Investigational Test of a New Sleep Supplement: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Investigational Test of a New Sleep Supplement (NCT03567343), a randomized na study. The primary outcome is Sleep Diary- Sleep Latency.

2. Clinical Question

3. Trial Design

The registry identifies NCT03567343 as a completed, randomized, quadruple-masked, parallel-group study sponsored by the University of Arizona. Its official title describes a randomized, double-blind, placebo-controlled study of a proprietary spearmint extract blend and sleep in healthy men and women. The record lists the study phase as not applicable. Enrollment was 105. The recorded start date is 2018-05-18, and primary completion is dated 2019-10-15.

Participants were assigned to the proprietary spearmint extract blend or control. The active-arm description specifies 500 mg/day taken by mouth every night, 30 minutes before bed. The control description specifies microcrystalline cellulose on the same schedule. Both descriptions include sleep diaries and Fitbit Charge 2 monitoring for 30 days. The active-arm text also mentions that a subset of participants would undergo something further, but that description is truncated in the registry record; its details are not reported in the registry record.

The parallel design means the listed groups are separate arms, while randomization is intended to allocate participants to those arms by chance. Masking is recorded as quadruple, although the registry record does not identify which roles were masked or describe how masking was maintained. Further details about allocation procedures, stratification, sequence generation and concealment are not reported in the registry record.

The primary outcome is sleep-diary sleep latency, measured as change from baseline after 30 days of supplementation. The registry lists additional sleep-diary and Fitbit outcomes, as well as mood and cognitive performance among the areas of interest. It records 17 outcome measures and 46 analyses. Those counts indicate that the posted results cover more than the primary endpoint, but they do not, by themselves, explain how outcomes were prioritized or how multiple analyses were handled.

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

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: Final analyses included only 43 participants (4 excluded) in the active group, and 46 participants (3 excluded), due to missing data.

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: Final analyses included only 43 participants (4 excluded) in the active group, and 46 participants (3 excluded), due to missing data.

6. Primary Endpoint

The registry lists the primary outcome as Sleep Diary- Sleep Latency, measured over Change from baseline after 30 days supplementation.

The primary endpoint is sleep-diary sleep latency: the time, in minutes, from lights being turned off until falling asleep. The registry describes the measure as weekly averages and defines the analysis timeframe as change from baseline after 30 days of supplementation. A change score expresses how the recorded measure differed from its baseline level over the stated period.

Sleep latency captures one specific aspect of sleep: how long it takes to fall asleep. It is not a measure of total sleep duration, sleep continuity, or sleep quality as a whole. That distinction matters because the registry also lists outcomes such as total sleep time, wake after sleep onset and sleep efficiency, measured through diaries or Fitbit monitoring. These outcomes describe related but non-identical features of sleep, and a result for one should not be treated as a result for all of them.

Using a mean summarizes the average change across the participants included in an analysis. This can make group-level patterns easier to describe, while also concealing differences among individual participants. The confidence interval communicates uncertainty around each estimated mean change. In the posted primary results, the estimated changes are negative in both groups, but their intervals extend across zero. Those are separate within-group estimates; they do not directly quantify whether the groups changed differently from one another.

The endpoint is based on a participant-completed diary, according to its description. The registry also names objective sleep monitoring for Fitbit outcomes, but it does not describe the sleep-latency endpoint as a Fitbit measure. Details of the diary instrument, its validation, scoring procedures and handling of daily entries are not reported in the registry record.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The statistical methods below explain the meaning of the methods named in the registry and the limits of the information recorded. Details of any analysis plan beyond what is stated in the registry record are not reported in the registry record.

What does a mean change describe?

Sleep latency is a continuous measure recorded in minutes. A mean change summarizes the average difference from baseline among the participants included in an arm's analysis. A negative mean change indicates that the average recorded latency was lower at the assessed follow-up than at baseline; a positive mean change would indicate a higher average. The reported means are group summaries, not predictions for an individual participant and not evidence that every participant changed in the same direction.

A mean is useful for describing the center of a distribution, but its interpretation depends on the measurement process and the pattern of participant data. The registry gives no distributional summaries beyond the means and confidence intervals for this primary outcome. Measures of variability among individual participants, checks of model assumptions and diagnostic results are not reported in the registry record.

What does a paired t-test test?

A paired t-test is generally used when measurements are paired within the same person, such as baseline and follow-up observations. It tests whether the average within-person difference is compatible with a specified null value, commonly no average change. In this study's posted analyses, a paired t-test is reported separately for each arm. Thus, each test addresses change within that arm rather than the difference in change between randomized arms.

A within-arm test is not a randomized comparison. Participants in both arms may change over time for reasons that are unrelated to the assigned intervention, including changes in measurement or circumstances during follow-up. A direct comparison between arms is needed to distinguish the pattern in one group from the pattern in the other. The registry record does not report an estimable direct comparison for the primary continuous endpoint.

In general, a t-test relies on assumptions about the data being analyzed, including the distribution of paired differences for a paired test. The registry names the method but does not report diagnostics or detailed assumption checks. It also does not describe adjustments or covariates used for the primary endpoint. Such methods should not be presumed.

How should the confidence intervals be read?

A confidence interval gives a range of estimates compatible with the data and the method used, under the method's assumptions. It reflects uncertainty in an estimate; it does not show the range of individual participants' outcomes. A wider interval indicates less precision than a narrower interval on the same scale. The reported intervals are 95% confidence intervals for the mean change within each arm.

For the active group, the mean change is -2.1 minutes and its 95% confidence interval runs from -5.03 to 0.82. For control, the mean change is -2.53 minutes and its 95% confidence interval runs from -6.3 to 1.24. Each interval includes zero. This means the intervals are compatible with no mean change as well as with changes in either direction. It does not establish that the groups are equivalent, and it does not tell whether a direct between-group effect is absent.

The registry labels both analyses as equivalence analyses. However, an equivalence conclusion requires a prespecified margin defining which differences are small enough to count as equivalent, along with an appropriate comparison and analysis. Such a margin and a between-group equivalence analysis are not reported in the registry record. The label alone is not enough to establish equivalence between the intervention and control.

What does randomization contribute?

In general, randomization is intended to make assigned groups comparable, on average, with respect to factors known and unknown at assignment. This supports causal interpretation of a properly conducted between-group comparison. It does not replace that comparison: separate within-arm tests do not use the randomized assignment to estimate the difference in outcomes attributable to assignment.

The registry identifies the trial as randomized and parallel, but additional details about the randomization procedure are not reported in the registry record. The record also does not state whether the primary analysis followed intention-to-treat principles, how adherence was considered, or how participants with missing endpoint data were handled beyond the stated exclusions.

What do the planning illustrations mean?

The registry includes a planning illustration based on an observed effect, with a two-sided significance level of 0.05, equal allocation between arms and a normal approximation. It describes the target effect as a stated fraction of the observed effect. Such calculations are planning scenarios, not evidence that the study had a particular probability of detecting an effect. They are not observed power, and they are not the trial's documented design justification. The specific design rationale and sample-size justification are not reported in the registry record.

Planning estimates based on an observed effect can be sensitive to the effect chosen as a reference. They do not resolve uncertainty in the posted primary results, supply a missing between-arm estimate, or establish that a future study would produce the same result. The registry's planning assumptions should therefore be kept distinct from the trial's observed outcomes and from any unreported protocol analysis plan.

10. Analyses Reported by the Sponsor

The sponsor's posted analyses are paired t-tests conducted separately within the proprietary spearmint extract blend and control arms. They report mean changes of -2.1 and -2.53 minutes, respectively, with the corresponding 95% confidence intervals. Those analyses are the sponsor's own reported results and take precedence over a simple recalculation if a discrepancy is present; the registry record does not flag a discrepancy.

These sponsor analyses do not estimate the randomized contrast between arms. The primary results available for each arm are not enough to reconstruct a valid unadjusted two-arm comparison from the posted summary statistics, and the registry record does not report an estimable direct comparison for this continuous outcome. There is therefore no recalculated between-arm estimate to set against the sponsor's analysis.

It is important to keep the scope of the sponsor's tests clear: each concerns within-arm change, not whether assignment to the blend produced a different change than assignment to control. The registry record does not report a separate sponsor analysis of that contrast, nor does it provide enough information to infer one. The presence of a reported analysis for each arm should not be mistaken for a reported treatment-versus-control test.

11. Multiplicity, Missing Data and Analysis Populations

The registry records 17 outcome measures, 15 posted secondary outcomes and 46 analyses. That breadth raises a general statistical issue: when many outcomes or analyses are examined, the chance of at least one apparently notable result can increase if each is assessed separately without accounting for multiplicity. This is a general concern, not evidence that any specific result in this trial is misleading.

The registry record does not report a multiplicity adjustment, a hierarchy for testing outcomes, or a prespecified strategy for distinguishing confirmatory from exploratory analyses. It also does not identify whether particular secondary outcomes were selected for emphasis before or after the data were examined. The protocol's multiplicity plan is not reported in the registry record. The number of posted measures and analyses alone cannot show how the investigators handled this issue.

Missing data are directly relevant to the primary outcome. The record says that the final analyses included 43 participants in the active group and 46 in control, with 4 and 3 excluded, respectively, due to missing data. It does not describe the amount or pattern of missing sleep-latency observations, why the data were missing, or whether missingness differed across follow-up weeks. The record does not describe imputation, sensitivity analyses or another method for assessing the effect of missing data.

The registry also does not specify whether the primary results use an intention-to-treat population, a per-protocol population, or another defined analysis population. The reported final analysis counts should therefore be read as the counts included in those posted analyses, not as proof of a particular population rule. Whether all randomized participants were included in other outcomes is not reported in the registry record.

These omissions matter because missingness and multiple testing can affect the precision and interpretation of results. Without details of the analysis plan, readers cannot determine whether the posted tests reflect prespecified decisions, how missing observations were treated, or whether results across multiple endpoints were adjusted or prioritized.

12. Safety Results

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

Deaths

Other (non-serious) adverse events

Participants with at least one non-serious event at or above the registry's 0% reporting threshold.

Most frequent non-serious adverse event terms

The registry reports serious adverse events affecting 0 participants in the proprietary spearmint extract blend group and 0 in control. It also reports 0 deaths in each group. For other adverse events, the record lists 6 affected participants among 53 at risk in the blend group and 9 among 52 at risk in control.

These figures describe the safety information posted in the registry record. The serious-event counts are not the same as the primary sleep-latency analysis counts: the record lists 53 and 52 participants at risk for adverse events, while the primary analysis included 43 and 46 participants. Different outcome populations or reporting processes may be involved, but the reason for the difference is not reported in the registry record.

A count of zero reported serious events does not establish that no serious event could occur, and the posted counts alone do not establish a comparative safety effect. The registry does not list serious adverse-event terms. Definitions, ascertainment procedures, timing, severity details and the nature of the other adverse events are not reported in the registry record. The counts should be understood as a limited summary of events recorded during the study, not as a complete characterization of safety.

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

The primary results describe average changes in sleep-diary latency within each randomized arm over the stated timeframe of 30 days. The mean change is -2.1 minutes in the proprietary spearmint extract blend group and -2.53 minutes in control. Each 95% confidence interval includes zero, so the posted arm-specific estimates are uncertain about whether there was an average change in either group.

Those findings do not establish that the intervention and control groups had the same change. Nor do they demonstrate that one group changed more than the other. The reported paired tests are within-arm analyses, and a direct two-arm estimate is not reported as estimable in the registry record. The result should therefore not be described as showing a comparative benefit, lack of benefit or equivalence.

The endpoint also represents one dimension of sleep, measured through a diary. The registry lists other sleep outcomes, including total sleep time, wake after sleep onset and sleep efficiency, as well as Fitbit-based measures. Results for sleep latency cannot stand in for all of these outcomes, and the record's broad mention of mood and cognitive performance does not mean that the primary endpoint measures either one.

Interpretation is further limited by the analysis population and missing-data information. The primary analysis included the listed subsets of the enrolled groups, with exclusions attributed to missing data. The registry does not give enough detail to assess the potential influence of those exclusions or to know whether the reported analysis population followed a prespecified rule.

In general, statistical uncertainty and clinical importance are different questions. A confidence interval describes uncertainty around an estimate, whereas judging importance requires context about what size of change matters and how the measure is interpreted. A threshold for clinically meaningful change in sleep latency is not reported in the registry record. The posted data alone do not support medical advice or a treatment recommendation.

14. Why This Trial Matters Statistically

This trial illustrates a basic distinction in randomized studies: a change within one arm is not the same quantity as a difference between arms. The posted primary estimates and paired t-tests describe within-group changes. Because the direct comparison is not reported as estimable, readers cannot use these results to quantify the randomized contrast for sleep latency. Recognizing that distinction helps prevent overinterpretation of separate arm-level results.

It also shows why uncertainty should be read alongside a point estimate. The confidence intervals for both reported mean changes include zero and cover values in both directions. A point estimate alone can make a result seem more definite than it is; the interval conveys that the data are compatible with a range of average changes under the analysis method.

The record offers another useful teaching example in how outcome breadth complicates interpretation. It lists many outcome measures and analyses, but does not report a multiplicity plan. A large set of results can be informative about the scope of measurement, while also making prespecification and transparent reporting important for judging which findings are confirmatory.

Finally, enrollment and analysis populations are not interchangeable. The registry records enrollment of 105, while the primary analysis description gives smaller final analysis groups and attributes exclusions to missing data. Understanding who contributed data to a particular result is essential when considering how broadly it applies. The absence of detail on missing-data methods and analysis populations constrains that assessment.

These are general statistical lessons drawn from the way this record reports its design and results. They do not determine whether the supplement changes sleep. They clarify which questions the posted analyses answer and which comparisons remain unreported.

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.

Several trial-specific limitations affect interpretation. The most consequential is that the registry posts separate within-arm paired t-tests for sleep latency but does not report an estimable direct comparison between the randomized groups. The primary results therefore do not quantify the difference in change that is central to a randomized intervention-versus-control question.

The primary analysis excludes participants because of missing data: the record lists 43 included and 4 excluded in the active group, and 46 included and 3 excluded in control. It does not describe the timing, reasons or pattern of missingness, or the methods used to address it. The primary analysis population is not identified as intention-to-treat or by another named rule. These gaps make it difficult to assess potential bias from incomplete outcome data.

The record lists many outcome measures and analyses but does not report how multiplicity was handled or which secondary analyses were prespecified. It also does not report the primary comparison's planned analysis, a clinically meaningful threshold for sleep-latency change, or the study's sample-size justification. Planning illustrations based on an observed effect do not fill those gaps.

Measurement details are limited. Sleep latency is described as a sleep-diary measure based on weekly averages, but the diary instrument, scoring details and procedures for summarizing entries are not reported. The record mentions Fitbit monitoring for study arms and lists Fitbit outcomes, but it does not describe the device-measurement algorithm or how diary and device results were reconciled.

Some descriptions are incomplete. The active-arm description ends after referring to a subset, and the control description is also truncated. The roles included in quadruple masking, allocation procedures, adherence, and the details of any protocol deviations are not reported in the registry record. The adverse-event summary gives participant counts but not event terms or detailed ascertainment information. These limitations do not invalidate the posted figures, but they restrict independent assessment of how they were generated and what they represent.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: Investigational Test of a New Sleep Supplement — NCT03567343.
• Registry record last updated 2024-07-18; status last verified 2024-02.
• Results first posted 2024-07-18.
• 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

NCT03567343 is recorded as a completed, randomized, quadruple-masked, parallel-group study sponsored by the University of Arizona. Enrollment was 105. It compared a proprietary spearmint extract blend with control in a study described as involving healthy men and women, with a 30-day regimen that included sleep diaries and Fitbit Charge 2 monitoring.

The primary outcome was change from baseline in sleep-diary sleep latency after 30 days. The registry reports separate mean changes and 95% confidence intervals for the two arms, based on final analysis groups of 43 and 46 participants. Both intervals include zero. The posted primary analyses are within-arm paired tests, and an estimable direct comparison between arms is not reported.

Missing-data methods, multiplicity handling, the primary comparison plan, and several design and measurement details are not reported in the registry record. The safety summary records no serious adverse events or deaths in either group and lists other adverse events for both groups. These findings should be interpreted within the limits of the posted analyses and descriptions.