Brief Cognitive Behavioral Therapy for Military Populations: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Brief Cognitive Behavioral Therapy for Military Populations (NCT02038075), a randomized na study. The primary outcome is Estimated Percentage of Participants Making Suicide Attempt During 24-month Follow-up.

2. Clinical Question

The study’s broader purpose included comparing direct markers of suicidality, such as suicide attempts, and indirect markers such as suicidal ideation, intent, anxiety, depression, hopelessness, substance abuse and sleep disturbance. The record also describes secondary investigation of risk factors and warning signs following an index attempt. Those broader aims do not change the specific primary outcome posted for this comparison: estimated percentage making a suicide attempt during 24-month follow-up.

3. Trial Design

This was a randomized, double-blind, parallel-group study with a recorded enrollment of 152. The registry lists the study as completed and gives a start date of 2011-01 and a primary completion date of 2014-09. Its phase is recorded as NA. The sponsor was the University of Utah.

Participants were assigned to Brief Cognitive Behavioral Therapy (BCBT) or Treatment As Usual (TAU). The posted arm counts are 76 for BCBT and 76 for TAU. BCBT was delivered in addition to TAU, so the comparison concerns adding the brief therapy program to usual care rather than replacing usual care. The TAU description says care could involve military clinicians and clinicians from the local community, with the care provider determined by the participant’s primary mental health care provider.

The registry describes BCBT as 12 outpatient individual psychotherapy sessions, scheduled weekly or biweekly. It reports a 90-minute first session and 60-minute subsequent sessions. Although the study is marked double-blind, the identities of the people who were masked and the practical masking procedures are not reported in the registry record. The allocation sequence and concealment procedures are also not reported in the registry record.

The primary outcome concerns suicide attempts during 24-month follow-up. Other outcomes listed include the Scale for Suicide Ideation, Beck Depression Inventory, Second Edition, Beck Hopelessness Scale and Beck Anxiety Inventory. The record lists 10 outcome measures in total, but the available description does not establish how every outcome was timed or analyzed. Details of the full analysis protocol, including prespecified model specifications, are not reported in the registry record.

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: 30 of 206 referred Soldiers were excluded (15 due to planned move in next 6 months, 15 refused). 24 of 176 consenting to participate were excluded (22 ineligible, 2 dropped before randomization).

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: intent to treat

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 Estimated Percentage of Participants Making Suicide Attempt During 24-month Follow-up, measured over 24 months.

The primary endpoint is titled “Estimated Percentage of Participants Making Suicide Attempt During 24-month Follow-up.” It expresses the outcome as an estimated percentage and is described for the intent-to-treat population. The posted arm estimates are 13.8% for BCBT and 40.2% for TAU, with 76 participants listed in each arm.

The endpoint description refers to the Suicide Attempt Self-Injury Interview (SASII), a clinician-administered interview for assessing factors involved in nonfatal suicide attempts and intentional self-injury. It covers features such as method, lethality, impulsivity, likelihood of rescue, suicidal intent, consequences and habitual self-injury. The registry description reports interrater reliability ranges from .87-.98 across items and a .92 correlation for rater classification of behavior as a suicide attempt or non-suicidal self-injury. It also describes high agreement in identifying and classifying suicide-related events against clinician therapy notes, patient diary cards and medical records for events requiring medical attention.

Using an interview-based endpoint can provide a structured way to characterize suicide-related events beyond a simple count. The posted percentage gives an accessible summary of the estimated proportion experiencing an attempt over follow-up. However, the registry record does not provide the estimation details behind the arm percentages, such as their uncertainty intervals, or explain how the estimates were derived. A separate Cox proportional hazards result is reported and addresses the timing of events through a time-to-event model.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The main statistical result is a Cox proportional hazards analysis. In general, survival analysis is used when the outcome is the time until an event, especially when participants can have different lengths of observation or some do not experience the event during follow-up. It uses both whether an event occurred and when it occurred, rather than reducing follow-up to a simple yes-or-no status alone.

A Cox model estimates a hazard ratio comparing groups. The hazard is the instantaneous event rate among participants who have not yet experienced the event. A hazard ratio of 1 would represent equal hazards under the model; values below or above 1 indicate the direction of the estimated relative hazard, depending on the group coding. The registry reports a Cox proportional hazard of .38 for the BCBT-versus-TAU comparison, with a 95% confidence interval of .16 to .87 and p = .02. The record does not describe the covariates, stratification, or other model details, so these should not be assumed.

The proportional hazards model assumes that the relative hazard between groups is stable over time. This is a general model assumption, not a reported finding about this trial. Whether the assumption was assessed, and what diagnostics were used, are not reported in the registry record. If the relative hazard changes over follow-up, a single hazard ratio may not fully characterize the pattern of treatment-group differences.

A confidence interval describes a range of estimates compatible with the observed data under the model and its assumptions, using the stated confidence level. It communicates precision as well as direction. Here the reported 95% confidence interval extends from .16 to .87. The interval does not include 1, the value representing equal hazards, but it still spans a range of possible relative hazard estimates. The p-value, .02, is another way of expressing evidence against a null hypothesis under the analysis assumptions. Neither the interval nor the p-value addresses whether the model assumptions hold or whether the endpoint captures every aspect of clinical experience.

The arm results are expressed as estimated percentages: 13.8% for BCBT and 40.2% for TAU. Such percentages are easy to interpret as summaries of the proportion estimated to experience an attempt during follow-up. They do not show when events occurred. The record does not give the underlying event counts or uncertainty intervals for these estimates, and it does not describe the exact estimation procedure. Therefore, the percentages cannot independently reproduce the Cox result.

Count outcomes can sometimes be analyzed with models that account for differences in exposure time. In general, count data may be over-dispersed when their variability exceeds what a simple Poisson model expects; ignoring over-dispersion can make uncertainty estimates too narrow. That general issue is not evidence that this trial’s data were over-dispersed. The reported primary analysis is Cox regression, and the registry record does not report a count model or a dispersion assessment.

The record’s sample-size illustration is described as planning based on the observed effect, with two-sided significance level 0.05, equal allocation, and a normal approximation; the effect to detect is specified as a fraction of the observed effect. For time-to-event planning, the illustration is described as using a total number of events according to Schoenfeld; other illustrations are expressed as participants per arm. These are planning illustrations, not observed trial results and not observed power. Their numeric values are not reported in the registry record used for this explanation. The trial’s own sample-size justification and assumptions are not reported in the registry record.

Randomization is intended, in general, to support a fair comparison by distributing prognostic factors across groups on average. The registry identifies the study as randomized, but does not report the sequence-generation method, allocation concealment, masking procedures, or whether the analysis adjusted for baseline characteristics. These details matter when judging the risk of bias and the reproducibility of the analysis.

The record lists 10 outcome measures and four named secondary measures, while only one analysis is listed for the primary comparison in the available results. When many outcomes or analyses are considered, the chance of a nominally unusual result can increase; multiplicity procedures can address that issue. Whether the trial used any such procedure, and which outcomes were designated in advance as primary or secondary beyond the posted information, are not reported in the registry record. The p-value should therefore be read as the reported result for the stated comparison, not as a correction for all possible outcomes.

10. Analyses Reported by the Sponsor

The registry reports the sponsor’s Cox regression result as a hazard ratio of .38, with a 95% confidence interval of .16 to .87 and p = .02. It also posts estimated percentages of 13.8% for BCBT and 40.2% for TAU. These are the reported results; the record does not include enough event-level or model detail to independently reconstruct the Cox analysis.

A simple unadjusted recalculation is not available for the posted primary endpoint. The endpoint is a count-type outcome, and the record indicates that values were missing or not estimable for a two-arm comparison of that form. Therefore, no recalculated comparison should be substituted for the sponsor’s time-to-event analysis. The posted percentages can be described as arm estimates, but they do not supply the counts and uncertainty information needed to reproduce the reported model result.

No discrepancy is flagged between a recalculation and the sponsor’s result. This does not establish that every modeling choice can be checked from the registry record; the Cox model specifications and diagnostics are not reported. The sponsor’s result should be interpreted as the analysis documented in the registry, with the limits of the available methodological detail in mind.

11. Multiplicity, Missing Data and Analysis Populations

The record lists 10 outcome measures and names four secondary measures: the Scale for Suicide Ideation, Beck Depression Inventory, Second Edition, Beck Hopelessness Scale and Beck Anxiety Inventory. The study description also refers to several other direct and indirect markers, including suicide attempts, suicidal ideation, intent, anxiety, depression, hopelessness, substance abuse, sleep disturbance and remission of psychiatric diagnoses. The registry record does not explain how the full outcome family was organized or which endpoints were formally designated as primary or secondary in the analysis plan.

When a study examines multiple outcomes or makes multiple statistical comparisons, the chance of at least one result appearing statistically notable by chance can increase. Statistical plans may address this through a prespecified hierarchy or adjustment procedure. Whether multiplicity was addressed, and by what method, is not reported in the registry record. The reported p = .02 should be understood as the value for the listed Cox comparison; it is not identified as adjusted for the other outcomes.

The primary outcome is described for the intent-to-treat population, and each arm has 76 participants listed. The record does not report how many participants had complete follow-up, how missing outcome information was handled, whether censoring occurred, or whether sensitivity analyses were conducted. It also does not report adherence to the assigned intervention, deviations from allocation, or details of any per-protocol analysis. These omissions matter because missing data and post-randomization differences can affect both estimated event proportions and time-to-event comparisons.

Interim analyses, stopping rules, and the prespecified statistical analysis plan are not reported in the registry record. Nor are the procedures for assessing proportional hazards or selecting covariates. These missing details do not change the posted estimates, but they limit assessment of whether the reported analysis followed a prespecified plan and how robust the result may be to analytic choices.

12. Safety Results

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

Other (non-serious) adverse events

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

Most frequent serious adverse event terms

Serious events with the highest arm-level rate: suicide death (1.3%).

The serious adverse event counts list 1 affected participant among 76 at risk in the BCBT group and 1 among 76 at risk in the TAU group. For other adverse events, the record lists 0 affected participants among 76 at risk in each arm. The record also lists “Suicide death” among the top serious adverse event terms, with a percentage of 1.3%.

The registry presentation does not clearly connect the “Suicide death” term and its 1.3% figure to a specific arm in the information available for this explanation. The group-level death fields are null, so a death count by group is not reported in the registry record. The term-level listing should not be treated as sufficient to establish arm-specific counts or to resolve how the percentage was calculated.

These safety figures are descriptive. The small number of reported serious events provides limited information about differences between groups, and the registry record does not provide a statistical comparison of adverse events. The posted counts should be read as reported event information, not as proof of an absence or presence of a safety difference.

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

The registry reports estimated suicide-attempt percentages of 13.8% for BCBT and 40.2% for TAU during 24-month follow-up, along with a Cox proportional hazards estimate of .38, 95% confidence interval .16 to .87, and p = .02. In direction, the model result favors a lower estimated hazard in the BCBT group relative to TAU, subject to the analysis coding and assumptions. The confidence interval conveys uncertainty around that relative hazard estimate.

These results concern the specific primary endpoint and follow-up period described in the registry. They do not establish effects on every outcome in the study description, including suicidal ideation, depression, anxiety, hopelessness, substance use, sleep disturbance or psychiatric diagnosis remission. Although several secondary measures are listed, detailed results for them are not included in the available record.

The result does not by itself establish the mechanism behind the observed group estimates, the durability of any differences beyond 24-month follow-up, or the effectiveness of BCBT in settings and populations not described by this trial. The record identifies active duty military personnel and a comparison of BCBT added to usual care against usual care alone. Generalizing beyond that context requires evidence not contained in the registry results.

Statistical evidence is not the same as certainty. The hazard ratio is an estimate, and the confidence interval indicates that its precision is limited. The registry does not report the full event data, missing-data handling, model diagnostics, or multiplicity procedures needed for a complete independent assessment. These limitations make the posted result informative but not exhaustive.

The study’s randomized design supports a comparative question, but details of allocation, masking, adherence and analysis implementation are not reported in the registry record. The findings should therefore be read as the posted results of this trial rather than as a guarantee of outcomes for an individual or a recommendation about care.

14. Why This Trial Matters Statistically

This trial is statistically useful because it reports both estimated arm percentages and a time-to-event model for suicide attempts. The percentages give a direct summary of the estimated proportion experiencing the endpoint, while the Cox result incorporates event timing. Seeing these different measures together helps illustrate why a proportion and a hazard ratio answer related but distinct questions.

The design also makes the comparison explicit: BCBT was added to usual care, while the control group received usual care. That distinction is important when interpreting the estimand. The result concerns the comparison of an added psychotherapy program with usual care alone, not BCBT in isolation or a comparison with no care.

At the same time, the record demonstrates why statistical interpretation needs more than a p-value. The Cox estimate and confidence interval provide information about direction and precision, but key details about the model, missing follow-up, proportional hazards assessment and multiplicity are not reported in the registry record. Transparent reporting of these details is important for judging reproducibility and the scope of the conclusions.

The endpoint is also a consequential event measured over extended follow-up. Time-to-event analysis can retain information about when events occur, while the registry’s estimated percentages provide a more readily understood summary. The difference between those summaries is a useful reminder that the choice of estimand and model affects what a result means.

Finally, the record lists several secondary psychological measures and broader study aims, but the primary posted result addresses suicide attempts. Keeping the endpoint-specific result separate from broader claims is central to sound interpretation of clinical trial evidence.

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.

The registry provides the primary arm estimates and Cox result but not the underlying number of suicide attempts in each group, the confidence intervals around the estimated percentages, or enough detail to reproduce the time-to-event model. The Cox covariates, handling of ties, censoring details and model diagnostics are not reported in the registry record. In particular, assessment of the proportional hazards assumption is not described.

Although the study is labeled double-blind, who was masked and how masking was implemented are not reported. The randomization sequence and allocation concealment procedures are also not described. The record does not report adherence, treatment crossover, intervention fidelity or the amount of usual care received in either group. These details are relevant to understanding the contrast between assigned groups and the implementation of BCBT.

The primary endpoint is an estimated percentage over 24-month follow-up, but the record does not describe the exact estimation method, the status of participants lost to follow-up, or how missing outcome data were handled. It also does not report sensitivity analyses. Without these details, the robustness of the estimates to incomplete outcome information cannot be assessed from the registry record.

The study lists 10 outcome measures and four named secondary measures, but the complete prespecified outcome hierarchy and any multiplicity procedure are not reported. The statistical analysis plan, interim analyses and sample-size justification for the trial itself are not reported in the registry record. A planning illustration based on the observed effect is not a substitute for the trial’s original design justification.

The available serious adverse event reporting gives one affected participant in each group, but the death fields are null and the serious-event term listing does not clearly assign its percentage to an arm. This restricts arm-specific interpretation of the serious event information. Finally, findings from this active duty military population and this specific comparison should not automatically be generalized to other populations or care arrangements.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: Brief Cognitive Behavioral Therapy for Military Populations — NCT02038075.
• Registry record last updated 2023-12-12; status last verified 2023-11.
• Results first posted 2016-05-30.
• 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.

19. Record Summary

NCT02038075, “Brief Cognitive Behavioral Therapy for Military Populations,” is listed as a completed randomized, double-blind, parallel-group study sponsored by the University of Utah. The registry reports enrollment of 152, with 76 participants in each group. BCBT was added to TAU; the comparator received TAU alone.

The primary outcome was the estimated percentage making a suicide attempt during 24-month follow-up in the intent-to-treat population. The posted estimates are 13.8% for BCBT and 40.2% for TAU. A Cox proportional hazards analysis reported a hazard ratio of .38, with a 95% confidence interval of .16 to .87 and p = .02.

The record lists four secondary measures and serious adverse event information for both groups. Important details—including missing-data handling, model diagnostics, multiplicity procedures and the trial’s own sample-size justification—are not reported in the registry record.