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 Metabolic Effects of Sleep Extension in People With Obesity (NCT03594994), a randomized na study. The primary outcome is Insulin Sensitivity.
| Feature | Registry value |
|---|---|
| ClinicalTrials.gov identifier | NCT03594994 |
| Phase | NA |
| Condition | Insulin Resistance, Obesity |
| Allocation | Randomized |
| Design model | Parallel |
| Masking | None (open label) |
| Enrollment | 31 |
| Arms in the primary outcome | 2 |
| Lead sponsor | Washington University School of Medicine |
| Results posted | Yes |
2. Clinical Question
The trial asks whether extending time in bed to support longer sleep changes whole-body insulin sensitivity in people with obesity, compared with maintaining usual sleep patterns. The primary outcome is the change from baseline in insulin sensitivity after 4-6 weeks of intervention. The record describes the measure as the insulin-stimulated rate of glucose disposal per kilogram of fat-free body mass during a hyperinsulinemic-euglycemic clamp.
Population
People with obesity; insulin resistance is also listed as a condition. The registry reports enrollment of 31 participants. Further eligibility details are not reported in the registry record.
Intervention
Behavioral sleep extension: participants were instructed to extend sleep duration to >7h/night. The brief description says the goal was to increase time in bed to achieve >7h sleep/night.
Comparator
Participants in the control group were instructed to maintain their usual sleep patterns. The record describes this as maintaining current sleep habits.
Primary question
How did mean insulin sensitivity change from baseline after 4-6 weeks in each group? The posted results give group-specific means and confidence intervals, but not a direct between-group estimate of the difference in change.
This framing distinguishes the measured outcome from the broader question of whether sleep duration changed or whether any metabolic changes would persist beyond the intervention period. Those outcomes and their measurement details are not reported in the registry record. The primary question is specifically about insulin sensitivity measured using a clamp, rather than a general measure of glucose metabolism or a clinical health outcome.
3. Trial Design
The registry describes a randomized, parallel-group trial with no masking. It lists the study as completed, with enrollment of 31. The listed start date is 2018-10-29, and the primary completion date is 2024-07-15. The phase is listed as NA. Washington University School of Medicine is named as the sponsor.
Participants were assigned to sleep extension or control. The sleep-extension description sets a target of >7h/night; the control group was to maintain usual sleep patterns. The brief description anticipated 15 participants in each group. For the posted primary outcome, the control group has 15 participants and the sleep-extension group has 14. The reason for the difference between these outcome counts and the described group sizes is not reported in the registry record. The safety denominators are 16 in control and 15 in sleep extension, so those denominators do not match the primary-outcome counts.
Random allocation means participants are assigned to groups by chance, which is generally intended to limit systematic differences between groups at the outset. In a parallel design, each participant is in one assigned group rather than moving between interventions. These general features help frame the comparison, but they do not establish how randomization was generated or concealed: those details are not reported in the registry record. Masking is listed as none, so the record does not describe blinded assignment or outcome assessment.
The intervention is behavioral, and follow-up for the primary endpoint is described as after 4-6 weeks. The registry record does not report detailed instructions for implementing sleep extension, methods for measuring adherence or actual sleep duration, or whether other care was controlled. Those details should not be inferred from the assigned target alone.
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
| Stage | Control | Sleep Extension |
|---|---|---|
| STARTED | 16 | 15 |
| COMPLETED | 15 | 14 |
| NOT COMPLETED | 1 | 1 |
Reasons for not completing this period:
| Reason | Control | Sleep Extension |
|---|---|---|
| Adverse Event | 1 | 0 |
| Withdrawal by Subject | 0 | 1 |
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 | Control | Sleep Extension | Total |
|---|---|---|---|---|
| Participants analysed | 15 | 14 | 29 | |
| Age, Continuous | years | 44 (SD 11) | 39 (SD 13) | 41.8 (SD 11.8) |
| Sex: Female, Male: Female | Participants | 12 | 10 | 22 |
| Sex: Female, Male: Male | Participants | 3 | 4 | 7 |
| Race/Ethnicity, Customized: White (Caucasian) | Participants | 8 | 7 | 15 |
| Race/Ethnicity, Customized: Black | Participants | 6 | 7 | 13 |
| Race/Ethnicity, Customized: Asian | Participants | 1 | 0 | 1 |
| Body mass index | kg/m2 | 34.8 (SD 5.6) | 35.5 (SD 6.9) | 35.2 (SD 6.2) |
| Fasting glucose | mg/dL | 94.9 (SD 5.0) | 95.5 (SD 4.1) | 95.2 (SD 4.5) |
| 120 min glucose | mg/dL | 133 (SD 29) | 127 (SD 22) | 130.1 (SD 25.4) |
| Homeostatic model assessment of insulin resistance | 22.5/(µU/mL * mmol/L) | 4.3 (SD 2.3) | 4.6 (SD 2.3) | 4.4 (SD 2.3) |
| Hemoglobin A1c | % | 5.4 (SD 0.3) | 5.4 (SD 0.5) | 5.4 (SD 0.4) |
| Total sleep time | hours | 6.1 (SD 0.5) | 5.8 (SD 0.5) | 6.0 (SD 0.5) |
6. Primary Endpoint
The registry lists the primary outcome as Insulin Sensitivity, measured over Change from baseline testing after 4-6 weeks of intervention.
| Element | Registry entry |
|---|---|
| Result type | Mean |
| Measure of spread | 95% Confidence Interval |
| Unit | µmol/kgFFM/min |
| Endpoint classification | Continuous Endpoint |
The primary endpoint is mean change from baseline in whole-body insulin sensitivity after 4-6 weeks of intervention. It is expressed in µmol/kgFFM/min. The registry defines it as the insulin-stimulated rate of glucose disposal per kilogram of fat-free body mass during a hyperinsulinemic-euglycemic clamp.
In general, insulin sensitivity describes how responsive the body is to insulin. A clamp is a controlled measurement approach: insulin is infused while glucose is managed to maintain the target condition, and the glucose-disposal rate provides an index of insulin action. Expressing glucose disposal per kilogram of fat-free body mass accounts for the amount of lean tissue used as the denominator. This is a physiological measure, not a direct report of symptoms, diabetes diagnosis, or long-term clinical events.
Using change from baseline focuses the outcome on the difference between a participant’s starting measurement and the follow-up measurement. The registry reports group means for that change and 95% confidence intervals. Such summaries describe the average change and its statistical precision; they do not by themselves quantify the comparative effect of assignment to sleep extension versus control. A direct between-group estimate and its uncertainty would be needed to express that comparison in the same endpoint units.
The primary result is therefore relevant to the trial’s stated metabolic question, but it represents a short-term measure under clamp conditions. The record does not report additional details about the clamp protocol or the clinical importance threshold for a change in this measure.
7. Why No Re-analysis Is Shown
8. Statistical Interpretation
For the primary outcome, the posted control-group mean change is -3.8 µmol/kgFFM/min, with a 95% confidence interval from -8.6 to 0.9. The sleep-extension group mean change is 0.9 µmol/kgFFM/min, with a 95% confidence interval from -5.0 to 6.8. These are group-specific changes from baseline, not a reported estimate of the difference between the groups.
In general, a confidence interval describes uncertainty around an estimated quantity under the statistical method used. The intervals indicate that the group mean changes are estimated with substantial uncertainty: each interval spans values on both sides of zero. The intervals overlap, but overlap between separate group confidence intervals is not a formal test of the between-group contrast. The registry record does not report a direct comparison estimate, its confidence interval, or a p-value for the primary endpoint. Consequently, these results alone do not establish whether the average change differed between assigned groups.
The number of participants contributing to the primary outcome is 15 in control and 14 in sleep extension. This is smaller than the overall enrollment of 31, and differs from the safety denominators. The registry does not report why outcome availability differs from enrollment or from the safety counts. The record also does not report an analysis population definition or a method for handling missing primary measurements. These omissions matter because the interpretation of a group mean depends partly on which participants contributed data and how unavailable observations were addressed.
Interpreting the endpoint as a change also requires caution about baseline and follow-up measurements. The registry gives the mean change, but does not report the baseline and follow-up means separately in the posted results. It is therefore not possible to assess those component distributions from these summaries alone. Nor should an individual-level conclusion be drawn from a group average.
The record lists 17 secondary outcomes, including measures of hepatic and adipose-tissue insulin sensitivity, insulin resistance, 24-hour glucose and insulin area under the curve, non-esterified fatty acids, body mass, and fat-free mass. Numerical results for these secondary outcomes are not reported in the registry record. The primary result should not be treated as a summary of all those metabolic measures.
The posted estimates are descriptive of the observed groups over the stated intervention timeframe. They do not establish a causal effect from the group means alone, nor do they determine whether any change is clinically important or sustained. Those conclusions would require a clearly reported between-group analysis, relevant clinical context, and information about follow-up and data completeness.
9. Statistical Methods Explained
The registry reports a continuous primary endpoint summarized as a mean with a 95% confidence interval. This choice gives a group-level estimate of change in the measured quantity and an interval describing uncertainty around that estimate. Because the outcome is continuous, values can vary across a range rather than falling into categories such as event or no event.
What does a mean change represent?
In general, a mean is the arithmetic average of observations. A mean change is based on each participant’s follow-up value relative to their baseline value, then summarized across the group. It describes the average direction and size of change in the participants whose observations contribute to the analysis. A mean does not show the full distribution: participants may have changes that differ considerably from the group average. The registry record does not provide individual participant values or a measure of spread for the posted endpoint results.
The posted means are -3.8 for control and 0.9 for sleep extension, in µmol/kgFFM/min. These represent within-group changes from baseline. A randomized trial’s central comparison is usually the contrast between groups, because the control group provides a reference for changes that could occur without the intervention. Group-specific changes alone are not equivalent to that contrast. The registry record does not report the numerical between-group contrast or a statistical test for it.
How should the confidence intervals be read?
In general, a confidence interval is a range produced by a statistical procedure to convey the precision of an estimate. A 95% confidence interval is not a statement that there is a 95% probability that the particular fixed parameter lies inside the interval. Rather, under the assumptions of the method, the procedure would cover the target parameter in 95% of repeated applications to comparable data. The actual interpretation depends on the model and sampling assumptions, which are not described in detail in the registry record.
The control mean has a 95% confidence interval from -8.6 to 0.9; the sleep-extension mean has a 95% confidence interval from -5.0 to 6.8. Both intervals include zero. For each separate group mean, this indicates that the interval includes no average change as well as changes in either direction. It is not a substitute for a test of the between-group difference. A confidence interval around that direct difference would more directly express the uncertainty relevant to the trial’s comparative question.
Why does the analysis population matter?
In general, an analysis population defines which randomized participants are included in an estimate and under what assignment or data rules. Analyses based on everyone as randomized and analyses restricted to participants with observed follow-up can answer different questions, especially if some outcomes are unavailable. The primary outcome counts are 15 and 14, while enrollment is 31 and the safety denominators are 16 and 15. The registry record does not explain this difference, identify the analysis population, or describe missing-data handling. The reason for absent or excluded outcome values is not reported in the registry record.
Missingness can affect interpretation if the availability of measurements is related to participants’ characteristics or outcomes. In general, methods for handling missing values rely on assumptions, and the assumptions should be described alongside results. No such method or assumption is reported in the registry record. It would therefore be inappropriate to assume that all randomized participants contributed to the primary analysis or that missing results had no effect on the estimates.
What do the planning illustrations represent?
In general, a sample-size calculation estimates how many participants would be needed under specified assumptions to detect a target effect with a chosen statistical procedure and power. The registry record describes planning illustrations based on the observed effect, with a two-sided significance level of 0.05, equal allocation, and a normal approximation. These are planning illustrations, not the trial’s own design justification and not observed power. The record does not report the trial’s original sample-size justification, and the numerical planning results are not reported in the registry record.
The stated illustration approach considers an effect set as a fraction of the observed effect. In general, planning for a smaller target effect requires more information than planning for a larger target effect, other assumptions being equal. That is a general statistical relationship, not a conclusion about the trial’s actual design. A planning calculation also depends on assumptions about variability and the analysis method; full protocol assumptions are not reported in the registry record. Sample-size illustrations cannot establish whether the study was adequately planned or whether its observed results are conclusive.
Why might variability differ from a normal model?
In general, over-dispersion means that observed variability is greater than a statistical model expects. When present, it can make standard errors too small and confidence intervals too narrow if the model does not account for it. For a continuous outcome such as insulin sensitivity, the appropriate model and its assumptions should be evaluated using the actual data and analysis plan. The registry record does not report evidence of over-dispersion, a diagnostic assessment, or the statistical model used to produce the posted confidence intervals. No claim about over-dispersion in this trial can therefore be made from the registry summaries alone.
The registry record does not report an interim-analysis plan, adjustment methods, or a detailed statistical analysis plan. Those items should not be inferred from the reported means or from the planning illustrations. The posted numerical summaries permit a limited description of group changes and their reported uncertainty, but not a reconstruction of the trial’s full analysis.
10. Multiplicity, Missing Data and Analysis Populations
The registry lists 17 secondary outcomes and 18 outcome measures in total. The secondary outcomes include several related metabolic measures, body mass, and fat-free mass. Numerical results for those secondary outcomes are not reported in the registry record. No analysis entries are reported in the registry record, and the record does not describe a prespecified approach to controlling multiplicity.
In general, testing many outcomes can increase the chance that at least one result appears statistically notable by chance. A multiplicity strategy may adjust interpretation or testing to account for the number and relationship of outcomes. Without a reported statistical analysis plan or adjustment method, the secondary-outcome list should be read as a set of measured or intended outcomes, not as evidence that each was analyzed or that any finding was confirmed. The registry does not report whether secondary outcomes were prioritized or how they relate to the primary endpoint.
The primary endpoint results include 15 control participants and 14 sleep-extension participants. The registry does not explain why these counts differ from the enrollment of 31 or the safety denominators of 16 and 15. It does not report the amount, reasons, or pattern of missing primary-outcome data, nor the method used to handle missingness. The analysis population and whether analyses followed randomized assignment are also not reported in the registry record.
In general, missing-data assumptions matter because conclusions can depend on who has an observed measurement and why other measurements are absent. A complete-case summary, an imputation method, and a model using available observations may rely on different assumptions. Since the registry record does not identify the approach used, those alternatives cannot be evaluated. No conclusion should be drawn that unavailable values were similar to observed values or that missingness was unrelated to outcome.
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 |
|---|---|---|---|
| Control | 0 | 16 | 0.0% |
| Sleep Extension | 0 | 15 | 0.0% |
Deaths
| Arm | Deaths (participants affected of at risk) |
|---|---|
| Control | 0 of 16 |
| Sleep Extension | 0 of 15 |
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 |
|---|---|
| Control | 1 of 16 |
| Sleep Extension | 0 of 15 |
Most frequent non-serious adverse event terms
| Event term | Organ system | Control | Sleep Extension |
|---|---|---|---|
| Unrelated medical issue | Musculoskeletal and connective tissue disorders | 1 of 16 | 0 of 15 |
The posted safety figures report zero participants with serious adverse events in each group: 0 of 16 in control and 0 of 15 in sleep extension. The record also reports zero deaths in both groups. For other, non-serious adverse events, the figures are 1 of 16 in control and 0 of 15 in sleep extension.
These counts use safety denominators of 16 and 15. They are not the same as the primary endpoint counts of 15 and 14. The registry record does not explain the difference in denominators. The figures describe participants affected, as recorded in the registry; they do not provide event-specific detail for non-serious events. The serious adverse event terms list is empty, so specific serious event types are not reported in the registry record.
In general, safety counts are descriptive and should be interpreted with their denominators and observation context. Zero recorded events does not establish that an event is impossible, and the small group counts limit how much can be concluded about uncommon harms. The registry record does not report the detailed safety-monitoring procedures, event definitions, or a comparative statistical analysis of adverse events. These counts alone cannot establish that the groups had equivalent safety profiles.
12. What the Result Does, and Does Not, Mean
The result directly describes mean change in insulin sensitivity within each randomized group over the stated 4-6 weeks. The control mean change is -3.8 µmol/kgFFM/min, and the sleep-extension mean change is 0.9 µmol/kgFFM/min. Their respective 95% confidence intervals are -8.6 to 0.9 and -5.0 to 6.8. The intervals indicate uncertainty around each group’s mean, and both include zero.
The registry does not report a direct estimate of the between-group difference in change, its confidence interval, or a p-value. Therefore, the posted results do not provide the main comparative statistic needed to quantify how the groups differed on the primary endpoint. Separate group means and intervals should not be treated as a substitute for that comparison. The record also does not report a threshold for a clinically meaningful change, so the practical importance of the observed group means cannot be determined from these summaries alone.
The trial concerns a physiological measure collected using a hyperinsulinemic-euglycemic clamp. It does not directly report a clinical event, diagnosis, symptom, or long-term outcome. Although the record lists other metabolic outcomes, their numerical results are not reported in the registry record. The primary endpoint should not be generalized to all metabolic outcomes.
The results also do not establish whether participants achieved the intended sleep target, how consistently they followed the instructions, or whether any measured changes persisted after the intervention. Those details are not reported in the registry record. The no-masking design and limited duration are part of the context for interpreting the findings, but the registry summary does not show how these features affected the estimates.
In general, randomized assignment can support a causal comparison when the design, follow-up, outcome ascertainment, and analysis are adequately reported and conducted. Here, the registry’s posted primary summaries are insufficient to reconstruct that comparative analysis. The data can be described as the observed group-specific changes and their reported precision; they should not be turned into a claim that the intervention works or does not work.
13. Why This Trial Matters Statistically
This trial is statistically interesting because its primary outcome is a continuous physiological measurement, summarized as a mean change and a confidence interval. The summaries show both the estimated direction of average change within each group and the uncertainty around those estimates. They also illustrate why a within-group change is not the same quantity as a randomized between-group contrast.
The enrollment is 31, while the primary outcome counts are 15 and 14 and the safety denominators are 16 and 15. Keeping these counts distinct is important: they describe different sets of recorded observations, and the registry does not explain the differences. A statistical interpretation depends not only on the point estimates but also on who contributed data and how unavailable measurements were handled.
The record lists multiple secondary metabolic outcomes, but their numerical results are not reported. This makes the primary outcome the only posted numerical efficacy result in the record. In general, a broad outcome list can raise questions about multiplicity and selective reporting; without the analysis plan and results for the listed outcomes, those questions cannot be resolved from the record.
The posted confidence intervals include zero for both group means, while the direct comparative estimate is not reported. That contrast highlights a recurring statistical point: uncertainty around two separate means does not itself give the uncertainty around their difference. Reporting the comparative estimate would be necessary to assess the primary randomized question directly.
The record also describes planning illustrations based on the observed effect. In general, those can be useful for illustrating how sample-size needs depend on target effect and assumptions, but they are not evidence of observed power or a substitute for the original sample-size rationale. The registry record does not report that rationale. Overall, the trial provides a useful example of why statistical interpretation needs clear estimands, analysis populations, missing-data methods, and between-group estimates—not only arm-level summaries.
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 registry record leaves several trial-specific details unreported. It does not give the eligibility criteria beyond the listed conditions, the method for generating or concealing random allocation, or detailed intervention and control procedures. It also does not report how actual sleep duration or adherence was measured. The target of >7h/night describes the assigned goal, not documented sleep achieved by participants.
The primary-outcome counts are 15 in control and 14 in sleep extension, compared with overall enrollment of 31. The safety denominators are 16 and 15. The reason these numbers differ is not reported in the registry record. Baseline and follow-up outcome summaries, participant-level variation, and the handling of missing measurements are also not reported. These omissions limit assessment of the composition and completeness of the primary analysis.
The posted endpoint results give group-specific mean changes and confidence intervals, but no direct between-group estimate, confidence interval for the contrast, or p-value. The record does not report the statistical model, analysis population, or a detailed analysis plan. It is therefore not possible to assess the reported comparative effect or independently evaluate all assumptions underlying the summaries.
The trial is described as having no masking and a 4-6 week intervention timeframe. The record does not provide details sufficient to assess how these features affected measurement or participant behavior. Secondary outcome names are listed, but their numerical results are not reported in the registry record. The record also does not report a clinical-importance threshold or longer-term follow-up for the primary outcome.
These are limits of what can be interpreted from the registry record; they do not establish that a particular design or analysis choice was inappropriate. The original protocol and fuller results are not reported in the registry record.
15. Related Tutorials
16. Related Statistical Calculators
17. Sources
- ClinicalTrials.gov: Metabolic Effects of Sleep Extension in People With Obesity — NCT03594994.
- Registry record last updated 2025-12-11; status last verified 2025-11.
- Results first posted 2025-12-11.
- 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 describes a completed, randomized, parallel-group trial without masking, sponsored by Washington University School of Medicine. Enrollment was 31. Participants were assigned to behavioral sleep extension, with a target of >7h/night, or to maintain usual sleep patterns. The primary outcome was mean change in whole-body insulin sensitivity after 4-6 weeks, measured using a hyperinsulinemic-euglycemic clamp.
Primary outcome results were posted for 15 control participants and 14 sleep-extension participants. Mean changes were -3.8 µmol/kgFFM/min and 0.9 µmol/kgFFM/min, respectively, with 95% confidence intervals of -8.6 to 0.9 and -5.0 to 6.8. A direct between-group estimate is not reported in the registry record. Safety figures report zero serious adverse events and zero deaths in each group; one other adverse event is reported in control and none in sleep extension.
The record lists 17 secondary outcomes, but their numerical results are not reported. Details on missing-data handling, analysis population, and the original sample-size justification are also not reported in the registry record.