Group-Based Intervention to Improve Mental Health and Adherence Among Youth Living With HIV in Low Resource Settings: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Group-Based Intervention to Improve Mental Health and Adherence Among Youth Living With HIV in Low Resource Settings (NCT04024488), a randomized na study. The primary outcome is Group-level Mean General Anxiety Disorder-7 (GAD-7) at 6 Months.

2. Clinical Question

The central question is whether youth living with HIV in resource-limited settings who were assigned to a group-based trauma-informed cognitive behavioral therapy program had different group-level anxiety scores at 6 months from youth assigned to a discussion control. The registry describes a multi-site, individually randomized, controlled study of an Indigenous Leader Outreach Model. Youth sessions were led by Indigenous Youth Leaders, and the intervention was adapted to local context through focus groups and pilot testing.

The registry identifies the primary endpoint as the mean group-level General Anxiety Disorder-7 score at 6 months. It reports a mean of 4.51 in the youth TI-CBT arm and 4.29 in the youth discussion control arm. The posted comparison is a mean difference of 0.22, with a 95% confidence interval from -1.23 to 1.68 and p = 0.74 in the sponsor-reported analysis. The interval includes differences in either direction, so the point estimate alone does not settle whether the groups differed.

The registry record does not report the study locations, detailed eligibility criteria, or the exact content of the discussion sessions. The trial status is terminated. The record does not state the reason for termination.

3. Trial Design

The registry describes a randomized, parallel-group, open-label study with enrollment of 484. Its brief summary calls it a multi-site, two-arm, individually randomized controlled study. The listed youth interventions form the primary comparison: youth TI-CBT versus youth discussion control. Caregiver TI-CBT and caregiver discussion control are also listed as behavioral interventions, but the primary outcome concerns youth, and caregivers did not have GAD-7 scores collected for that outcome.

Youth in the TI-CBT arm were offered six 2-hour group sessions led by Indigenous Youth Leaders during weeks 1 to 8 and a 2-hour booster group session at 6 months. Youth in the discussion control arm were offered six 2-hour discussion group sessions, also led by Indigenous Youth Leaders during weeks 1 to 8, plus a 2-hour booster discussion session at 6 months. The registry says the intervention was delivered in a group format. Its arm description for the discussion control is incomplete, so additional details about that program are not reported in the registry record.

For caregivers, the listed programs involved two 2-hour group sessions on separate weeks within 8 weeks of the youth’s first group session, followed by a 2-hour booster session at 6 months. The caregiver TI-CBT sessions were led by adult study staff; the registry describes the caregiver discussion control as discussion sessions. The caregiver arm descriptions are incomplete in the registry record.

The primary endpoint uses groups as the analytic units because the registry identifies intracluster correlation among participants in the same group. This is important: people receiving sessions together may have more similar outcomes than people in different groups. Although randomization is described as individual, the primary analysis description says the group-level mean was the analytic unit.

The recorded start date is 2024-04-06, and the primary completion date is 2025-03-27. The study status is terminated. The reason for termination, numbers assigned to each arm at randomization, allocation concealment procedures, and any changes to the design are not reported in the registry record. Masking is listed as none.

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: No data for focus groups or pilot tests is reported because participant data was not collected by focus groups and the pilot consent form stated that collected data would not be analyzed. The data reported for all group-level measures reflect an aggregate of the 15 groups in each arm, instead of the individual participants in each arm.

Overall Study

Analysis population for the primary outcome, as defined in the registry: Youth participants who completed month 6 assessments. GAD-7 scores were not collected for caregivers. Primary outcomes were analyzed with groups as the analytic units due to the presence of intracluster correlation among participants in the same group.

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: Limited sociodemographic characteristics were collected from caregivers at baseline.

6. Primary Endpoint

The registry lists the primary outcome as Group-level Mean General Anxiety Disorder-7 (GAD-7) at 6 Months, measured over at 6 months.

The primary endpoint is the group-level mean General Anxiety Disorder-7 score at 6 months among youth who completed month 6 assessments. The GAD-7 asks about the frequency of anxiety symptoms over the past 2 weeks using 7 items. Each item is scored from 0, meaning not at all, to 3, meaning nearly every day; the item responses are summed to produce a score from 0 to 21, with higher scores indicating more severe symptoms. The registry notes that a score of 10 or more indicates moderate to severe anxiety symptoms.

A continuous score retains information across its range rather than reducing responses to a yes-or-no classification. The endpoint is therefore a comparison of mean scores, not a comparison of the proportions meeting the registry’s threshold. The posted group-level means are 4.51 in the youth TI-CBT arm and 4.29 in the discussion control arm, with standard deviations of 2.1 and 1.7, respectively.

Using a group-level mean is consistent with the registry’s stated concern that participants within the same group may have correlated outcomes. Treating those participants as fully independent could make the uncertainty appear smaller than it is. The record specifically says groups were the analytic units for primary outcomes because of intracluster correlation among participants in the same group.

The endpoint measures reported anxiety symptoms at the 6-month assessment. It does not, by itself, measure every aspect of mental health, adherence, or longer-term experience. The registry lists other mental health and ART adherence outcomes separately. Whether the GAD-7 data were complete for all eligible youth, how missing assessments were handled, and whether any adjustment was made for baseline scores are not reported in the registry record.

7. Results Posted in the Registry

• Youth TI-CBT Intervention Arm: mean 4.51 versus 4.29 on control; difference 0.22 (95% CI -0.25 to 0.69), standardized effect d = 0.12; z-test p = 0.361.

8. Secondary Efficacy Endpoints

The registry also posts results for secondary outcomes. For each outcome below, the endpoint type, the test applied to the posted summary statistics, the result and its interpretation are given in turn. Secondary endpoints are usually not adjusted for multiple testing, so they support rather than confirm the primary finding.

Group-level Mean General Anxiety Disorder-7 (GAD-7) at IPL (Immediately Post-Last Group Session)

• Endpoint type: Continuous: a mean value per participant, compared as a difference in means.
• Test: Unadjusted two-sample z-test for the difference in means, using the posted standard deviations.
• Result:
  • Youth TI-CBT Intervention Arm: mean 4.92 versus 4.63 on control; difference 0.29 (95% CI -0.13 to 0.71), standardized effect d = 0.17; z-test p = 0.176.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Youth TI-CBT Intervention Arm and the control. The interval gives the range of effects that remain compatible with the data.

Group-level Mean Patient Health Questionnaire-9 (PHQ-9) at IPL (Immediately Post Last Group Session)

• Endpoint type: Continuous: a mean value per participant, compared as a difference in means.
• Test: Unadjusted two-sample z-test for the difference in means, using the posted standard deviations.
• Result:
  • Youth TI-CBT Intervention Arm: mean 5.81 versus 6.13 on control; difference -0.32 (95% CI -0.89 to 0.25), standardized effect d = -0.14; z-test p = 0.270.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Youth TI-CBT Intervention Arm and the control. The interval gives the range of effects that remain compatible with the data.

Group-level Mean UCLA Post-Traumatic Stress Disorder-Reaction Index (UCLA PTSD-RI) at IPL (Immediately Post-Last Group Session)

• Endpoint type: Continuous: a mean value per participant, compared as a difference in means.
• Test: Unadjusted two-sample z-test for the difference in means, using the posted standard deviations.
• Result:
  • Youth TI-CBT Intervention Arm: mean 20.32 versus 19.00 on control; difference 1.32 (95% CI -0.96 to 3.60), standardized effect d = 0.16; z-test p = 0.257.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Youth TI-CBT Intervention Arm and the control. The interval gives the range of effects that remain compatible with the data.

The posted secondary outcomes extend the assessment beyond the 6-month primary anxiety endpoint. They include anxiety, depression, post-traumatic stress, a composite mental health measure, ART adherence, and viral load at the immediate post-last-group-session assessment and at 6 months. The reported set therefore addresses several related but distinct dimensions rather than repeating a single measure.

Across the secondary comparisons with results described in the registry record, the point estimates do not all point in the same direction. For the posted immediate post-session measures of anxiety, depression, and post-traumatic stress, sponsor-reported confidence intervals include zero. These outcomes do not provide a consistent, precise pattern that would override the uncertainty in the primary anxiety comparison. Their relationship to the primary finding is contextual: they offer additional outcome-specific information, not an independent confirmation of a single overall effect.

There are multiple secondary outcomes and assessment times. When many comparisons are examined, the chance of finding an apparently notable result by chance can increase. The registry record does not report a multiplicity adjustment or a hierarchy for interpreting secondary tests. For that reason, unadjusted secondary results generally carry less weight than a prespecified primary outcome and should be regarded as exploratory unless the analysis plan establishes otherwise.

The registry lists 12 outcome measures and 18 analyses, and identifies 8 secondary outcomes as posted. It does not provide a complete multiplicity strategy or enough detail to explain which analyses were planned as confirmatory. Some secondary outcome results are not included in the listed detailed summaries. Their numerical results and analysis methods are not reported in the registry record, so no conclusion about those outcomes can be drawn from their names alone.

In combination, the secondary outcomes show the breadth of outcomes tracked, but they do not resolve the precision limitations of the primary result. Their interpretation depends on the completeness of reporting, the analysis population, treatment of missing observations, and any adjustment for multiple comparisons—details that are not fully reported in the registry record.

Secondary outcomes registered for this trial, for reference (only those above were recalculated):

• Group-level Mean General Anxiety Disorder-7 (GAD-7) at IPL (Immediately Post-Last Group Session)
• Group-level Mean Patient Health Questionnaire-9 (PHQ-9) at IPL (Immediately Post Last Group Session)
• Group-level Mean UCLA Post-Traumatic Stress Disorder-Reaction Index (UCLA PTSD-RI) at IPL (Immediately Post-Last Group Session)
• Group-level Mean Composite Mental Health Measure at IPL (Immediately Post-Last Group Session)
• Group-level Mean ART Adherence at 6 Months
• Viral Load at 6 Months
• Group-level Mean ART Adherence at IPL (Immediately Post-Last Group Session).
• Viral Load at IPL (Immediately Post-Last Group Session)

9. Statistical Interpretation

10. Statistical Methods Explained

The primary outcome is continuous: it is a score with a mean and standard deviation in each arm. A mean difference expresses the separation between two arm averages on the score’s original scale. It is interpretable in GAD-7 score points, unlike a standardized effect that expresses the difference relative to score variability.

What does a confidence interval express?

A confidence interval describes uncertainty around an estimated difference under the assumptions of the analysis. A narrower interval indicates greater precision, while a wider interval indicates that a broader range of differences remains compatible with the data and method. The sponsor-reported 95% interval for the primary comparison runs from -1.23 to 1.68. Because it spans zero, the result is compatible with no difference as well as with differences in either direction. The interval is not a guarantee that the true value lies within its limits, and it does not by itself say whether any possible difference matters clinically.

How should the p-value be read?

A two-sided p-value summarizes how unusual the observed result, or a more extreme one, would be if the null hypothesis of no difference held and the analysis assumptions were appropriate. It is not the probability that the null hypothesis is true, and it does not measure the size or importance of a difference. The sponsor reports p = 0.74 for the first primary analysis. This is not evidence that the two arms are identical; it indicates that the reported data do not provide strong evidence against the null under that analysis. The confidence interval supplies important complementary information about the range of estimates supported by the analysis.

Why does group-level analysis matter?

The registry says primary outcomes were analyzed with groups as the analytic units because participants in the same group may have intracluster correlation. This means outcomes from members of a shared group may be more alike than outcomes from members of separate groups. If that dependence is ignored, a calculation that treats every person as an independent observation can understate uncertainty. The group-level approach recognizes the shared setting of the intervention and the potential dependence of observations.

The registry describes individual randomization, but also specifies group-level means as the units for primary analysis. These descriptions concern different aspects of the design: randomization assigns participants, while the chosen analysis unit determines how outcome summaries are compared. The full statistical model, any adjustments for clustering, and the number of groups included in the primary analysis are not reported in the registry record.

What is the recalculation?

The recalculated comparison is a simple, unadjusted calculation from posted arm-level summary statistics. It uses the posted means, standard deviations, and arm sample sizes to give a descriptive mean difference and an approximate uncertainty summary. It does not recreate the sponsor’s complete analysis or any unreported adjustments. The recalculated primary mean difference is 0.22, matching the first sponsor-reported point estimate, while the confidence interval and p-value differ. This distinction is important because matching point estimates do not imply that the underlying analyses used the same information or assumptions.

The posted arm counts for the primary endpoint are 127 in the youth TI-CBT arm and 124 in the youth discussion control arm. Those are endpoint summary counts, not a statement of how many participants were originally randomized or completed every study procedure. The registry lists enrollment of 484, but does not provide a complete account of movement from enrollment through randomization, follow-up, and analysis.

What does over-dispersion mean in general?

In general, over-dispersion means that observed variability is greater than a model’s assumptions allow. When data are clustered, ordinary independent-observation methods may fail to represent the extra variation associated with shared groups. Methods that account for clustering or use cluster-level summaries can produce more appropriate uncertainty estimates when their assumptions fit the design. The registry’s stated group-level analytic unit reflects the relevance of clustering; the specific variance estimator and model details are not reported in the registry record.

The GAD-7 is a summed score, and the posted analysis treats its group-level mean as continuous. A mean difference describes the average separation but does not show the full distribution of individual or group scores. The registry does not report score distributions, baseline-adjusted estimates, or a minimally important difference for this endpoint. As a result, the mean comparison should be interpreted alongside its uncertainty and the definition of the scale, not as a complete description of every participant’s experience.

The record also lists sponsor-reported two-sided t-tests for the primary outcome. A t-test compares means while accounting for uncertainty under its specified assumptions. The registry entries do not provide enough detail to determine the exact sample unit, variance assumptions, degrees of freedom, or population used for each of the three reported tests. Those analytic details are not reported in the registry record, so the numerical results can be described but not fully audited from the posted summary alone.

Finally, p-values and confidence intervals depend on the analysis population and handling of incomplete observations. The primary population is described as youth who completed month 6 assessments. The number eligible for that assessment, reasons for non-completion, and method for handling missing outcomes are not reported in the registry record. The results consequently describe the reported analysis and should not be treated as a complete account of all enrolled participants’ outcomes.

11. Confidence Intervals and Precision

Precision concerns how tightly an estimate is determined by the data and analysis. For the primary comparison, the sponsor-reported mean difference is 0.22 GAD-7 score points, with a 95% confidence interval from -1.23 to 1.68. That interval is broad enough to include a difference favoring either arm as well as no difference. The point estimate is therefore more informative when read together with the interval than when considered alone.

The interval does not establish that the groups are equivalent. Establishing equivalence would require an appropriate equivalence design and a prespecified margin, neither of which is reported in the registry record. Nor does an interval spanning zero prove that there is no potentially important difference. The record does not report a clinically meaningful difference threshold for the GAD-7 endpoint, so the practical importance of values within the interval cannot be determined from the posted results alone.

The sponsor’s first reported interval is wider than the simple recalculation’s interval, which runs from -0.252 to 0.692. The two summaries use different uncertainty calculations; the recalculation is a simple unadjusted result from arm-level statistics and does not reproduce the study’s full analysis. Since the primary endpoint description says groups were the analytic units because of intracluster correlation, the sponsor-reported uncertainty should be considered the main trial analysis. The registry record does not provide enough detail to reconcile all differences between the reported and recalculated uncertainty estimates.

The three sponsor-reported primary analyses have intervals from -1.23 to 1.68, -1.62 to 2.34, and -1.88 to 2.01. Each spans zero. The registry does not explain why the estimates differ, so their varying precision cannot be attributed to particular populations or model choices from the record alone. In general, more variability, fewer independent analytic units, and greater clustering can reduce precision; that general principle does not identify which factor produced a particular interval in this study.

The sample-size illustrations use the observed differences as planning inputs, but they do not change the precision of the completed study or resolve the limits of the reported analysis. In interpreting the primary result, the key point is that the posted estimate is modest while the sponsor’s interval leaves substantial uncertainty about the size and direction of a difference.

12. 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 estimate how many participants per arm would be needed under the listed assumptions for a future two-arm comparison. They are not total event counts: the endpoint is continuous, and the figures are participants per arm. The planning assumptions are a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. The target effect is set to the stated fraction of the observed effect. The trial’s own original design assumptions are not reported in the registry record.

The 80% power column represents the planned sample size for a design with an 80% chance of detecting the specified effect if that effect is real and the planning assumptions hold. The 90% power column uses a higher chance of detection and therefore requires more participants for the same target effect. These are planning illustrations, not observed power.

For each outcome, planning for 75% or 50% of the observed effect requires more participants per arm than planning for 100% of the observed effect. Smaller effects are harder to distinguish from random variation, so they need larger samples to achieve the same target power under otherwise similar assumptions. The illustrations should not be read as a recommendation for a future study without considering its scientific question, analysis method, expected clustering, missing data, and meaningful effect size.

The calculations use the observed effect as their reference, which can be an unstable basis for planning, especially when an estimate is uncertain. The displayed figures also use equal allocation and a normal approximation, whereas the registry describes a group-based intervention and group-level primary analysis. The record does not say that the original trial was designed using these assumptions. These figures describe hypothetical planning scenarios rather than the study’s actual design rationale.

13. 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 = 0.220; pooled standard deviation σ = 1.913.
2. n = 2 × (1.96 + 0.842)² × 1.913² / 0.220² = 1,186.8, rounded up to 1,187 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 <- 0.22                      # observed difference in means
sigma <- 1.912892215078                     # 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(4024488)
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 = 1,187 (80% power), 1,589 (90% power)
 75% of observed effect: n per arm = 2,110 (80% power), 2,825 (90% power)
 50% of observed effect: n per arm = 4,748 (80% power), 6,356 (90% power)
power.t.test: 1187.7 
Simulated power at n = 1,187 per arm: 0.797 (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.

14. Analyses Reported by the Sponsor

The sponsor reports three two-sided t-test analyses for the primary endpoint. The first gives a mean difference of 0.22, a 95% confidence interval from -1.23 to 1.68, and p = 0.74. The other reported comparisons give differences of 0.36 and 0.07, with confidence intervals from -1.62 to 2.34 and -1.88 to 2.01, and p-values of 0.71 and 0.94. The registry does not explain why there are three analyses or how their populations or specifications differ.

The simple unadjusted recalculation from arm-level summaries also estimates a mean difference of 0.22, but gives a confidence interval from -0.252 to 0.692 and p = 0.361. The matching point estimate does not make the uncertainty calculations interchangeable. The recalculation does not reproduce the sponsor’s full analysis and does not incorporate details that are not available in the posted arm summaries. In particular, the registry says groups were the analytic units for primary outcomes because participants within a group may be correlated.

The registry’s discrepancy indicator is false. The sponsor’s reported results remain the trial’s own analyses; the recalculated result is best treated as a descriptive check, not as a replacement. The registry does not report sufficient analysis detail to explain the differences in uncertainty or the three sponsor-reported estimates.

15. Multiplicity, Missing Data and Analysis Populations

Multiplicity arises when a study evaluates several outcomes, time points, or statistical comparisons. Even when each individual test uses a conventional significance threshold, testing many hypotheses can increase the chance of at least one apparently unusual result arising by chance. Common approaches include prespecifying a primary outcome, adjusting for multiple tests, or labeling secondary analyses as exploratory. The registry identifies a primary endpoint and lists many secondary outcomes, but a multiplicity adjustment, testing hierarchy, and prespecified interpretation strategy are not reported in the registry record.

The record lists 12 outcome measures, 18 analyses, and 8 posted secondary outcomes. These counts indicate that results are not limited to a single comparison. Without a reported multiplicity plan, secondary p-values should not be treated as if they were the only tests conducted. The registry’s primary endpoint remains the main focus, while secondary findings provide supporting context and require cautious interpretation.

The primary analysis population is described as youth who completed month 6 assessments. The registry does not report the number of youth eligible for that assessment, the number with missing GAD-7 data, reasons for missingness, or how missing values were handled. It also does not report whether any participants were excluded after randomization or whether the analyses followed an intention-to-treat approach. These points matter because results among assessment completers may differ from results among all randomized participants if missingness is related to outcomes or group assignment.

Baseline characteristics, baseline-adjusted analyses, and detailed population definitions for the three sponsor-reported primary tests are not reported in the registry record. The record says groups were the analytic units for primary outcomes because of intracluster correlation, but does not give the full statistical model, variance method, or group-level attrition details. These omissions limit independent assessment of the precision and robustness of the analyses.

The study status is terminated, but the reason for termination and any relationship between termination and outcome availability are not reported. The record also does not describe interim analyses or stopping rules. No inference about those processes should be made from the status alone.

16. 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 serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: suicidal ideation (0.8%), hand fracture (0.8%), suicide attempt (0.8%).

The registry’s adverse-event summary lists a TI-CBT intervention group with 128 at risk for serious adverse events and 2 affected, and a discussion control group with 126 at risk and 1 affected. The registry lists no deaths in either group. These are the reported serious-event figures; they do not describe the full circumstances, timing, relatedness, or duration of events.

The listed serious terms include suicidal ideation, hand fracture, and suicide attempt, each shown as 0.8%. The registry does not provide enough detail in this summary to determine whether a term refers to a unique participant, repeated events, or the number of events in another reporting unit. It also does not report a causal assessment in the posted adverse-event summary. These entries should therefore be read as reported safety information, not as proof that an intervention caused or prevented an event.

For other adverse events, the registry lists 4 affected among 128 at risk in the TI-CBT group and 1 among 126 at risk in the discussion control group. Event definitions, severity categories, follow-up completeness, and the method for collecting adverse events are not reported in the registry record. Comparisons of these counts should be cautious because the summary does not supply the context needed to account for exposure or event ascertainment.

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

The posted primary result describes a small difference between the observed group-level mean GAD-7 scores at 6 months: 4.51 in the youth TI-CBT arm and 4.29 in the discussion control arm. The sponsor-reported mean difference is 0.22, with a 95% confidence interval from -1.23 to 1.68 and p = 0.74. The interval includes zero and a range of differences in either direction, so the result does not precisely establish a separation between the arms.

That does not establish that the interventions are equivalent, nor does it prove that no difference exists. The registry does not report an equivalence margin or a clinically important difference threshold for the GAD-7. The confidence interval is the relevant guide to statistical precision, while the score definition helps explain what is measured; neither alone determines the practical importance of every possible value in the interval.

The primary measure concerns anxiety symptoms reported over the past 2 weeks at the month 6 assessment. It does not summarize all mental health outcomes, ART adherence, viral load, or longer-term outcomes. The registry lists those as separate outcomes. Nor does a group-level mean describe the range of individual responses within each group.

The trial was randomized and open label, and the primary analysis used groups as analytic units due to intracluster correlation. These features provide important context for interpreting the comparison. The registry does not report detailed missing-data handling, all analysis specifications, or the reason the study was terminated. Those gaps constrain what can be concluded about generalizability and robustness from the registry record alone.

The appropriate reading is therefore narrow: the posted results show no clear, precise difference on the reported primary endpoint under the sponsor’s analysis. They do not establish a broader conclusion about every outcome, every setting, or every participant. The finding is an account of the recorded comparison, not medical advice or a judgment about what any individual should do.

18. Why This Trial Matters Statistically

This trial raises statistical issues that are common in behavioral interventions delivered in groups. Participants who meet in the same group may share experiences, leaders, and session context, so their outcomes may be correlated. The registry explicitly identifies intracluster correlation and says that groups were the analytic units for primary outcomes. That makes the unit of analysis central to understanding the uncertainty, not merely a technical detail.

The study also distinguishes individual randomization from group-based delivery and group-level analysis. Clear reporting of the relationship among assignment, intervention groups, and analysis units helps readers judge whether the reported standard errors and confidence intervals represent the design. The registry states the rationale for group-level analysis, but detailed model specifications and variance methods are not reported. This illustrates why reporting both the design and the analysis is important.

The primary result shows why estimates should be read with confidence intervals rather than by focusing only on point estimates or p-values. A mean difference of 0.22 has a sponsor-reported interval that includes zero and values on either side. The data therefore do not provide a precise estimate of the direction or size of the difference. The distinction between absence of strong evidence and evidence of equivalence is especially important when the interval is not tied to a reported equivalence margin.

The record also shows the interpretive challenges created by multiple outcomes and analyses. Mental health scores, adherence, and viral load address different questions and may be measured at different times. Without a reported multiplicity plan, secondary findings need more cautious interpretation than a clearly designated primary endpoint.

Finally, the posted planning illustrations should not be confused with the trial’s original design rationale. They are based on observed effects and specified hypothetical assumptions, while the original sample-size assumptions are not reported in the registry record. Separating post hoc planning scenarios from actual design decisions is important for transparent statistical interpretation.

19. 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 gives an incomplete account of several aspects needed to evaluate the primary result. Although it describes the study as multi-site, the locations are not reported. Detailed eligibility criteria, baseline characteristics, and the balance of participant characteristics between arms are also not reported in the registry record. These omissions make it difficult to assess how closely the analyzed participants represent the broader population of youth living with HIV in resource-limited settings.

The primary endpoint population is defined as youth who completed month 6 assessments, but the record does not report the number of missing assessments, reasons for missingness, or how missing data were handled. It does not state whether the analysis included all randomized participants or whether participants were excluded. The primary group-level analytic approach is described, but the number of groups in each analysis and full model details are not reported.

The registry includes three sponsor-reported primary analyses with different estimates, but does not explain their populations or specifications. It also does not report a multiplicity strategy for the 12 outcome measures and 18 analyses. These gaps limit independent assessment of how the primary result was selected and how secondary results should be weighed.

The descriptions of several arms are incomplete, including details of the youth discussion control and caregiver interventions. The registry does not report the reason for termination, so the status alone cannot explain whether follow-up, recruitment, or analysis was affected. The date and enrollment entries also do not provide a full participant flow from enrollment through the endpoint analysis.

Finally, the registry does not report a clinically meaningful difference threshold for the GAD-7 or the original sample-size justification. Without those details, the practical significance of the confidence interval and the intended precision of the trial cannot be assessed against stated design goals.

20. Related Tutorials

21. Related Statistical Calculators

22. Sources

• ClinicalTrials.gov: Group-Based Intervention to Improve Mental Health and Adherence Among Youth Living With HIV in Low Resource Settings — NCT04024488.
• Registry record last updated 2026-05-22; status last verified 2026-04.
• Results first posted 2026-05-22.
• 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.

23. Record Summary

The registry describes NCT04024488, IMPAACT 2016, as a randomized, open-label, parallel-group study of a group-based intervention intended to address mental health and ART adherence among youth living with HIV in resource-limited settings. The listed enrollment is 484, and the study status is terminated. The primary outcome is the group-level mean GAD-7 score at 6 months among youth who completed that assessment.

The youth TI-CBT arm had a posted mean of 4.51, compared with 4.29 in the youth discussion control arm. The sponsor-reported mean difference is 0.22, with a 95% confidence interval from -1.23 to 1.68 and p = 0.74. The interval includes zero and is compatible with differences in either direction. A simple unadjusted recalculation gives the same point estimate but different uncertainty estimates.

The registry lists additional mental health, ART adherence, viral load, and safety outcomes. It does not report the reason for termination, the trial’s original sample-size assumptions, full analysis details, or complete missing-data and multiplicity procedures.