Non-Pharmacological Treatments for Insomnia in Chronic Traumatic Brain Injury: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Non-Pharmacological Treatments for Insomnia in Chronic Traumatic Brain Injury (NCT03261674), a randomized na study. The primary outcome is Insomnia Severity Index (ISI).

2. Clinical Question

The registry record describes a randomized comparison of two behavioral interventions for insomnia among Veterans with chronic mild traumatic brain injury. The primary question is whether the groups differed in their mean change in Insomnia Severity Index (ISI) score from baseline at week 8 after treatment. The reported outcome is a change score, so the group means describe change over the specified period, rather than the groups’ average score at a single time point.

The registry record frames the trial as an assessment of the relative efficacy of the two non-pharmacological interventions. The reported comparison is a difference between groups, not a comparison with no treatment or with a medication. The record does not establish how participants were recruited, how the interventions were delivered, or which individuals were masked.

3. Trial Design

This was a randomized, parallel-group trial with single masking. The registry record lists CBT-I and ABT-I as behavioral interventions and describes the study as assessing their relative efficacy for insomnia in Veterans with chronic mild traumatic brain injury. The enrollment figure is 110. The record lists the trial as completed, with a start date of 2018-06-28 and a primary completion date of 2023-09-30.

The primary outcome was the ISI, measured as change from baseline at week 8 after treatment. For this analysis, the record states that only participants who completed the end-of-treatment ISI survey were included. The displayed group sizes are 47 for CBT-I and 39 for ABT-I. These are the numbers represented in the primary outcome summaries, not the randomized enrollment figure.

Randomization supports a between-group comparison, but the registry record does not report the sequence-generation method, allocation concealment, or the identity of the masked parties. It also does not report the original sample-size justification, the protocol’s planned analysis model, or whether interim analyses were planned. Those design details cannot be inferred from the observed results.

The sponsor is the VA Office of Research and Development. The record lists one outcome measure and no secondary outcomes posted. It lists no analyses in the registry. The result summaries include group means and standard deviations for the primary outcome and a mean-difference estimate with a confidence interval and unadjusted p-value. The record does not describe additional treatment components or the timing and number of therapy sessions, so the comparison should be understood as a comparison of the named assigned interventions rather than as evidence about specific treatment procedures.

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: Of the 110 participants who were randomized, only participants who completed the end of treatment Insomnia Severity Index survey were included in this analysis.

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 Insomnia Severity Index (ISI), measured over Change from baseline at week 8 after treatment.

The primary endpoint is the change in ISI score from baseline at week 8 after treatment. The ISI is a seven-item measure of a participant’s subjective experience of insomnia severity. Its total score ranges from 0 to 28, with higher scores indicating more severe symptoms. The registry description says scores above 15 indicate clinical insomnia and scores above 22 indicate severe symptoms.

A change score expresses how much the recorded score moved relative to baseline. In the posted results, both group mean changes are negative, indicating that the mean week 8 score was lower than the corresponding baseline score in each analyzed group. Since lower ISI scores indicate less severe symptoms, a negative change is in the direction of a lower score. This description concerns the measured score, not a clinical recommendation or a conclusion about treatment benefit.

The ISI is used because the registry describes it as a reliable subjective measure of insomnia severity and a sensitive measure of symptom change. It captures participants’ reported experience, rather than an objective sleep measure. The endpoint is a mean of a continuous measure, accompanied by standard deviations. The record does not report a threshold for a clinically important change or a separate responder outcome, so the group means alone do not show how many individuals crossed a particular threshold.

7. Results Posted in the Registry

• ABTI: mean -4.72 versus -9.70 on control; difference 4.98 (95% CI 3.08 to 6.88), standardized effect d = 1.12; z-test p = <0.001.

8. Statistical Interpretation

9. Statistical Methods Explained

The primary measure is continuous: each participant’s ISI score is numeric, and the registry reports a mean and standard deviation for each group. A group mean summarizes the average change, while the standard deviation describes how much individual changes vary around that group average. These summaries are not interchangeable: the mean describes the center of the observed values, and the standard deviation describes their spread.

What does the mean difference represent?

The comparison is expressed as ABT-I minus CBT-I. The reported group means are -4.72 for ABT-I and -9.7 for CBT-I, giving a posted mean difference of 4.98. A positive result in this ordering means the ABT-I mean is higher than the CBT-I mean. Because both group means are negative, the difference reflects a less negative average change in ABT-I. The direction depends on the order of subtraction; it should not be read without noting which group is subtracted from which.

A difference in group means describes an average contrast. It does not mean that every participant had a change equal to the group mean, or that every participant’s response followed the same direction. Individual outcomes can vary, as the group standard deviations indicate. Nor does the average contrast provide the proportion of participants whose scores improved, worsened, or crossed a clinical threshold.

How should the p-value be understood?

The unadjusted p-value is <0.001. In general, a p-value measures how incompatible the observed data are with a specified null hypothesis, under the assumptions of the statistical test. It is not the probability that the null hypothesis is true, and it does not measure the size or clinical importance of the estimated difference. The record does not describe the complete test specification or analysis model, so interpretation should stay with the posted label and result.

The word unadjusted is relevant because the record does not report a correction for multiple comparisons. It lists one outcome measure and zero analyses. A reader should not assume that the reported p-value accounts for a larger family of outcomes or analyses that may have been considered outside the registry record. Whether any such work occurred is not reported in the registry record.

Why do confidence intervals matter?

A confidence interval gives a range of values compatible with the observed data under a statistical procedure and its assumptions. The interval for the mean difference runs from 3.085 to 6.875. It describes uncertainty about the estimated group contrast, rather than a range containing a fixed proportion of individual participant changes. A narrower interval indicates more precision than a wider interval in the same measurement units. The interval’s interpretation depends on the analysis assumptions being appropriate.

For a continuous outcome, the variability of individual scores and the number of analyzed participants both affect the precision of a mean comparison. Greater variability tends to make a group contrast harder to estimate precisely, while additional observations can improve precision, provided the design and analysis are suitable. These are general statistical principles; the registry record does not supply a detailed model or explain any adjustments made in the analysis.

What does over-dispersion mean in general?

Over-dispersion is a general concern for outcomes whose variation is greater than expected under a particular probability model, often in count data. It can make standard errors too small if the analysis model fails to account for the extra variation. The primary outcome in this trial is a continuous ISI change score, and the registry record reports standard deviations. It does not report an over-dispersion assessment or a model for count data. The concept is therefore a general statistical point, not a reported feature of this outcome analysis.

The group summaries and comparison can be read as an unadjusted mean-based contrast. The registry record does not report a covariate-adjusted estimate, a repeated-measures model, a sensitivity analysis, or an analysis of all randomized participants. It is not possible to infer those methods from the displayed summary statistics. Where detailed model choices or diagnostic checks are important, they are not reported in the registry record.

10. Confidence Intervals and Precision

The estimated mean difference is 4.98 ISI units, with a 95% confidence interval from 3.085 to 6.875. In practical terms, the interval expresses uncertainty around the estimated average contrast, in the same units as the ISI change. It is not an interval for individual participants’ changes and does not mean that every plausible value within it is equally likely.

The interval lies above zero, so the estimates compatible with this interval all have the same direction under the reported ordering, ABT-I minus CBT-I. That direction corresponds to a less negative mean change for ABT-I than for CBT-I. The unadjusted p-value is <0.001, which is consistent with the interval excluding zero for this reported comparison. These are related descriptions of the evidence, but neither determines whether the size of the difference is important to patients.

Precision is affected by the amount of variation in scores and by the number of observations contributing to the estimate. The analyzed groups included 47 and 39 participants, respectively, and their standard deviations were 4.42 and 4.5. The registry record reports no confidence limits for the separate group means. The comparison interval is the available uncertainty summary for the mean difference.

Confidence intervals rely on the statistical procedure and assumptions used to construct them. The registry record does not report enough detail to evaluate all model assumptions or any adjustment strategy. It also does not report how missing endpoint measurements were handled beyond limiting this analysis to survey completers. Thus the interval quantifies uncertainty around the reported comparison, but does not capture every source of uncertainty about generalization, missingness, or analysis choices.

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 participants per arm, not total event counts. They are illustrative calculations based on the observed mean difference, not the trial’s original sample-size design. The listed assumptions are a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. Each planning scenario uses a stated fraction of the observed effect to detect.

For 100% of the observed effect, the calculation gives 13 participants per arm for 80% power and 17 participants per arm for 90% power. For 75% of the observed effect, it gives 23 participants per arm for 80% power and 30 participants per arm for 90% power. For 50% of the observed effect, it gives 51 participants per arm for 80% power and 68 participants per arm for 90% power.

In general, a power target is the probability of detecting an effect of a specified size under the assumptions used for a design calculation, if that effect is truly present. The 80% column describes planning for an 80% chance of detecting the stated effect; the 90% column describes planning for a 90% chance. Higher planned power requires more participants under otherwise unchanged assumptions. Smaller effects also need more participants to detect, which is why the figures increase as the assumed fraction of the observed effect decreases.

These numbers are planning illustrations, not observed power for the completed trial. The original design assumptions and sample-size justification are not reported in the registry record. The calculations do not show that the trial was designed using the observed result, and they should not be interpreted as a retrospective assessment of whether the trial was adequately powered.

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 = 4.980; pooled standard deviation σ = 4.456.
2. n = 2 × (1.96 + 0.842)² × 4.456² / 4.980² = 12.6, rounded up to 13 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 <- 4.98                      # observed difference in means
sigma <- 4.45636836982975                     # 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(3261674)
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 = 13 (80% power), 17 (90% power)
 75% of observed effect: n per arm = 23 (80% power), 30 (90% power)
 50% of observed effect: n per arm = 51 (80% power), 68 (90% power)
power.t.test: 13.6 
Simulated power at n = 13 per arm: 0.779 (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 lists one outcome measure, the ISI, and no secondary outcomes posted. It lists zero analyses. The primary result includes an unadjusted p-value. A multiplicity adjustment is therefore not reported in the registry record. The record does not establish whether additional outcomes, time points, subgroups, or analyses were considered outside the entries posted in the registry.

Multiplicity arises when several statistical tests are conducted: the chance of at least one apparently notable result can increase as the number of tests grows. In general, analysis plans may address this by defining a primary comparison in advance or adjusting inference across a specified family of tests. The registry record does not describe such a plan or identify a multiplicity procedure. No particular adjustment should be assumed.

The endpoint analysis included participants who completed the end-of-treatment ISI survey. The record states that 110 participants were randomized, while the outcome summaries include 47 participants in CBT-I and 39 in ABT-I. It does not explain why the other randomized participants are not included in these summaries. Reasons for missing surveys, timing of missingness, and whether missingness differed by treatment group are not reported in the registry record.

The record also does not report imputation, a missing-data model, sensitivity analyses, or an intention-to-treat analysis. Without this information, the results can be described as the reported analysis among survey completers, but the effect of missing outcome data on the comparison cannot be assessed. If completion were related to participants’ outcomes or treatment assignment, a completer-only estimate could differ from an estimate based on all randomized participants; whether that occurred is not reported in the registry record.

The treatment assignment is randomized, but the reported endpoint population is defined by completion of a survey after assignment. This distinction matters: randomization describes how groups were formed, while the analysis population describes who contributed outcome data. The record does not give sufficient detail to determine whether the analysis preserved all benefits of randomization or how excluded observations were handled.

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 5% reporting threshold.

The serious adverse event figures list 47 participants at risk in the CBT-I group and 39 in the ABT-I group. The count of participants affected by a serious adverse event is 0 in each group. The registry record also lists 0 deaths in each group. For other adverse events, the participants-at-risk counts are 47 for CBT-I and 39 for ABT-I, and the participants-affected counts are 0 in both groups.

These are counts in the registry’s reported groups. A count of zero means no affected participants were recorded in the relevant category in these entries; it does not prove that an event is impossible or that no event occurred outside the reported collection and follow-up. The registry record does not describe event definitions, the duration of safety follow-up, how events were collected, or whether all randomized participants contributed safety information.

The record lists no serious adverse event terms. The figures should therefore be read as the posted counts, not as a complete assessment of safety or a comparison establishing that one intervention is safer. The group sizes and the scope of reporting matter when interpreting zero counts.

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

The posted results describe mean ISI change among participants who completed the end-of-treatment survey. The average change was negative in both groups, and the between-group mean difference was 4.98 units in the direction of a less negative mean change for ABT-I than for CBT-I. The confidence interval and unadjusted p-value quantify the statistical uncertainty and evidence for that reported contrast under the analysis represented by the registry results.

The result does not show how every participant’s symptoms changed. Group means conceal individual variation, and the record does not provide responder counts or the distribution of changes. Although the registry description gives score ranges and symptom thresholds, no threshold for a clinically important change is reported. Statistical evidence for a difference is not, by itself, evidence that the difference is meaningful to patients.

The comparison is between two behavioral interventions, not between an intervention and no treatment. The result therefore does not establish the effect of either intervention relative to no care, usual care, or another comparator. The outcome is a subjective insomnia severity measure at week 8 after treatment; the registry record does not report longer-term outcomes or a broader set of clinical measures.

Generalization is also limited by the available description. The intended population is Veterans with chronic mild traumatic brain injury and insomnia, and the endpoint analysis includes only survey completers. The registry record does not provide demographic details, recruitment information, or the reasons for missing endpoint measurements. It is not possible to determine from the posted summary whether the observed estimate applies equally to other populations or settings.

Finally, the result should not be treated as a complete account of treatment effects or harms. The record lists no secondary outcomes, no analyses, and no detailed analysis plan. It provides counts for adverse events, but not the fuller collection and follow-up details needed to interpret those counts in depth.

16. Why This Trial Matters Statistically

Several statistical features are important for reading this comparison. First, the estimand is a between-group difference in mean change, rather than a within-group test of whether either group changed from baseline. Both groups have negative mean changes, but the primary comparative question is whether their average changes differ. The reported contrast makes that comparison directly.

Second, the outcome is a subjective continuous measure. A mean and standard deviation give a concise account of each group’s central tendency and variation, while the confidence interval expresses uncertainty about the contrast. Those summaries are useful, but they do not show the individual outcome distribution or the proportion of participants with a meaningful change.

Third, randomization and endpoint completion are distinct. Assignment was randomized, but the primary outcome analysis was limited to participants who completed the week 8 survey. The difference between randomized enrollment and analyzed group counts makes the analysis population especially important. The registry record does not explain missingness or report an analysis that includes all randomized participants.

Fourth, the reported p-value is unadjusted, and the registry record lists one outcome and no analyses. The absence of posted additional outcomes does not establish what was or was not explored elsewhere. The original statistical plan and sample-size rationale are not reported, so the observed estimate should not be used to reconstruct those design decisions.

Finally, the planning illustrations depend on the observed effect and stated assumptions. They help explain how required sample size changes with target effect and power, but they are not evidence of the trial’s original planning or its achieved power. Keeping the observed comparison, its precision, the analysis population, and the planning illustrations separate supports a more careful statistical reading.

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.

The primary outcome analysis is restricted to participants who completed the end-of-treatment ISI survey. While 110 participants were randomized, the group summaries include 47 participants assigned to CBT-I and 39 assigned to ABT-I. Reasons for the difference between randomized enrollment and the analyzed counts are not reported in the registry record. The consequences of missing endpoint data therefore cannot be evaluated from the posted results.

The registry record does not report a detailed statistical analysis plan, the original sample-size justification, a model description, covariate adjustments, or sensitivity analyses. It also does not report whether interim analyses were planned or conducted. The displayed comparison can be interpreted as the reported mean-based result, but the record does not permit an independent assessment of all analysis choices.

The record identifies single masking but does not identify who was masked. Since the primary endpoint is a subjective report, the absence of detail about masking limits assessment of the measurement context. The record does not describe intervention delivery, treatment adherence, or the number and timing of sessions, so those aspects cannot be considered when interpreting the observed group contrast.

The endpoint is measured at week 8 after treatment. Longer-term outcomes, additional clinical measures, and a threshold for clinically important change are not reported. The registry lists no secondary outcomes and no analyses. The adverse event entries provide counts but do not describe collection methods or the duration of safety follow-up. These limitations restrict interpretation to the reported endpoint and recorded safety counts.

18. Related Tutorials

19. Related Statistical Calculators

20. Sources

• ClinicalTrials.gov: Non-Pharmacological Treatments for Insomnia in Chronic Traumatic Brain Injury — NCT03261674.
• Registry record last updated 2024-10-31; status last verified 2024-10.
• Results first posted 2024-10-31.
• 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

NCT03261674 is a completed randomized, single-masked, parallel-group trial sponsored by the VA Office of Research and Development. It compares CBT-I with ABT-I for insomnia among Veterans with chronic mild traumatic brain injury. Enrollment was 110, and the primary outcome was change in ISI score from baseline at week 8 after treatment.

The endpoint analysis included 47 CBT-I participants and 39 ABT-I participants who completed the end-of-treatment survey. Mean changes were -9.7 and -4.72, respectively. The reported mean difference, ABT-I minus CBT-I, was 4.98, with a 95% confidence interval from 3.085 to 6.875 and an unadjusted p-value of <0.001.

The registry lists no secondary outcomes and no analyses. It does not report the original sample-size rationale, detailed analysis plan, or handling of missing endpoint data. The adverse event entries report zero affected participants in the serious and other adverse event categories for both groups, and zero deaths.