SweetDreams Sleep Study for Children With Autism: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for SweetDreams Sleep Study for Children With Autism (NCT04452045), a randomized na study. The primary outcome is Change in Children's Sleep Habits Questionnaire (CSHQ).

2. Clinical Question

In practical terms, the study asks whether an online adaptation can be tested in this population and how participants’ questionnaire scores changed during the recorded treatment timeframe. Its stated aim includes preliminary information about feasibility and effectiveness, but the registry’s posted primary results are group summaries rather than a complete statistical evaluation. The findings should therefore be read as descriptive evidence about the recorded scores, not as a definitive estimate of the intervention’s comparative effect.

3. Trial Design

The registry describes SweetDreams as a randomized, single-masked, crossover study sponsored by the University of California, San Francisco. It lists enrollment of 20 and a completed status. The recorded study start is 2020-09-01, and primary completion is recorded as 2021-12-30. Phase is listed as NA.

A crossover design generally allows participants to receive more than one condition in an ordered sequence, so results may depend on period, sequence, and within-participant comparisons. That is general information about crossover studies, not a description of details confirmed for this trial. The registry record does not report the treatment sequence, the timing of periods, any washout arrangements, or whether the posted change summaries account for repeated observations from the same participant.

The posted outcome summaries are organized under SweetDreams and Waitlist Control. The SweetDreams group has n=9, and the Waitlist Control group has n=11. These are the group counts attached to the posted summaries; the registry does not explain how these counts relate to crossover sequence or period participation. It also does not describe which people were masked, how masking was maintained, or whether masking was assessed.

The intervention is described as access to an online adaptation of existing evidence-based sleep interventions, available through mobile and computer access. The waitlist condition is described as receiving future access to that adaptation. The primary outcome timeframe is baseline to end of treatment, with average exposure of 1 month. The registry record does not report further detail about intervention content, adherence, or the timing and conduct of outcome assessments.

For interpretation, the registry’s design label and the available outcome summaries should be kept distinct. Randomization and crossover describe the planned structure, while the posted summaries show means and standard deviations by listed group. The record does not include enough analysis detail to determine whether those summaries represent period-specific, sequence-specific, or participant-level crossover estimates.

4. Participant Flow and Accounting

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

Overall Study

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 Change in Children's Sleep Habits Questionnaire (CSHQ), measured over baseline to end of treatment (average exposure 1 month).

The primary endpoint is change in the Children’s Sleep Habits Questionnaire (CSHQ). The registry describes it as a validated questionnaire with a total score ranging from 33 to 99, where higher values represent more sleep problems. The posted timeframe is baseline to end of treatment, with average exposure of 1 month. Change scores are intended to describe how a participant’s measured score differs across that timeframe; the registry does not specify the exact calculation convention used for the posted summaries.

A questionnaire outcome can capture reported sleep-related behavior in a structured way, making it useful when the study question concerns changes that may not be represented by a single clinical event. A validated instrument has undergone evaluation for its measurement purpose, but that does not by itself show that a particular change is important to participants or that it was caused by the intervention. No threshold for a meaningful change is reported in the registry record.

The scale direction matters when reading the posted means: higher scores correspond to more sleep problems, so a negative change describes a movement toward a lower score. That direction is a description of the measure, not proof of a treatment effect. The record gives mean change and standard deviation for each listed group, but no confidence intervals, significance tests, or individual score distributions. The primary endpoint should therefore be read alongside the uncertainty and analysis information that is absent from the registry record.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The registry gives a randomized, single-masked, crossover design label and reports mean changes with standard deviations for a continuous primary measure. Those features describe the study’s broad structure and the format of the posted results. They do not identify the statistical model used to compare conditions. No analysis is listed in the registry record, and the exact analysis plan is not reported there.

What a mean change summarizes

A mean is a way to describe the central value of a set of observations. Here, the posted outcome is change in CSHQ score from baseline to end of treatment, with average exposure of 1 month. A mean change combines the participant changes into a group summary. It can help readers see the general direction of scores, but it conceals the individual values and does not show whether nearly all participants changed similarly or whether the mean reflects a more varied pattern.

The standard deviation is a summary of spread around the mean. A larger spread indicates that participant changes vary more widely around the group average. It is not an estimate of the uncertainty in the mean, and it is not itself an interval for the intervention-versus-control effect. The standard deviations are 12.3 for SweetDreams and 12.7 for Waitlist Control. The registry does not post the participant-level observations needed to inspect the distributions directly.

Why the crossover structure matters

In general, crossover studies can use within-participant comparisons, because participants may experience more than one condition. Those comparisons account for the fact that observations from the same person tend to be related. Depending on the design, period and sequence may also need attention. A simple comparison of independent group means does not necessarily represent the appropriate analysis for a crossover study. The registry record does not report how the crossover structure was handled, whether the posted summaries pool periods, or whether the analysis used participant-level data.

The group labels and counts in the posted results are SweetDreams, n=9, and Waitlist Control, n=11. The record does not explain whether these counts refer to participants assigned to a sequence, participants contributing outcome data in a period, or another analysis population. It is therefore not possible to reconstruct the analysis population or identify how incomplete crossover participation may have been handled.

What an inferential comparison would add

A treatment comparison would estimate the contrast between conditions and quantify uncertainty around that estimate. A confidence interval is a range of values compatible with the data and the analysis assumptions; its width conveys precision. A hypothesis test can summarize how compatible the observed data are with a specified null model, but a p-value does not measure the size or importance of an effect. These are general explanations. The registry record includes no confidence interval or p-value for the primary result, and no inferential comparison is posted.

For a crossover design, the comparison should reflect the repeated measurements and any relevant period or sequence structure. Other design details, including how outcomes were recorded and how missing observations were treated, can affect the estimate and its uncertainty. The registry record does not report a model, covariates, estimation method, or assumptions for the primary outcome analysis. It is not appropriate to infer these from the design label alone.

Reading the planning illustrations

The sample-size section on this page is a planning illustration based on the observed effect, not the trial’s own design justification. Its stated assumptions are a two-sided significance level of 0.05, equal allocation between arms, and a normal approximation; the effect to detect is a stated fraction of the observed effect. Sample-size illustrations describe hypothetical planning under those assumptions. They are not observed power, do not show the probability that this completed study detected an effect, and do not replace the missing analysis of the posted trial results.

The registry record does not report the trial’s own sample-size justification or the assumptions used to determine enrollment. The total enrollment is listed as 20, while the posted outcome summaries have the group counts stated above. Those numbers should not be taken as evidence of a particular design target or a power calculation. The results do not include enough information to independently assess the adequacy of the planning assumptions for a crossover design.

What cannot be inferred

Because no analysis is posted, readers cannot determine from this record whether the sponsor used a model that respected randomization, crossover structure, or baseline measurements. Nor can they determine whether there were interim analyses, whether any multiplicity adjustment was planned, or whether a prespecified statistical plan governed the outcome analysis. These items are not reported in the registry record. General statistical principles can explain why they matter, but they cannot fill in those missing trial-specific details.

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("SweetDreams", "Waitlist Control")
n   <- c(9, 11)
m   <- c(-8.78, -0.59)
sp  <- c(12.3, 12.7)
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  SweetDreams  9  -8.780  12.300  -18.235  0.675
CARM  Waitlist Control  11  -0.590  12.700  -9.122  7.942

11. Multiplicity, Missing Data and Analysis Populations

The registry lists 2 outcome measures and identifies change in CSHQ as the primary outcome. It lists change in Pittsburgh Sleep Quality Index (PSQI) as a secondary outcome. The record does not include PSQI result values or a posted analysis. It also does not report a multiplicity strategy for handling more than one outcome or multiple comparisons. Without a stated strategy, readers cannot tell whether any reported inference was intended to be confirmatory or exploratory; in this record, no inferential analyses are posted.

Missing-data handling is not reported in the registry record. The record does not describe how missing questionnaires, incomplete treatment periods, or participants without an end-of-treatment measure were handled. It also does not state whether the displayed means use all randomized participants, participants with available measurements, or another analysis population. The listed group counts, n=9 and n=11, are attached to the outcome summaries, but the analysis-population definition is not reported.

These omissions matter in a crossover study because participants may contribute measurements under more than one condition, and missing observations can affect within-participant comparisons. General methods for handling missing data rely on assumptions about why data are missing and how observed information relates to unobserved values. The registry record does not identify any such assumptions or methods, so their appropriateness cannot be assessed.

The record also does not report an intention-to-treat, per-protocol, or other analysis population. It is not possible to infer the use of any of these approaches from the randomized design label or the posted group counts. As a result, the analysis population and the possible influence of missing outcome data remain uncertain.

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.

The posted safety figures show no serious adverse events among the 9 participants at risk in the SweetDreams group and none among the 11 participants at risk in the Waitlist Control group. Other adverse events are also recorded as 0 in each group, and deaths are recorded as 0 in each group. No top serious adverse-event terms are listed.

These entries describe the events recorded in the registry for the stated groups. A count of zero means that no event of that category is shown in the posted results; it does not establish that no event occurred outside the record’s capture or reporting. The record does not provide additional detail about surveillance procedures, event definitions, reporting completeness, or the time at risk beyond the group counts.

Safety summaries should be interpreted in relation to the size and scope of the study and the way events were collected. The listed at-risk counts are limited, and the registry record does not present an analysis of event rates or a comparison with uncertainty. These figures therefore describe the posted adverse-event counts; they should not be treated as conclusive evidence about safety in a broader population.

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

The posted data show mean decreases in CSHQ scores for both listed groups over the baseline-to-end-of-treatment timeframe. The SweetDreams mean change is -8.78, with standard deviation 12.3 and n=9. The Waitlist Control mean change is -0.59, with standard deviation 12.7 and n=11. Because higher CSHQ scores indicate more sleep problems, negative changes correspond to lower group scores at the end of the recorded timeframe.

Those summaries do not establish that SweetDreams caused a change or that its change differed reliably from the waitlist condition. The registry record contains no confidence interval, p-value, or adjusted comparison estimate for the primary endpoint. It also does not describe how the crossover structure was included in the analysis. Without this information, the precision and inferential strength of a comparison cannot be evaluated from the posted results.

The record does not report a threshold for a meaningful CSHQ change. Therefore, the group means cannot be classified from this information as representing a clinically important improvement. The study summary describes an aim to gather preliminary feasibility and effectiveness data, but the posted outcome information is not a complete assessment of either. No formal feasibility result is included in the information, and the listed PSQI secondary outcome has no reported result value.

These limitations do not make the summaries meaningless: they describe the direction and spread of recorded group changes. But descriptive statistics answer a narrower question than an analyzed randomized comparison. They do not determine whether changes are attributable to the intervention, whether a difference is compatible with chance, or whether results would generalize beyond the participants represented in the posted summaries.

The most defensible reading is therefore bounded: the registry reports negative mean CSHQ changes in the two listed groups, along with their standard deviations and counts. It does not report enough statistical detail to support a precise comparative conclusion. The study is best understood from these entries as a completed preliminary investigation with limited posted outcome analysis, not as a definitive estimate of comparative effect.

14. Why This Trial Matters Statistically

This trial raises several important statistical questions because its registry design is crossover while the posted primary outcome is presented as group-level mean change. In a crossover study, participants can contribute information under more than one condition, and a suitable comparison may need to account for the relationship between measurements from the same participant. Seeing the design label alongside the summaries highlights why a group mean and standard deviation are not always enough to understand the intended contrast.

The study also illustrates the distinction between feasibility aims and formal effectiveness analysis. The registry summary describes an online adaptation intended to improve access to sleep health care and to collect preliminary feasibility and effectiveness data. The posted primary results give questionnaire summaries, but no analysis estimate or uncertainty interval. Reporting the analysis model, the analysis population, and the handling of missing outcome data would help readers understand how the summaries relate to the randomized design.

Another statistical lesson is that scale direction and uncertainty must be read together. The CSHQ assigns higher scores to more sleep problems, so negative mean changes have a clear descriptive direction. Yet the standard deviations describe variation, not the precision of a treatment comparison. A confidence interval or other comparison estimate would help convey what the data support about the contrast, while a prespecified meaningful-change threshold would give context for interpreting its practical importance. Neither is reported in the registry record.

Finally, planning illustrations should not be mistaken for trial evidence. The sample-size section uses stated hypothetical assumptions and an effect based on observed results. Such illustrations can show how planning inputs relate to sample-size requirements, but they do not constitute observed power or a substitute for the trial’s own design rationale. The registry record does not report that rationale. Together, these points make this study useful for understanding how design, analysis, and reporting determine what can be learned from a small set of posted summaries.

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 registry record leaves important trial-specific details unreported. Although it identifies the study as randomized, single-masked, and crossover, it does not describe the sequence of conditions, period structure, or whether there was a washout. It does not identify who was masked. These omissions limit the ability to understand how the design was implemented and how the posted group summaries correspond to participants’ experiences across conditions.

The record gives no analysis model, comparison estimate, confidence interval, or p-value for the primary endpoint. It does not state whether baseline scores, period effects, or participant-level correlations were incorporated. It also does not report a prespecified statistical analysis plan, interim analyses, or the sample-size justification. These details are not reported in the registry record and should not be inferred from general practice.

Outcome reporting is limited to the primary group summaries in the information. The secondary outcome is listed as change in PSQI, but its result is not reported in the registry record. There is no reported meaningful-change threshold for CSHQ, and no detailed account of questionnaire completion, adherence, or timing beyond the baseline-to-end-of-treatment timeframe and average exposure of 1 month.

Missing-data methods and analysis populations are also not reported. The posted counts do not explain whether all randomized participants contributed to every outcome or how incomplete participation was handled. Because the record lists enrollment of 20 and separate outcome-summary counts, it is not possible to infer the reasons for the counts or the disposition of each participant.

Finally, this completed study has limited posted safety information and no detailed event-collection description. The adverse-event counts are recorded, but the registry does not describe the surveillance process or provide uncertainty estimates. Taken together, these limitations mean that the record supports a descriptive reading of the reported changes, while leaving substantial uncertainty about comparative analysis, missing data, and broader applicability.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: SweetDreams Sleep Study for Children With Autism — NCT04452045.
• Registry record last updated 2025-12-02; status last verified 2025-11.
• Results first posted 2025-12-02.
• 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

NCT04452045 is a completed randomized, single-masked, crossover study sponsored by the University of California, San Francisco. It lists enrollment of 20 and evaluates SweetDreams, an online adaptation of sleep interventions accessible by mobile device and computer, against a waitlist condition with future access. The primary outcome is change in CSHQ score from baseline to end of treatment, with average exposure of 1 month.

The posted mean changes are -8.78 for SweetDreams and -0.59 for Waitlist Control, with standard deviations of 12.3 and 12.7 and group counts of n=9 and n=11. The record lists change in PSQI as a secondary outcome but reports no result value for it in the information. No inferential analysis, confidence interval, or p-value is posted for the primary outcome.

The registry records zero serious adverse events, other adverse events, and deaths in each listed group. It does not report the crossover analysis details, missing-data methods, analysis populations, or sample-size justification. The results therefore offer descriptive questionnaire summaries, but not a fully documented estimate of the comparative effect.