Friendship Bench Adaptation to Improve Mental Health & HIV Care Engagement Outcomes Among PLWH and PWID in Vietnam: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Friendship Bench Adaptation to Improve Mental Health & HIV Care Engagement Outcomes Among PLWH and PWID in Vietnam (NCT04790201), a randomized na study. The primary outcome is Study Retention (Study Feasibility).

2. Clinical Question

The study asks whether a Friendship Bench counseling approach adapted for people living with HIV who inject drugs in Vietnam can be delivered by professional or trained lay counselors, and whether participants can be retained through study completion. Retention is the designated primary feasibility measure. The registry also lists outcomes related to HIV viral suppression, HIV visit attendance, and mental health symptoms.

The project is described as an adaptation and pilot intended to generate evidence for designing a fully powered trial. Its central statistical question is therefore whether the posted feasibility and outcome estimates are informative enough to guide later study design, while recognizing that estimates from a pilot can be imprecise. The record does not state a formal hypothesis for comparing the two counseling approaches with enhanced usual care.

3. Trial Design

This was a randomized, parallel-group study with no masking reported. The registry lists an enrollment of 77 and a completed study. Its stated dates are 2022-02-28 for study start and 2023-07-31 for primary completion. The phase is listed as NA. The sponsor is the University of North Carolina, Chapel Hill.

There were three study arms: Friendship Bench delivered by a professional counselor, Friendship Bench delivered by a lay counselor, and enhanced usual care. The counseling arm descriptions specify 6 weekly sessions. The enhanced usual care description mentions general training of HIV providers and clinics on common mental disorder identification and management. More detail about what the counseling sessions contained, how the lay counselors were trained, or what additional enhanced usual care activities involved is not reported in the registry record.

The primary outcome population is described as 75 eligible participants enrolled. The record says two randomized participants, one in the professional-counselor arm and one in enhanced usual care, were found ineligible shortly after enrollment and excluded. For the posted retention outcome, the professional-counselor and lay-counselor arms each list 25 participants, with 25 retained in each; enhanced usual care lists 25 participants, with 23 retained.

The registry describes the project as a pilot to inform the design of a fully powered trial. The rationale for the enrollment target, randomization method, allocation concealment, eligibility criteria in full, and any stratification factors are not reported in the registry record. The record also does not report a masking procedure; the study is listed as having no masking.

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: Two of the randomized participants (1 in Friendship Bench delivered by Professional Counselor arm and 1 in Usual Care arm) were found to be ineligible for the study shortly after enrollment and were excluded from all analyses.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: The analysis population includes the total number of people approached for study eligibility and screening. Two of the randomized participants (1-Friendship Bench Delivered by Professional Counselor and 1-Enhanced Usual Care) were found to be ineligible for the study shortly after enrollment and were excluded from all analyses.

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: Two of the randomized participants (1-Friendship Bench Delivered by Professional Counselor and 1-Enhanced Usual Care) were found to be ineligible for the study shortly after enrollment and were excluded from all analyses.

6. Primary Endpoint

The registry lists the primary outcome as Study Retention (Study Feasibility), measured over Through study completion, an average of 12 months.

The primary endpoint is study retention, expressed as a count of participants. The registry defines it as the number of people enrolled at baseline who remained enrolled through study completion. The stated timeframe is through study completion, an average of 12 months. Retention is used as a feasibility measure because it describes whether participants can be kept engaged in the pilot long enough to complete study participation.

The posted counts are 25 retained among 25 listed participants in the professional-counselor arm, 25 among 25 in the lay-counselor arm, and 23 among 25 in enhanced usual care. The comparisons against enhanced usual care give a risk difference of 0.08 for each counseling arm. The recalculated unadjusted p-value for each comparison is 0.149. These are descriptive estimates from the posted arm counts; they do not establish why participants remained or left the study, nor do they by themselves demonstrate feasibility in other settings.

Retention is distinct from treatment adherence, completion of every counseling session, HIV-care engagement, and improvement in mental health. The registry lists those topics among other outcomes, but the primary endpoint alone cannot answer them. The record does not report a prespecified threshold for what retention level would be considered feasible, or a formal decision rule linking retention to progression to a later trial.

7. Results Posted in the Registry

• Friendship Bench Delivered by Professional Counselor: 100.0% (25/25) versus 92.0% (23/25) on control; risk difference 8.0 percentage points (95% CI -2.6 to 18.6), risk ratio 1.09 (0.97–1.22); unadjusted two-proportion z-test p = 0.149.
• Friendship Bench Delivered by Lay Counselor: 100.0% (25/25) versus 92.0% (23/25) on control; risk difference 8.0 percentage points (95% CI -2.6 to 18.6), risk ratio 1.09 (0.97–1.22); unadjusted two-proportion z-test p = 0.149.

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.

Number of Participants Achieving HIV Viral Suppression

• Endpoint type: Binary: the number of participants with the event, compared as proportions.
• Test: Unadjusted two-proportion z-test, with a Wald 95% confidence interval for the risk difference.
• Result:
  • Friendship Bench Delivered by Professional Counselor: 76.0% (19/25) versus 91.3% (21/23) on control; risk difference -15.3 percentage points (95% CI -35.6 to 5.0), risk ratio 0.83 (0.65–1.07); unadjusted two-proportion z-test p = 0.155.
  • Friendship Bench Delivered by Lay Counselor: 87.5% (21/24) versus 91.3% (21/23) on control; risk difference -3.8 percentage points (95% CI -21.3 to 13.7), risk ratio 0.96 (0.79–1.17); unadjusted two-proportion z-test p = 0.672.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Friendship Bench Delivered by Professional Counselor and the control. The interval gives the range of effects that remain compatible with the data.
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Friendship Bench Delivered by Lay Counselor and the control. The interval gives the range of effects that remain compatible with the data.

Total CMD Symptoms Score for Participants

• 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:
  • Friendship Bench Delivered by Professional Counselor: mean 27.28 versus 32.30 on control; difference -5.02 (95% CI -18.79 to 8.75), standardized effect d = -0.20; z-test p = 0.475.
  • Friendship Bench Delivered by Lay Counselor: mean 23.90 versus 32.30 on control; difference -8.40 (95% CI -20.76 to 3.96), standardized effect d = -0.38; z-test p = 0.183.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Friendship Bench Delivered by Professional Counselor and the control. The interval gives the range of effects that remain compatible with the data.
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Friendship Bench Delivered by Lay Counselor and the control. The interval gives the range of effects that remain compatible with the data.

Change in CMD Symptoms

• 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:
  • Friendship Bench Delivered by Professional Counselor: mean -24.20 versus -13.40 on control; difference -10.80 (95% CI -25.14 to 3.54), standardized effect d = -0.42; z-test p = 0.140.
  • Friendship Bench Delivered by Lay Counselor: mean -23.50 versus -13.40 on control; difference -10.10 (95% CI -25.70 to 5.50), standardized effect d = -0.36; z-test p = 0.205.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Friendship Bench Delivered by Professional Counselor and the control. The interval gives the range of effects that remain compatible with the data.
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Friendship Bench Delivered by Lay Counselor and the control. The interval gives the range of effects that remain compatible with the data.

The secondary outcomes broaden the study beyond retention to include HIV viral suppression, HIV visit attendance, and common mental disorder symptoms, among other listed measures. Posted summaries include viral suppression at 6 months after enrollment and symptom measures at 6 weeks after enrollment. The registry lists additional secondary outcomes, including HIV visits and depressive and anxiety symptoms, but full comparative results for every listed measure are not displayed in the posted summaries described in the record.

Across the posted comparisons, the estimates do not form a uniformly consistent pattern. Viral suppression contrasts differ in size between the two counseling arms, while symptom comparisons generally have wide uncertainty around their estimated differences. The registry’s outcome list also includes change in symptoms, which should not be conflated with a single follow-up score: a change measure and a follow-up mean answer different questions.

These outcomes are relevant to the study’s broader aims, but they do not change the interpretation of the primary feasibility endpoint. Retention indicates whether participants stayed enrolled; it does not establish improvement in viral suppression, care engagement, depression, anxiety, or drug-use outcomes. Likewise, secondary outcome patterns cannot compensate for uncertainty in the retention comparison.

The secondary comparisons are unadjusted, and the registry does not report a plan for controlling the chance of false-positive findings across multiple outcomes or contrasts. When many outcomes are examined, isolated nominal findings can arise by chance even if no true effects are present. Without a stated hierarchy, multiplicity procedure, or confirmatory plan, secondary results are best regarded as exploratory evidence for shaping future questions rather than as definitive tests.

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

• Number of Participants Achieving HIV Viral Suppression
• Proportion of Scheduled HIV Visits That Were Attended in the 12-Month Follow-Up Period
• Total CMD Symptoms Score for Participants
• Change in CMD Symptoms
• Mean Depressive Disorder Score Among Participants
• Absolute Reduction in Depressive Symptoms
• Mean Anxiety Disorder Score Among Participants
• Absolute Reduction in Anxiety Symptoms

9. Statistical Interpretation

10. Statistical Methods Explained

The comparisons are simple, unadjusted recalculations from the posted arm summaries. The registry record does not report the sponsor’s statistical analysis results, and no reported analyses are listed. The study is randomized and parallel-group, but details of the planned statistical methods are not reported in the registry record. These recalculations should not be mistaken for an adjusted analysis or for a substitute for any sponsor analysis.

How to read the retention comparison

For a binary outcome such as retained or not retained, a risk is the proportion of participants with the outcome in an arm. A risk difference compares the two arm-specific risks by subtraction. A value above zero means the event was more common in the named counseling arm than in enhanced usual care; a value below zero would mean it was less common. For retention, the posted risk difference is 0.08 for each counseling arm versus enhanced usual care. The registry also gives arm counts, which are the most direct description of how many participants were retained.

A risk difference is expressed on the probability scale and is often easier to interpret as an absolute contrast than a relative measure. It does not describe the probability that the intervention caused an outcome. Because the primary endpoint is a count with a binary participant status underneath, the count and the arm denominator need to be read together. A count alone does not indicate the retention rate without its denominator.

What the p-values say

The reported p-values for the primary comparisons are unadjusted recalculations. A p-value measures how incompatible the observed data are with a specified null model, under the assumptions of the calculation. It is not the probability that the null hypothesis is true, and it does not measure the size or practical importance of a difference. The p-values of 0.149 do not establish that the counseling arms and enhanced usual care are the same. The registry record does not report a prespecified significance threshold or a statistical analysis plan.

The recalculations are not adjusted for other participant characteristics, and the registry record does not provide covariate-adjusted estimates. The record does not state whether the original analysis was intended to follow intention-to-treat principles, use a per-protocol population, or use another analysis population beyond the primary endpoint description. These distinctions matter because exclusions and missing outcome information can affect comparisons.

Reading continuous outcomes

For a continuous outcome, a mean difference compares the reported group means. Here, the posted symptom outcomes use means and standard deviations. A standard deviation describes variability among individual observations; it is not the uncertainty around the group mean. The recalculated mean differences compare each counseling arm with enhanced usual care. Their signs reflect the arithmetic direction of the contrast, but whether a higher or lower score is favorable depends on the scale and its interpretation. The registry record does not give enough scale detail to interpret the clinical meaning of a particular score change.

Standard errors and confidence intervals reflect uncertainty in estimated means and contrasts. A standard error concerns the precision of an estimate, while a standard deviation describes spread among observations. Treating one as the other would misstate variability. The posted continuous outcome comparisons use the reported means, standard deviations, and group sizes; the registry does not report adjusted models or details of distributional checks.

Planning illustrations are not observed power

The planning illustrations use the observed effect as a starting point, assume a two-sided significance level of 0.05, equal allocation between two arms, and a normal approximation. They show participants per arm for 80% or 90% power when planning for 100%, 75%, or 50% of the observed effect. For example, the viral-suppression comparison for the professional-counselor arm lists 91 participants per arm for 80% power when planning for 100% of its observed effect, and 146 when planning for 75% of that effect. For the lay-counselor comparison, the corresponding illustrations are 1,027 and 1,754 participants per arm. These are planning illustrations, not observed power, and they are sensitive to assumptions about the effect to be detected.

The registry does not report the original sample-size justification or state that these planning assumptions were used to design the pilot. It also does not report assumptions about loss to follow-up, adjustment for multiple comparisons, or whether the future trial would use the same outcome definitions. The illustrations should therefore be treated as one way to explore possible study sizes, not as a definitive design recommendation.

General statistical point: over-dispersion matters when observed variation is greater than a model’s assumptions allow, because standard errors can then be too small and confidence intervals too narrow. No over-dispersion parameter is reported for the posted primary endpoint, which is a participant count. The registry record does not report any model-based analysis that would permit assessment of over-dispersion.

11. Confidence Intervals and Precision

A confidence interval gives a range of effect estimates compatible with the observed data under the method’s assumptions. Its width conveys precision: a wider interval indicates greater uncertainty about the magnitude and direction of the comparison. It is not a range that guarantees the true effect lies inside it, nor is it a probability statement about the effect after observing the data.

For the primary retention comparisons, the reported recalculated confidence intervals extend from a negative value to a positive value. They therefore include no difference. The point estimate of 0.08 is not precise enough to rule out either a small difference in the opposite direction or a larger positive difference. The same estimate and interval are reported for both counseling arms against enhanced usual care. That similarity should not be interpreted as evidence that the two counseling models have identical effects.

The secondary estimates are also uncertain. For viral suppression, each comparison’s interval includes no difference. The continuous symptom comparisons likewise have intervals that include no difference. These intervals communicate uncertainty about comparative effects and are more informative than a binary interpretation of whether an unadjusted p-value crosses a threshold.

The trial’s arm sizes are limited, which contributes to uncertainty, especially for outcomes with few events or substantial variation. The registry record does not report an analysis accounting for possible missing outcome data, nor does it provide a detailed account of the precision expected in the original design. The planning illustrations use selected fractions of observed effects; they do not remove uncertainty from the pilot estimates on which those illustrations are based.

12. Multiplicity, Missing Data and Analysis Populations

The registry record does not report how multiplicity was handled. It lists multiple secondary outcomes and multiple comparisons, but does not state whether a primary-secondary hierarchy, family-wise error procedure, false discovery approach, or other adjustment was planned. The displayed recalculated p-values are unadjusted. As a general statistical matter, testing many outcomes increases the chance that at least one result appears notable by chance; unadjusted secondary findings therefore require cautious interpretation.

The record does not report the planned approach to missing data, including how missing follow-up outcomes were to be handled or whether sensitivity analyses were planned. The number of participants contributing to some secondary outcome summaries differs across arms and outcomes. The registry does not explain all of these differences, so they should not be assigned a cause without further information. The adverse-event denominators also differ from the primary outcome denominators; the reason is not reported in the registry record.

The primary population description states that 75 eligible participants were enrolled and that two randomized participants found ineligible shortly after enrollment were excluded. The record does not describe whether any other exclusions occurred for secondary outcomes, nor does it give a complete accounting of participants at each follow-up. It does not report whether the analysis was intention-to-treat, per-protocol, or based on available cases for each endpoint. These omissions limit assessment of whether the comparisons preserve the randomized groups and how missing observations may affect interpretation.

The registry gives no statistical analysis plan, interim-analysis information, or detailed rules for handling protocol deviations. These items are not reported in the registry record. Their absence should not be filled in by assuming standard practice.

13. Safety Results

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

Deaths

Other (non-serious) adverse events

Participants with at least one non-serious event at or above the registry's 5% reporting threshold.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: pancreatitis (3.8%).

The adverse-event summaries list serious events affecting 1 participant in the professional-counselor arm, 0 in the lay-counselor arm, and 0 in enhanced usual care. The listed numbers at risk are 26, 25, and 26, respectively. The summaries also list other adverse events affecting 2 participants in the professional-counselor arm, 1 in the lay-counselor arm, and 0 in enhanced usual care. These counts describe the posted record; the different denominators and the reason for them are not explained.

One death is listed in the lay-counselor arm, with 0 listed in the other arms. The record does not state a cause or relationship to study participation, so none should be inferred. Pancreatitis is listed among the most frequent serious adverse-event terms at 3.8%; the registry record does not provide further context in the posted summary to interpret that term’s distribution across arms.

These figures should not be used to compare safety conclusively. The study is a pilot, event counts are small, and the registry summary does not report a detailed adverse-event analysis, event timing, severity breakdown beyond the serious classification, or attribution. Absence of a listed event in an arm does not establish absence of risk.

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

The primary result describes retention during this pilot: the posted counts show that all listed participants in both Friendship Bench arms remained enrolled, compared with 23 listed as retained in enhanced usual care. The recalculated risk differences are 0.08 for each counseling arm relative to enhanced usual care, and both comparisons have p = 0.149. The uncertainty intervals include no difference. These findings do not establish a reliable difference in retention between study arms.

The result does not show that Friendship Bench counseling improves mental health, HIV viral suppression, HIV visit attendance, or other clinical outcomes. Those outcomes are separate from retention, and the posted secondary comparisons are exploratory and imprecise. Nor does the result show that the professional and lay delivery approaches perform equally; the reported comparisons do not directly test one counseling arm against the other.

The record describes the work as an adaptation and pilot intended to inform a future fully powered trial. In that context, the results can be useful as preliminary estimates and as a record of how retention and other outcomes were observed in this setting. They do not establish the effect that would be expected in a larger study or in other populations and service settings.

General statistical point: a nonsignificant test does not prove no effect, just as a point estimate in a favorable direction does not prove benefit. The effect estimate, its confidence interval, the outcome definition, and the study context all matter. The registry does not report a predefined threshold for an acceptable retention rate or a confirmatory decision rule, so the result cannot be translated into a definitive judgment about whether the intervention should be adopted or advanced.

15. Why This Trial Matters Statistically

This study is statistically relevant because it evaluates feasibility in a population and setting that the project describes as important for future research: people living with HIV who inject drugs in Vietnam and experience common mental disorder symptoms. It examines two counselor delivery models within a randomized parallel-group pilot, alongside enhanced usual care. That structure can generate comparative estimates and reveal what a later trial may need to measure.

Retention is an appropriate kind of feasibility outcome for a pilot because the ability to keep participants engaged affects whether a larger study can obtain follow-up data. The posted counts suggest that retention was high across the arms, while the uncertainty around the comparative estimates cautions against treating the observed differences as stable. The contrast between feasibility information and evidence of efficacy is central: the primary result concerns remaining enrolled, not whether a health outcome improved.

The secondary outcome summaries may help identify outcomes for a future study and indicate where estimates remain especially uncertain. Planning illustrations based on alternative fractions of observed effects also make clear that the required study size can change substantially with the effect assumed. Those calculations are not a substitute for a design grounded in justified effect assumptions, anticipated missingness, and a prespecified analysis plan.

The pilot therefore has value as an early evidence-generating study, while its results remain preliminary. A later trial would need clearly stated primary and secondary hypotheses, a transparent approach to missing data and multiplicity, and sufficient precision for its intended questions. The registry record does not report whether such plans have been finalized.

16. 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 several limitations for statistical interpretation. It does not report a statistical analysis plan, sample-size justification, interim analyses, or a prespecified threshold for judging feasibility. The project is described as a pilot intended to inform a fully powered trial, but the record does not state a formal progression criterion.

The posted primary comparisons are simple unadjusted recalculations. The record does not provide sponsor-reported comparative analyses, adjusted estimates, or enough detail to assess the methods originally intended. The two randomized participants excluded after being found ineligible are described, but the registry does not explain how other exclusions, protocol deviations, or missing follow-up data were handled.

Several outcomes are listed, but comparative results are not available for every listed secondary measure in the record. The registry also does not report a multiplicity strategy. This makes isolated secondary findings difficult to interpret as confirmatory evidence. The record’s symptom measures are described as scores, but fuller information about scale interpretation and clinically meaningful change is not reported.

The adverse-event summaries use denominators that differ from those shown for the primary outcome, and the reason is not reported. Event attribution, timing, and detailed safety analyses are also not reported. Finally, the study was conducted among people seeking services at participating clinics in Hanoi, Vietnam; the registry record does not establish whether the observed estimates apply to other settings or populations.

17. Related Tutorials

18. Related Statistical Calculators

19. Sources

• ClinicalTrials.gov: Friendship Bench Adaptation to Improve Mental Health & HIV Care Engagement Outcomes Among PLWH and PWID in Vietnam — NCT04790201.
• Registry record last updated 2024-08-13; status last verified 2024-05.
• Results first posted 2024-08-13.
• 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.

20. Record Summary

NCT04790201 is a completed, randomized, parallel-group study with no masking reported. It enrolled 77 participants and compared Friendship Bench counseling delivered by professional counselors, Friendship Bench counseling delivered by lay counselors, and enhanced usual care in Hanoi, Vietnam. The project was described as an adaptation and pilot intended to inform a later fully powered trial.

The primary feasibility outcome was retention through study completion, an average of 12 months. The posted counts were 25 of 25 in each counseling arm and 23 of 25 in enhanced usual care. Each unadjusted recalculated risk difference versus enhanced usual care was 0.08, with p = 0.149. The registry also lists secondary HIV and mental health outcomes and adverse-event summaries.

Comparative estimates are preliminary and imprecise. The registry record does not report a full statistical analysis plan, a missing-data strategy, or multiplicity handling.