Studying the Modification of Attention Bias Remotely After Trauma: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Studying the Modification of Attention Bias Remotely After Trauma (NCT04888169), a randomized na study. The primary outcome is Change in PTSD Checklist-5 (PCL) Score (Before/After Training).

2. Clinical Question

3. Trial Design

The registry classifies this as a randomized, double-masked, parallel-group trial. The study is marked completed, with a reported enrollment of 872. The brief summary describes a planned remote trial with four broad conditions: attention bias modification, attention control training, placebo neutral attention training, and a final control condition involving daily questions. The intervention details and outcome groups show word- and face-stimulus versions for the training conditions. These outcome groups should be distinguished from the brief summary’s broader description.

The primary outcome is change in PTSD Checklist-5 (PCL-5) score from baseline to post-training. Its timeframe is listed as baseline and 4 weeks, while the outcome description names Day 1 and Day 28. The registry’s brief summary also describes assessments at baseline, after one and two weeks of training, and at eight-week follow-up. The relationship between those planned assessment points and the posted primary endpoint is not further explained in the registry record.

The brief summary states that investigators planned to randomize 1,897 participants and aimed for 1,232 completers, described as 176 per condition. The reported enrollment is 872, while the posted primary-outcome groups contain smaller group counts. The reasons for the difference between the planned totals and posted outcome counts are not reported in the registry record. The trial’s original sample-size justification and design assumptions are not reported in the registry record.

The listed dates are a start date of 2023-04-20 and primary completion on 2025-06-30. A study phase is listed as NA. Sponsor information identifies the University of California, San Francisco. These record details describe the registered study; they do not establish how closely the eventual conduct matched every element of the planned design.

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.

Only the first 6 of 7 groups are shown.

Baseline - Post-Training

Post-Training - Follow-Up

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: Baseline measures (i.e., age, gender, race/ethnicity) were collected for all participants who completed screening surveys.

Only the first 6 of 8 groups are shown.

6. Primary Endpoint

The registry lists the primary outcome as Change in PTSD Checklist-5 (PCL) Score (Before/After Training), measured over Baseline, 4 weeks.

The primary endpoint is the mean change in PTSD Checklist-5 score between baseline and post-training. The PCL-5 is a self-report measure of the presence and severity of 20 symptoms associated with post-traumatic stress disorder. Each item is scored from 0, “Not at All,” to 4, “Extremely,” and the total score ranges from 0 to 80. The registry states that scores of 33 and above indicate clinically significant PTSD symptoms, and that higher scores indicate worse symptom severity. It also describes a 10 point change, up or down, as clinically significant.

A continuous score preserves information across the scale rather than reducing participants to a symptom threshold alone. Comparing mean change can describe the average direction and size of score movement within each group. It does not show how many individual participants experienced a change of a particular size, nor does it show the distribution of changes around the mean. The posted outcome provides a mean and standard deviation for each group, but individual-level outcome distributions are not reported in the registry record.

The endpoint description identifies baseline as Day 1 and post-training as Day 28. The primary outcome timeframe labels the visits as baseline and 4 weeks. The registry does not state the subtraction convention for the change score; thus, interpreting the sign as a specific subtraction order requires caution. In any case, a change score describes a difference over time, not a comparison of final scores alone.

7. Results Posted in the Registry

• Attention Bias Modification - Word Stimuli: mean -4.54 versus -4.65 on control; difference 0.11 (95% CI -4.83 to 5.05), standardized effect d = 0.01; z-test p = 0.965.
• Attention Bias Modification - Face Stimuli: mean -5.37 versus -4.65 on control; difference -0.72 (95% CI -5.56 to 4.12), standardized effect d = -0.05; z-test p = 0.771.
• Attention Control Training - Face Stimuli: mean -6.73 versus -4.65 on control; difference -2.08 (95% CI -6.37 to 2.21), standardized effect d = -0.16; z-test p = 0.341.
• Placebo Attention Training - Word Stimuli: mean -3.81 versus -4.65 on control; difference 0.84 (95% CI -3.68 to 5.36), standardized effect d = 0.06; z-test p = 0.716.

8. Statistical Interpretation

9. Statistical Methods Explained

The primary outcome is continuous: it is a score change on the PCL-5, summarized by a mean and standard deviation for each group. A mean gives the arithmetic average of the recorded changes, while a standard deviation describes how spread out individual recorded changes are around that average. A group mean alone does not reveal whether most participants had similar changes or whether the mean reflects a range of individual experiences. The registry record does not provide participant-level observations or a fuller distributional summary.

What does the mean difference compare?

The posted pairwise estimates compare each listed group’s mean change with the Attention Control Training - Word Stimuli group. The difference is expressed in PCL-5 score points. A value of zero would mean equal group means for the recorded change; a negative or positive estimate indicates the direction of the numerical difference under the comparison’s stated ordering. For example, the estimate for Attention Control Training - Face Stimuli is -2.08, while the estimate for Control - Questions is 1.99. These are differences between group averages, not changes within a single participant.

The comparator group itself has a mean change of -4.65 and a standard deviation of 15.1 among 72 participants. The estimates use the posted group means and standard deviations. The registry record does not describe covariate adjustment, stratification, repeated-measures modeling, or another statistical model for the reported primary outcome. It also does not provide the sponsor’s analysis code or detailed analysis specification.

How should the p-values be read?

A p-value summarizes how compatible the observed data are with a specified null model in which the compared group means have no difference, subject to the assumptions of the test. It is not the probability that the null hypothesis is true, and it does not describe the magnitude or clinical importance of a difference. The posted values are explicitly unadjusted. They do not account for the set of multiple comparisons unless an adjustment is described, and no such adjustment is reported in the registry record.

The estimates and their confidence intervals give more information about magnitude and precision than a p-value alone. An interval spanning zero is compatible with no mean difference as well as with differences in either direction within its limits, under the analysis assumptions. All six posted intervals span zero. That does not demonstrate equivalence: an equivalence conclusion requires a prespecified margin and an analysis designed to assess whether the interval is sufficiently narrow within that margin. Such a margin is not reported in the registry record.

Why do spread and sample size matter?

Standard deviations in the outcome groups range from 11.71 to 15.1. The spread is substantial relative to the point estimates of the between-group differences. When participant outcomes vary considerably, a mean difference is estimated less precisely unless the sample provides enough information to distinguish the average difference from that variability. Unequal group sizes also affect precision. The reported groups have different counts, from 71 to 87, rather than identical numbers in every arm.

The posted analyses are based on summary statistics and are described as simple, unadjusted recalculations. They do not account for individual-level data, covariates, within-person correlations across visits, missing-outcome mechanisms, or any clustering or design features that might be relevant. Whether such features applied or were considered in the sponsor’s analysis is not reported in the registry record. If outcomes are skewed or contain influential values, mean-based summaries may also be affected; distributional diagnostics are not reported in the registry record.

What is not documented about the analysis?

The registry record lists zero analyses and two outcome measures, with no posted secondary outcomes. It does not report the original analysis plan, the trial’s multiplicity strategy, an interim-analysis plan, the approach to missing data, or the analysis population used for the sponsor’s own primary analysis. The exact definition of the analysis set for each outcome group is not reported in the registry record. These omissions limit assessment of whether the posted comparisons reflect the prespecified primary analysis or a particular subset of the randomized participants.

The displayed mean differences and confidence intervals are useful descriptive comparisons, but they should not be mistaken for a full account of the trial’s statistical methods. The registry record does not report whether the sponsor’s primary analysis used a different model or adjustment. No discrepancy is flagged in the posted record, but absence of a discrepancy flag does not supply the missing analysis details.

10. Confidence Intervals and Precision

A confidence interval gives a range of values compatible with the observed data and the statistical procedure used, under its assumptions. It communicates uncertainty around an estimated mean difference; it is not a guarantee that the true difference lies within the particular interval, nor does it show the range of individual participants’ changes.

The posted intervals are broad relative to several of the point estimates. For attention bias modification with word stimuli, the estimated difference is 0.11, with a confidence interval from -4.833 to 5.053. For attention bias modification with face stimuli, the estimate is -0.72 and the interval runs from -5.559 to 4.119. The interval for attention control training with face stimuli is -6.365 to 2.205. These ranges include zero and allow for differences in either direction.

The placebo word-stimuli comparison has an estimate of 0.84 and a confidence interval from -3.682 to 5.362; the placebo face-stimuli comparison has an estimate of -0.14 and a confidence interval from -4.593 to 4.313. For the questions-only control, the estimate is 1.99 and the interval is -2.439 to 6.419. Each interval spans zero. The data therefore leave uncertainty about the direction and size of the mean difference for each comparison.

Precision depends on both outcome variability and the number of participants contributing data. The observed group standard deviations range from 11.71 to 15.1, and group counts range from 71 to 87. These summaries help explain why a small point estimate should not be interpreted as proof that the groups are equivalent. A narrow interval can rule out a wider range of effects more readily than a broad one; the posted intervals do not establish a prespecified equivalence or noninferiority conclusion.

The intervals are unadjusted pairwise results. The registry record does not report a multiplicity adjustment, a sponsor model, or details of how missing observations were handled. Those factors could affect interpretation of inferential precision. Confidence intervals describe uncertainty in these comparisons under the calculations used, not uncertainty about every possible analysis or all enrolled participants.

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 sample-size displays are planning illustrations based on the observed effect for each comparison, not the trial’s original design calculation. They give participants per arm for 80% power and participants per arm for 90% power. These columns represent different chances of detecting the specified effect if it were the true effect and if the planning assumptions held; they do not report the number enrolled in the actual trial.

The planning assumptions are a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. The effect used for each illustration is the stated fraction of the observed effect. The rows describe planning for 100%, 75%, or 50% of the observed effect. The trial’s own original design assumptions and sample-size justification are not reported in the registry record.

Smaller effects require more participants to distinguish them from random variation at the same significance level and target power. This is why the illustrations increase as the assumed effect is reduced from 100% to 75% and then 50% of the observed effect. The figures can become very large when the observed difference is close to zero; they should be understood as consequences of the chosen effect assumption, not as recommended recruitment targets.

The trial’s observed group counts and reported enrollment are separate from these illustrations. The sample-size figures are not observed power and do not indicate the probability that the trial’s results are correct. They are hypothetical planning calculations based on the posted estimates and specified 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 = 0.110; pooled standard deviation σ = 15.080.
2. n = 2 × (1.96 + 0.842)² × 15.080² / 0.110² = 295,028.5, rounded up to 295,029 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 <- 0.11                      # observed difference in means
sigma <- 15.0801551057734                     # 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(4888169)
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 = 295,029 (80% power), 394,960 (90% power)
 75% of observed effect: n per arm = 524,496 (80% power), 702,151 (90% power)
 50% of observed effect: n per arm = 1,180,114 (80% power), 1,579,838 (90% power)
power.t.test: 295029.4 
Simulated power at n = 295,029 per arm: 0.803 (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 study includes multiple training conditions and the posted results contain six pairwise comparisons against Attention Control Training - Word Stimuli. The registry record also identifies two outcome measures, although no secondary outcomes are posted. When several comparisons or outcomes are examined, the chance of at least one nominally unusual result can increase if no adjustment is made. A multiplicity strategy can address that issue, but an adjustment method or hierarchy of outcomes is not reported in the registry record.

The p-values accompanying the displayed comparisons are unadjusted. The registry record does not say whether the sponsor used a different multiplicity procedure in its own analysis, nor does it provide a statistical analysis plan. The six comparisons should therefore be interpreted as separate unadjusted estimates rather than as a family-wise adjusted set of confirmatory tests. The record also does not report whether there were other planned comparisons that were not posted.

Missing-data handling is another important unknown. The posted outcome counts range from 71 to 87 across groups, while the overall enrollment is reported as 872 and the brief summary describes larger planned randomization and completion totals. The registry record does not explain these differences or report how many randomized participants lacked a Day 28 outcome. It also does not state whether missing values were imputed, whether analyses used available cases, or whether sensitivity analyses assessed assumptions about missingness.

The analysis population is not described in sufficient detail to determine whether the outcome summaries use all randomized participants, participants who completed training, or another set. The exact population definition is not reported in the registry record. Nor is the handling of protocol deviations, withdrawals, or incomplete training specified.

These gaps matter because missing outcomes and multiple testing affect how confidently results can be generalized from the displayed comparisons. They do not change the arithmetic of the posted means, but they limit evaluation of the confirmatory status and robustness of the inferential conclusions.

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.

The registry reports serious adverse events for each listed group. The serious-event counts are zero in every group, and the number at risk ranges from 123 to 127. The registry also reports zero other adverse events in every group, with the same group-specific at-risk counts. Deaths are reported as zero in each group. No serious adverse-event terms are listed.

These figures indicate that no serious adverse events or other adverse events were recorded in the registry’s safety results for the reported groups. They are event counts, not estimates of the probability of an event, and they do not establish that no events occurred outside the registry’s reporting scope. The safety at-risk counts differ from the group counts shown for the primary outcome. The registry record does not explain the difference or provide further information about safety-event definitions, collection procedures, or follow-up completeness.

Zero reported events should be read as an absence of reported events in these results, not as proof that the interventions have no risks. The record’s safety reporting does not describe the clinical seriousness criteria or provide event-level details beyond the counts.

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

The posted outcome describes mean changes in PCL-5 score and compares six groups with Attention Control Training - Word Stimuli. The estimated mean differences are modest in numerical size relative to the reported standard deviations, and each confidence interval spans zero. The unadjusted p-values do not provide clear evidence of a difference from the reference group in these comparisons. At the same time, the intervals allow for differences in either direction, so the results do not establish that the groups are equivalent.

The findings concern a particular continuous symptom score measured from baseline to post-training. They do not, by themselves, indicate how many participants crossed the registry’s threshold of 33, experienced a change of 10 points, or had a change considered meaningful by another criterion. Group means do not show individual response patterns. The registry record does not report the number of participants achieving a clinically significant change, or the distribution of individual changes.

The comparisons are between specific app conditions, including word- and face-stimulus variants, placebo training, and a questions-only condition. They should not be collapsed into a general conclusion about all remote attention training or all possible approaches to post-traumatic stress symptoms. The outcome record also does not establish whether the observed group means represent every randomized participant, since the analysis population and missing-data approach are not reported.

The registry marks the study completed and reports enrollment of 872. The brief summary describes higher planned randomization and completion totals, while the primary outcome groups have smaller counts. Without an account of participant flow and outcome availability, it is not possible to determine from the registry record why these numbers differ or how that affects the estimates.

Accordingly, the posted results are best read as imprecise, unadjusted comparisons of average recorded score changes in the groups with outcome data. They neither establish equivalence nor support a broad causal conclusion beyond the comparisons and outcome reported.

16. Why This Trial Matters Statistically

This study is statistically informative because it places several related attention-training formats alongside placebo training and a questions-only control in a randomized, parallel-group design. The outcome groups separate word and face stimuli for the training conditions, allowing the posted comparisons to be read as specific contrasts rather than as one pooled estimate for a broad intervention category.

The results also illustrate why effect estimates and confidence intervals matter alongside p-values. Several estimates are close to zero, but their intervals span zero and include differences in either direction. A nonsignificant result is not evidence that groups are identical; assessing equivalence would require a defined margin and a suitably designed analysis, neither of which is reported in the registry record.

The trial’s planned enrollment and completer targets in the brief summary differ from the reported enrollment and the smaller outcome-group counts. That makes transparent participant flow and analysis-population definitions especially important for understanding precision and generalizability. The registry record does not explain these counts or specify how missing outcome data were handled.

The study is also a useful example of the distinction between a planning illustration and the original design. The sample-size figures based on observed effects are hypothetical calculations under stated assumptions; they are not evidence of the trial’s planned power or of its ability to detect a meaningful effect. The trial’s own sample-size justification and analysis plan are not reported in the registry record.

Overall, the posted statistics show the limits of drawing strong conclusions from group means when uncertainty remains broad and key analysis details are unavailable. They highlight the value of prespecified comparisons, clear missing-data reporting, and reporting of individual-level outcome patterns alongside group summaries.

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 registry record has limited detail about participant flow. Although enrollment is reported as 872, the primary-outcome group counts range from 71 to 87, and the brief summary describes planned randomization of 1,897 participants and 1,232 completers. The record does not explain why these figures differ, whether the posted groups represent all randomized participants with available outcome data, or how many participants completed the training.

The trial’s detailed statistical analysis plan is not reported in the registry record. This includes the prespecified model, adjustment variables, multiplicity approach, missing-data methods, and analysis population. The posted pairwise results are unadjusted recalculations from summary statistics and may not capture features of a sponsor analysis that are not documented.

The registry does not report participant-level distributions, the proportion meeting a clinically significant change threshold, or results for the other planned assessment times described in the brief summary. The endpoint’s subtraction convention is also not specified. This limits interpretation of the sign of the change and assessment of whether average changes reflect broad shifts or a varied set of individual outcomes.

The primary outcome is self-reported, and the record does not provide details about adherence to training, completeness of remote assessments, or potential effects of missing responses. While double masking is listed, which participants or study personnel were masked and how masking was maintained are not reported in the registry record.

Finally, all six posted confidence intervals include zero and are not especially narrow relative to the observed group variation. These results leave uncertainty about the magnitude and direction of differences. The registry record does not report a prespecified equivalence margin, so a lack of clear evidence for a difference should not be taken as evidence that the conditions are equivalent.

18. Related Tutorials

19. Related Statistical Calculators

20. Sources

• ClinicalTrials.gov: Studying the Modification of Attention Bias Remotely After Trauma — NCT04888169.
• Registry record last updated 2026-10-30; status last verified 2026-10.
• Results first posted 2026-10-30.
• 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

NCT04888169 is a completed randomized, double-masked, parallel-group remote study sponsored by the University of California, San Francisco. The record lists enrollment of 872 and a primary outcome of mean change in PCL-5 score between baseline and post-training. The primary endpoint describes assessments on Day 1 and Day 28, with the timeframe also listed as baseline and 4 weeks.

Posted group means range from -6.73 to -2.66, with group counts from 71 to 87. Six unadjusted mean-difference comparisons use Attention Control Training - Word Stimuli as the reference group. Every confidence interval spans zero. The registry reports zero serious adverse events, zero other adverse events, and zero deaths in all listed groups.

The brief summary describes larger planned enrollment and completion totals than the reported enrollment and outcome-group counts. The registry record does not explain the differences or report the detailed analysis plan, missing-data methods, or analysis populations.