Individually Tailored Lighting System to Improve Sleep in Older Adults: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Individually Tailored Lighting System to Improve Sleep in Older Adults (NCT01855126), a randomized na study. The primary outcome is Sleep Disturbance.

2. Clinical Question

3. Trial Design

NCT01855126 is a completed, randomized crossover study of individually tailored lighting for sleep disturbance. The registry lists the phase as NA, the sponsor as Rensselaer Polytechnic Institute, and enrollment as 46. The recorded start is 2013-03, and primary completion is recorded as 2017-04. The study status is COMPLETED.

The planned comparison involved a blue-light mask and a red-light mask. The registry describes two 8-week intervention/control periods. Each participant was to wear either the blue-light intervention mask or the red-light control mask every night during a period; the order of the two conditions was randomized by the study biostatistician. This within-participant crossover structure means that, in principle, each participant could contribute observations under both conditions. The record does not state how many participants completed both periods or how the posted summaries relate to the enrolled population.

Masking is listed as TRIPLE. The registry does not identify which people or roles were masked, how masking was maintained, or whether the different light colors could be distinguished. Those details should not be inferred from the masking label alone. The brief study description says that the device was intended to deliver individually timed light while participants were asleep, but the posted arm descriptions are truncated and do not provide a complete account of device operation.

The primary endpoint was assessed at baseline and week 8 of a lighting intervention. Blue Light Mask has a posted mean change of -1.9 units, with a standard deviation of 0.5 and n of 34; Red Light Mask has a posted mean change of -1.3 units, with a standard deviation of 0.5 and n of 34. The registry separately reports enrollment as 46. The relationship between these posted group counts and overall enrollment, including any crossover completion or missing outcome information, is not reported in the registry record.

The registry lists four outcome measures but posts no secondary outcomes in the record summarized for this trial. It reports no analysis entries. Consequently, the trial design is more fully described than the statistical analysis plan or participant flow.

4. Participant Flow and Accounting

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

Before assignment: 46 participants were enrolled into the study. These participants were randomized to begin the study with the active intervention, or the placebo intervention. After a 2 week washout period, participants were exposed to the opposite intervention.

First Intervention (8 Weeks)

Reasons for not completing this period:

Washout Period (2 Weeks)

Second Intervention (8 Weeks)

Reasons for not completing this period:

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 Sleep Disturbance, measured over Baseline (week 0) and week 8 of lighting intervention.

The primary outcome is change in the Pittsburgh Sleep Quality Index (PSQI) score. The scale ranges from 0 to 21, and the registry notes that a score over 5 is indicative of sleep disturbance. The outcome is described in units on a scale, with a mean and standard deviation reported for each mask condition.

Using change rather than a score at a single time point can describe how an individual's measured score differs between baseline and the end of an intervention period. Comparing changes between conditions can help distinguish the observed condition contrast from baseline score levels, although the analysis method matters, especially in a crossover study. The registry does not report enough detail to establish precisely how baseline measurements were used in the comparison or whether participant-level pairing was incorporated.

The registry states that a larger difference indicates a better outcome. The posted mean changes are -1.9 for the Blue Light Mask and -1.3 for the Red Light Mask. Because the record does not define the sign convention for the between-condition contrast, it is important to retain the condition-specific values rather than assume that the positive reported mean difference directly denotes benefit in one direction. The endpoint provides a standardized way to describe sleep-related symptoms, but a score change is not itself an explanation of the mechanism behind any observed difference.

The PSQI is used as a measure of reported sleep quality and disturbance. The registry does not report additional information about how the questionnaire was administered, how missing items were handled, or whether the meaning of a change was defined in advance for this trial. Clinical interpretation beyond the posted scale description is therefore limited.

7. Results Posted in the Registry

• Red Light Mask: mean -1.30 versus -1.90 on control; difference 0.60 (95% CI 0.36 to 0.84), standardized effect d = 1.20; z-test p = <0.001.

8. Statistical Interpretation

9. Statistical Methods Explained

Statistical methods determine how measurements become an estimate and how uncertainty around that estimate is assessed. The registry reports condition-specific mean changes and standard deviations, a mean difference, a confidence interval, and an unadjusted p-value. It does not report a complete statistical analysis plan or analysis entry. The interpretation must therefore distinguish the descriptive results from what can be established about the inferential method.

What does a mean change describe?

A mean is an average across the observations included in a group summary. Here, the outcome is change in PSQI score from baseline to week 8 of a lighting intervention. The posted mean changes are -1.9 units for the Blue Light Mask and -1.3 units for the Red Light Mask. Those values describe the average reported change in each condition, not the experience of every participant. A mean can conceal variation in individual responses, including people whose scores changed in different directions.

The standard deviation describes the spread of observations around a mean. The reported standard deviation is 0.5 for each condition. It is a description of variability in the outcome data and is not the same thing as the uncertainty in the estimated mean or in the difference between conditions. The standard deviation alone does not show whether the distribution is symmetric, whether there are unusual observations, or how participant-level changes varied across both periods.

What does the between-condition contrast represent?

The registry reports a mean difference of 0.6 for Red Light Mask relative to Blue Light Mask. A contrast is a numerical comparison of two conditions, and its interpretation depends on the order of subtraction and on how change is coded. The arm means are signed changes, while the registry's note that a larger difference indicates a better outcome does not fully define the coding of the reported contrast. The contrast should therefore be reported with its condition ordering, rather than translated into a claim about which condition had a better result.

Because this is a crossover design, the same participant may have measurements under both mask conditions. A paired analysis can use the fact that observations from the same person are related. Treating the two sets of observations as if they came from unrelated people can produce an inappropriate estimate of uncertainty when within-person correlation is present. The registry does not report whether the comparison used paired participant-level data, a model for repeated measurements, or a simpler calculation from group summaries. The posted group counts do not establish how many participants contributed complete paired outcomes.

Why can period and order matter?

In a crossover trial, outcomes may depend not only on the current condition but also on when that condition was received, the order of conditions, or effects that persist from an earlier period. Randomizing condition order can help prevent systematic assignment of one condition to a particular sequence, but it does not by itself show how an analysis addressed period or carryover effects. The registry says that the order was randomized, but does not report a statistical model, a washout procedure, or an assessment of period or carryover effects. Their handling is not reported in the registry record.

Triple masking is recorded, but the statistical consequences of masking depend on who was unaware of assignments and whether the masking was successful. Those operational details are not reported in the registry record. Masking does not replace appropriate handling of repeated observations or missing outcomes.

How should the p-value be understood?

The reported unadjusted p-value is p < 0.001. In general, a p-value measures how incompatible the observed data are with a specified null model, conditional on the test assumptions. It does not measure the size or importance of an effect, and it is not the probability that a finding is due to chance. The registry does not identify the test statistic, null hypothesis formulation, degrees of freedom, or assumptions used for this p-value. It also does not say whether the calculation accounts for the crossover design. The value is therefore best described as an unadjusted reported test result.

Adjustment for multiple testing may be relevant when many outcomes or contrasts are examined. The registry lists four outcome measures but reports no secondary posted outcomes and no analysis entries. It does not say how many tests were planned or conducted, or whether a multiplicity strategy was specified. A reader cannot determine from the record whether the reported p-value was selected from a larger set of analyses.

What does the confidence interval add?

A confidence interval gives a range of values compatible with the data under a specified method and confidence level. It conveys uncertainty in an estimated contrast in a way that the point estimate alone cannot. The registry reports an interval for the mean difference, and both bounds are positive. The confidence level and calculation method are not reported in the registry record, so the interval's exact coverage interpretation cannot be verified from the record. It also reflects only uncertainty captured by its calculation; it does not automatically account for bias from missing data, unreported analytic choices, or an incorrect model.

The registry includes pooled standard deviation and a standardized effect-size field in the comparison record. Such standardization expresses a mean contrast relative to a measure of spread, which can aid comparison across scales. It does not resolve the contrast's direction, determine clinical importance, or substitute for a crossover model. Those derived measures should be treated as descriptive comparison statistics rather than evidence of the sponsor's prespecified analysis, which is not reported in the registry record.

Overall, the available statistics support a description of the posted means and the registry's unadjusted contrast. Reproduction and full appraisal would require the participant-level analysis approach, the handling of period and sequence, missing-data procedures, and a prespecified testing strategy. These details are not reported in the registry record.

10. Confidence Intervals and Precision

The reported mean difference between the Red Light Mask and Blue Light Mask conditions is 0.6. The registry also reports a confidence interval for this contrast, with both bounds above zero. The interval gives a range of estimates compatible with the data under the method used to calculate it. Its bounds are not repeated as rounded values because the registry reports more decimal places than the permitted display precision. The confidence level and calculation method are not reported in the registry record.

Precision concerns how tightly an estimate is determined by the observed data and the analysis assumptions. A narrow interval generally indicates less sampling uncertainty than a wide interval, while still allowing for uncertainty that the model captures. The reported interval is consistent with the reported unadjusted p < 0.001 because it does not include zero. That correspondence does not establish that the analysis appropriately handled paired measurements, period effects, or other design features.

The group summaries include standard deviations of 0.5 in each condition, but these do not by themselves determine the precision of a crossover contrast. In a crossover design, the association between observations from the same participant can affect the uncertainty of the difference. The registry does not report the within-participant correlation, the variance calculation, or the exact analysis model. The stated interval should therefore be interpreted as the uncertainty reported for the registry contrast, not as an independently verified interval for a particular crossover analysis.

Precision also differs from clinical importance. An estimate may be precise without demonstrating that its magnitude matters to participants, and a broad interval can leave a range of possible effects compatible with the data. The registry provides no threshold for a clinically meaningful change. The PSQI score description notes that a score over 5 indicates sleep disturbance, but that threshold does not establish a meaningful threshold for change or for a between-mask contrast.

11. Planning a Similar Trial: Sample Size

This is a planning illustration, not a re-evaluation of the original trial. It asks: if a new study wanted to detect a difference like the one observed, or a smaller one, how large would it need to be? It is deliberately not “observed power” of the finished trial, which is not a meaningful quantity.

The planning figures are illustrations based on the observed effect, not the trial's original design assumptions. For a scenario using 100% of the observed effect, the planning entries are 11 participants per arm for 80% power and 15 participants per arm for 90% power. For 75% of the observed effect, the entries are 20 participants per arm for 80% power and 26 participants per arm for 90% power. For 50% of the observed effect, the entries are 44 participants per arm for 80% power and 59 participants per arm for 90% power.

These are participants per arm, not total events. The calculation assumes a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. The effect to detect is the stated fraction of the observed effect. The trial's own original sample-size assumptions and justification are not reported in the registry record.

In general, 80% power describes a design that would detect the specified effect in 80% of repeated studies under the stated assumptions if that effect and model were correct. The 90% column describes the corresponding design target of detecting that effect in 90% of repeated studies under those assumptions. These percentages are planning probabilities, not probabilities that the trial result is true. The figures are not observed power.

As the effect used for planning becomes smaller, more participants are needed to have the same target power, as the posted scenarios illustrate. The crossover structure of the trial should be kept in mind: the displayed assumptions specify equal allocation between two arms, while the trial description refers to two conditions received in randomized order. The planning calculation is an illustration based on its stated assumptions, not evidence that the original study was designed or analyzed using those assumptions.

12. How the Sample Size Is Derived

Power is the probability that a study of a given size rejects the null hypothesis when a specific true effect exists. It depends on four things: the effect to detect, the variability of the outcome, the significance level (here a two-sided α of 0.05, so z_1−α/2 = 1.960) and the desired power (z_1−β = 0.842 for 80% and 1.282 for 90%).

Formula

n is the number of participants in each arm, Δ is the difference in means to detect and σ is the common standard deviation.

Worked calculation for this trial

1. Observed difference Δ = 0.600; pooled standard deviation σ = 0.500.
2. n = 2 × (1.96 + 0.842)² × 0.500² / 0.600² = 10.9, rounded up to 11 participants per arm for 80% power.

The smaller-effect rows in the table above repeat the same calculation with the effect multiplied by 0.75 and 0.5 (for ratios, applied to the logarithm of the ratio).

R code

alpha <- 0.05
z_a   <- qnorm(1 - alpha / 2)          # two-sided
delta <- 0.6                      # observed difference in means
sigma <- 0.5                     # pooled standard deviation
n_per_arm <- function(power, frac = 1) {
  ceiling(round(2 * (z_a + qnorm(power))^2 * sigma^2 / (delta * frac)^2, 6))
}
for (f in c(1, 0.75, 0.5))
  cat(sprintf("%3.0f%% of observed effect: n per arm = %s (80%% power), %s (90%% power)\n", 100 * f,
              format(n_per_arm(0.80, f), big.mark = ","), format(n_per_arm(0.90, f), big.mark = ",")))

# Check against R's built-in function (t-based, so very slightly larger)
cat("power.t.test:", round(power.t.test(delta = delta, sd = sigma, power = 0.80)$n, 1), "\n")

# Simulation: exact two-sample t-test using simulated means and variances
set.seed(1855126)
n    <- n_per_arm(0.80)
reps <- 20000
m_diff <- rnorm(reps, delta, sigma * sqrt(2 / n))             # difference in sample means
s2     <- sigma^2 * rchisq(reps, 2 * (n - 1)) / (2 * (n - 1))   # pooled sample variance
tstat  <- m_diff / sqrt(s2 * 2 / n)
cat(sprintf("Simulated power at n = %s per arm: %.3f (%d replicates)\n", format(n, big.mark = ","), mean(abs(tstat) > qt(0.975, 2 * (n - 1))), reps))

Output from running the code

100% of observed effect: n per arm = 11 (80% power), 15 (90% power)
 75% of observed effect: n per arm = 20 (80% power), 26 (90% power)
 50% of observed effect: n per arm = 44 (80% power), 59 (90% power)
power.t.test: 11.9 
Simulated power at n = 11 per arm: 0.766 (20000 replicates)

The simulation draws many sets of group means and variances at the planned sample size and counts how often the two-sample t-test rejects the null. It should land close to the target power; at very small required sample sizes it can fall below it, because the normal approximation is rough there.

13. Multiplicity, Missing Data and Analysis Populations

The registry record does not report a statistical analysis plan, and it lists no analysis entries. It identifies four outcome measures, while no secondary outcomes are posted in the record summarized for this trial. It is therefore not possible to determine how many outcomes or contrasts were prespecified for formal testing, whether the primary outcome was selected before seeing the data, or whether any adjustment for multiple testing was planned or used.

Multiplicity arises when several outcomes, time points, subgroups, or comparisons are tested. In general, examining more tests can increase the chance of finding at least one small p-value even when there is no true effect. A prespecified testing strategy can clarify which findings are confirmatory and how the error rate is controlled. The registry does not report such a strategy. The posted p-value is explicitly unadjusted, and its relation to any broader set of analyses is not reported in the registry record.

Missing-data handling is also not reported. The registry gives n of 34 for each mask condition and enrollment of 46, but it does not explain participant flow, completion of both crossover periods, reasons for missing observations, or whether the summary counts represent all available outcome data. No imputation method, sensitivity analysis, or missingness assumption is described. A reader cannot determine whether incomplete follow-up could have affected the means or the contrast.

The analysis population is not specified. The record does not say whether results include all randomized participants, only those who completed both conditions, or another defined population. In a crossover study, eligibility for a paired analysis can depend on having outcome data from both periods; the registry does not report whether that was required. The relationship between randomized sequence, adherence to the nightly mask use, and inclusion in the endpoint analysis is not reported in the registry record.

These omissions do not invalidate the posted descriptive values, but they limit assessment of bias and inferential strength. Full interpretation would require the planned outcome hierarchy, analysis population, participant flow, missing-data methods, and multiplicity approach.

14. 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 Blue Light group and 0 participants in the Red Light group. It also reports 0 deaths in each group. The listed number at risk for serious adverse events is 34 in each group. These are the posted serious-event counts and should be read as counts, not as proof that no serious event could occur outside the recorded observation or reporting process.

The record separately reports other adverse events affecting 13 participants in the Blue Light group and 12 in the Red Light group, with 34 at risk in each group. Those other-event counts are not serious adverse-event counts. No serious adverse-event terms are listed. The registry does not provide event descriptions, timing, severity grading for other events, exposure-adjusted rates, or a detailed account of how adverse events were collected. It also does not clarify how participants who experienced events across both crossover periods were handled in the group summaries.

These safety figures describe what the registry posted; they do not establish that the masks are safe in general or that one condition has a safety advantage. Differences in follow-up, exposure, event definitions, or reporting could affect observed counts, and those details are not reported in the registry record.

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

The registry reports a mean PSQI change of -1.9 units during the Blue Light Mask condition and -1.3 units during the Red Light Mask condition. It reports a mean difference of 0.6, a confidence interval with positive bounds, and an unadjusted p < 0.001. These results describe a statistical contrast in the posted outcome summaries. Because the sign convention for the between-condition estimate is not fully explained, the direction of that contrast should not be translated into a claim of benefit without qualification.

The result does not establish why the score changes differed, whether light exposure altered a circadian mechanism, or whether the difference was important to participants. The study summary gives a proposed rationale involving light timing and dim light melatonin onset, but the primary outcome is PSQI change. No melatonin outcome is included in the primary result described in the registry record.

The result also does not establish how well the lighting system would perform for all older adults with sleep disturbance. Participant characteristics and eligibility details are not reported in the registry record, and the statistical analysis details needed to assess crossover effects are absent. The recorded enrollment and the condition-specific outcome counts are not reconciled in the registry. The result should therefore be understood as a posted comparison, not as a complete account of who contributed data or how the estimate was adjusted.

Statistical significance and clinical importance are distinct. A small p-value indicates incompatibility with a specified null model under the test assumptions; it does not say that a difference is large, meaningful, durable, or likely to apply in another setting. The registry gives the PSQI scale range and its score threshold for sleep disturbance, but it does not report a threshold for a meaningful change. The observed means alone cannot supply that missing criterion.

The safest conclusion is descriptive: the record reports different mean changes and an unadjusted statistical contrast between mask conditions. Its incomplete reporting of the analysis, missing data, and contrast coding prevents a stronger interpretation about direction, robustness, or practical significance.

16. Why This Trial Matters Statistically

This trial is statistically interesting because its randomized crossover design compares two lighting conditions within a study that describes participants receiving both conditions in randomized order. Crossover designs can make efficient use of within-person comparisons, because stable characteristics of an individual are shared across conditions. But that efficiency depends on collecting and analyzing repeated observations appropriately. A group-level contrast without a clearly described paired analysis leaves uncertainty about whether the posted inference uses the design's key feature.

The record also illustrates why the estimate, its uncertainty, and the analysis assumptions must be considered together. The mean contrast is 0.6 and the unadjusted p-value is p < 0.001, while the confidence interval is entirely on one side of zero. Those statistics are consistent with a clear numerical separation under the reported calculation. They cannot resolve unanswered questions about the sign convention, period effects, participant flow, or the confidence interval's construction.

There is a notable reporting gap between enrollment of 46 and n of 34 in each condition summary. Since this is a crossover study, the condition-specific counts should not automatically be treated as independent group sizes or added to infer the number of participants. The record does not explain how these counts relate to randomized participation or completion of both periods. That distinction matters for precision and for assessing missing-data bias.

Finally, the registry distinguishes a primary outcome from a broader set of listed outcomes, but reports no secondary posted outcomes or analysis entries. The lack of an analysis plan and multiplicity details makes it difficult to assess how the reported test fits within the intended analysis strategy. The trial is a useful example of why transparent reporting of design-specific models, participant flow, endpoint definitions, and analysis decisions is essential for interpreting statistical results.

17. Limitations

• The posted summary statistics are the only data these results rest on; patient-level data, full protocols and statistical analysis plans are not available.
• Recalculated results are simple, unadjusted calculations from posted summaries. They do not replace the sponsor's prespecified analysis.

Several limitations are specific to the information recorded for this trial. First, the participant population is not described in sufficient detail to assess eligibility, baseline sleep characteristics, or representativeness. The study summary identifies older adults with early-awakening insomnia as the intended context, but the registry does not report detailed characteristics of those enrolled.

Second, the registry describes a crossover schedule but does not report whether the analysis accounted for paired observations, treatment order, period, or possible carryover. It also does not describe a washout procedure. These omissions matter because outcomes from the two conditions may be related within a participant, and outcomes may vary by period or sequence.

Third, the reported enrollment of 46 and condition-specific n of 34 are not reconciled. The record does not describe the numbers randomized to each sequence, the number completing each period, adherence to nightly mask use, or reasons for missing outcome data. It is therefore not possible to determine how much of the enrolled population contributed to the posted comparison or whether missingness could have influenced it.

Fourth, the registry does not fully specify the primary comparison's sign convention, statistical model, confidence-interval method, or test assumptions. The p-value is unadjusted, while the record does not describe multiplicity control or the broader set of tests performed. A complete protocol-level analysis plan is not reported in the registry record.

Finally, the posted outcome is a change in PSQI score, but the registry does not define a threshold for a clinically meaningful change. The masking label and intervention summary also leave practical details of implementation and blinding unreported. These limitations restrict conclusions to the posted numerical summaries and their stated statistical contrast.

18. Related Tutorials

19. Related Statistical Calculators

20. Sources

• ClinicalTrials.gov: Individually Tailored Lighting System to Improve Sleep in Older Adults — NCT01855126.
• Registry record last updated 2019-01-29; status last verified 2019-01.
• Results first posted 2019-01-29.
• 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.

21. Record Summary

NCT01855126 is a completed randomized crossover study sponsored by Rensselaer Polytechnic Institute. The registry lists enrollment as 46 and describes two 8-week mask conditions, blue light and red light, with randomized order. Masking is listed as triple. The study concerns sleep disturbance in older adults and evaluates change in Pittsburgh Sleep Quality Index score from baseline to week 8.

The posted mean changes are -1.9 for the Blue Light Mask and -1.3 for the Red Light Mask, with a reported mean difference of 0.6 and unadjusted p < 0.001. The record reports no serious adverse events or deaths in either group. It does not report analysis entries, a full analysis plan, detailed missing-data handling, or how the condition-specific counts relate to enrollment and crossover completion.