as estimates of a treatment contrast. They are best interpreted as small-sample descriptive results from completers, not as a definitive estimate of the effect of the intervention.
9. Statistical Methods Explained
The posted primary outcome is summarized using the median and full range for each group. A median is the central value after observations are ordered. It is useful for describing a small sample when the values may be unevenly distributed, because an unusually high or low observation has less influence on the median than on the arithmetic mean. The median does not, however, convey the full shape of the data or the uncertainty around a treatment contrast.
A full range reports the lowest and highest observed values. It conveys the span of the observations but not how many values lie near either end, how tightly values cluster, or how likely the same range would be in another sample. The ranges in this trial should therefore be read alongside the small endpoint group sizes, rather than as a precise characterization of the underlying population.
For a treatment comparison, descriptive summaries are usually complemented by an estimate of the contrast and a measure of uncertainty. A confidence interval is a statistical interval constructed using a method intended to capture the underlying quantity at a stated long-run rate under its assumptions. It indicates the precision of an estimate; it is not a guarantee that the true value lies within any particular reported interval. No confidence interval for the primary group contrast is reported in the registry record.
A p-value is a measure, under a specified statistical model and null hypothesis, of how incompatible the data are with that hypothesis. It is not the probability that the null hypothesis is true, nor does it describe the size or clinical importance of a difference. The registry record does not report a primary endpoint p-value or a statistical test. The group medians alone cannot substitute for such an analysis.
What do the endpoint summaries establish?
They establish the posted endpoint medians and observed ranges for the completer population. They do not show baseline-adjusted changes, individual trajectories, or a model-based comparison. A contrast of endpoints can be affected by differences that were already present before treatment; the record does not report baseline endpoint values or an adjustment for them. The record also does not explain whether the reported summaries were prespecified as the primary analysis.
How should the planning illustration be read?
A sample-size calculation is a planning exercise about how many participants might be needed under specified assumptions. It is not an analysis of the trial's observed evidence and is not observed power. The planning assumptions described in the registry record use a two-sided significance level of 0.05, equal allocation, and a normal approximation, with the effect to detect expressed as a fraction of the observed effect. The record does not report a sample-size justification from the trial's original design. Any planning illustration based on an observed effect should be treated cautiously when the underlying estimate comes from a very small endpoint population.
In general, a planning calculation is sensitive to the effect size chosen, the variability assumed, the statistical test, and the allocation. Smaller target effects generally require more information to detect than larger target effects under otherwise comparable assumptions. These are general properties of sample-size planning, not results established by this trial. Such calculations do not show that a future study will achieve its planned power, and they do not repair uncertainty in the observed estimate.
Why does the small endpoint population matter?
With few observations, each participant's score can have substantial influence on group summaries and estimated contrasts. This makes results sensitive to chance variation and makes precise estimation difficult. It can also limit the ability to assess whether standard statistical assumptions are reasonable. The median and range remain valid descriptions of the posted observations, but population-level conclusions require more than descriptive summaries.
Over-dispersion is a general statistical issue in which observed variability exceeds what a model assumes. If a model understates variability, it can produce confidence intervals that are too narrow and tests that appear more decisive than they should. The registry record does not report a model for this endpoint, so over-dispersion is not a reported finding in this trial. It is a general consideration when selecting models for outcome data, not a basis for adding an unreported analysis.
The record identifies the outcome as a score and reports medians with full ranges. It does not state whether any statistical method was used to account for the repeated assessments, the phase structure, or the washout period. The primary endpoint summaries should therefore be understood as the posted endpoint description, not as evidence that these design features were modeled.
10. Multiplicity, Missing Data and Analysis Populations
The registry record lists the ABC Lethargy/Social Withdrawal Subscale as the primary outcome and lists four secondary outcomes: the ABC Stereotypic Behavior Subscale, sleep duration, clinician-rated global improvement, and parent-reported global improvement. It records five outcome measures and no posted analyses. The record does not describe a procedure for controlling multiplicity, such as adjusting the interpretation of tests across multiple outcomes or time points. Because no inferential analyses are posted, there are no reported multiplicity-adjusted or unadjusted test results to interpret.
Repeated ratings were scheduled every 5 weeks for a total of 5 times, but the registry says only the endpoint is reported because of the small N. It does not report results for earlier visits or an analysis using the repeated ratings. The record also does not state how the endpoint timing was handled analytically or whether any repeated-measures method was planned.
The primary outcome population is described as participants who completed the trial. Endpoint group sizes are 2.0 for Group A and 6.0 for Group B. The registry record does not report reasons for missing endpoint data, a missing-data strategy, or whether any values were imputed. It does not identify an intention-to-treat population or provide a separate per-protocol analysis. Consequently, the posted summaries cannot establish how participants without reported endpoint scores were handled or whether missingness could have affected the results.
These omissions limit interpretation of the endpoint summaries. They do not, by themselves, establish that a particular analysis or data-handling method was or was not used; such details are not reported in the registry record.
11. Safety Results
The registry reports serious adverse events as affected participants divided by participants at risk.
| Arm | Participants with serious adverse events | Participants at risk | Percentage |
|---|---|---|---|
| Group A | 0 | 4 | 0.0% |
| Group B | 0 | 8 | 0.0% |
Deaths
| Arm | Deaths (participants affected of at risk) |
|---|---|
| Group A | 0 of 4 |
| Group B | 0 of 8 |
Other (non-serious) adverse events
Participants with at least one non-serious event at or above the registry's 0% reporting threshold.
| Arm | Participants affected of at risk |
|---|---|
| Group A | 2 of 4 |
| Group B | 2 of 8 |
Most frequent non-serious adverse event terms
| Event term | Organ system | Group A | Group B |
|---|---|---|---|
| difficulty falling asleep | Nervous system disorders | 2 of 4 | 2 of 8 |
The serious adverse event figures are reported separately by group. Group A lists 4 participants at risk, 0 affected by a serious adverse event, and 0 deaths. Group B lists 8 participants at risk, 0 affected by a serious adverse event, and 0 deaths. These counts describe the safety information posted for the listed groups; they do not estimate the rate of rare events with precision.
The record also lists other adverse events for both groups: 2 affected participants in Group A and 2 in Group B, with 4 and 8 participants at risk, respectively. The registry record does not provide the terms for serious adverse events, and the list of top serious terms is empty. These other-event counts should not be confused with serious adverse events.
Zero recorded serious events in a small study does not demonstrate that a serious event is impossible or that a treatment is free of risk. Small samples have limited ability to characterize uncommon harms. The posted counts also do not provide details about event timing, severity beyond the serious/non-serious grouping, relationship to study medication, or the way adverse events were collected. Those details are not reported in the registry record.
12. What the Result Does, and Does Not, Mean
The posted results describe endpoint scores among the participants who completed this trial. The Group A median was 6.0, with a full range of 0.0 to 12.0; the Group B median was 5.5, with a full range of 1.0 to 33.0. Those values are a snapshot of observed scores at the reported endpoint, not a complete account of the treatment experience across the trial.
The results do not establish that the intervention changed lethargy or social withdrawal, nor do they establish that it did not. The registry record provides no statistical comparison, confidence interval, or baseline-to-endpoint change analysis. It is therefore not possible to assess the precision of a group contrast using the posted endpoint summaries alone. Close medians do not prove that the groups are equivalent, and a broad range does not by itself show whether a treatment effect is present.
The findings also do not answer every question named in the study summary. The record lists sleep duration and other behavioral or global-improvement measures as secondary outcomes, but it does not post analyses for them. The primary endpoint is a caregiver-rated behavior subscale, rather than a direct measure of sleep. The posted results cannot be used to infer unreported outcomes or to determine whether changes in sleep and behavior were related.
The trial's terminated status and the small endpoint population are important context. The registry record does not give the reason for termination or explain why endpoint group sizes differ from the enrollment figure. Without those details, the effect of incomplete follow-up on the observed summaries cannot be assessed.
Overall, these are descriptive results from a small completed-participant population. They can inform understanding of what was recorded, but they do not provide a precise or definitive estimate of the intervention's effect. They should not be read as proof of benefit, proof of no benefit, or proof of safety.
13. Why This Trial Matters Statistically
This trial illustrates the difference between a registered study question and the evidence available for interpretation. The study was randomized and quadruple-masked, but the primary endpoint results are descriptive medians and ranges from a small group of participants who completed the trial. Randomization and masking are valuable design features, yet they do not remove the need for an adequately informative analysis and transparent reporting.
It also demonstrates why the analysis population matters. The registry records enrollment of 12, while the endpoint group sizes are 2.0 and 6.0. The record does not explain this difference or report how missing endpoint data were handled. A reader needs both the observed outcome summaries and an account of who contributed to them to assess how broadly the results may apply.
Reporting the median and full range helps show the spread of observed scores, but these summaries do not quantify uncertainty about the group contrast. Without a reported effect estimate and confidence interval, readers cannot judge the precision of that contrast. Without a posted inferential analysis, they also cannot evaluate the evidence against a specified null hypothesis. Statistical restraint is therefore important: descriptive values should remain descriptive.
Finally, a planning illustration based on an observed effect is not observed power and does not overcome limitations in a small study. The distinction between design assumptions, observed outcomes, and inferential conclusions matters when considering whether further research may be informative. The record offers a limited endpoint snapshot, while leaving key questions about completion, analysis, and uncertainty unanswered.
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 primary endpoint summaries are based on participants who completed the trial, and the endpoint group sizes are 2.0 and 6.0. The registry record lists enrollment of 12 but does not explain the difference, report reasons for non-completion, or describe missing-data handling. This makes it difficult to assess whether the completer summaries could differ from results for all enrolled participants.
Only the endpoint at Visit 5 is reported, despite repeated ratings every 5 weeks for a total of 5 times. The registry record does not provide earlier outcome values, baseline scores, or trajectories. It therefore does not show change over time or whether the groups differed before the endpoint. The specified endpoint timing spans at least 18 wks and up to a maximum of 27 wks, including 2 wks of washout, so the endpoint reflects assessment after differing overall durations as described in the record.
The intervention descriptions include different treatment periods and a Phase 2 alternative for Group B. The registry record does not fully explain the treatment assignment or provide a statistical analysis that accounts for the phases. It also does not report who was masked, the analysis population, the method used for the primary comparison, or a confidence interval.
The study is listed as terminated, but the reason is not reported in the registry record. No primary endpoint inferential analysis is posted, and the listed secondary outcomes do not have posted analyses. These gaps limit interpretation beyond the descriptive endpoint values and adverse-event counts.
15. Related Tutorials
16. Related Statistical Calculators
17. Sources
- ClinicalTrials.gov: Pilot Study of Sleep Therapy and Biomarkers in Children With Autism Spectrum Disorders — NCT02487082.
- Registry record last updated 2024-06-06; status last verified 2024-05.
- Results first posted 2024-06-06.
- 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.
Continue through the Clinical Biostats statistical pathway
Use this trial's endpoint as a starting point for deeper study of effect measures, confidence intervals, and sample-size planning.
18. Record Summary
The registry record describes a randomized, quadruple-masked, parallel Phase 2 study sponsored by Stony Brook University. It concerns melatonin and donepezil compared with placebo in children with autism spectrum disorders. Enrollment is listed as 12. The study started in 2015-06, has a primary completion date of 2020-12-31, and is listed as terminated.
The primary outcome is the ABC Lethargy/Social Withdrawal Subscale at the endpoint, reported for participants who completed the trial. The medians are 6.0 in Group A and 5.5 in Group B, with full ranges of 0.0 to 12.0 and 1.0 to 33.0. No statistical analyses are posted. The registry record also lists four secondary outcomes and safety counts, but does not report a primary inferential comparison, reasons for termination, or an account of missing endpoint data.