Community Health Worker Training to Reduce Depression and Substance Use Stigma in TB/HIV Care in South Africa: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Community Health Worker Training to Reduce Depression and Substance Use Stigma in TB/HIV Care in South Africa (NCT05282173), a randomized na study. The primary outcome is Training Feasibility.

2. Clinical Question

The broader study purpose was to examine implementation and preliminary effectiveness, including changes in CHW stigma towards patients with depression and substance use. The listed secondary outcomes are CHW stigma towards substance use and CHW stigma towards depression. These are distinct from the primary feasibility endpoint: attendance describes participation in training, while stigma outcomes concern attitudes. The registry record does not report results for either secondary outcome, so the extent or direction of any change cannot be described.

3. Trial Design

The registry identifies this as a randomized, cluster-randomized crossover study with no masking. It lists enrollment of 82 participants and describes three sequences. In each sequence, CHWs were monitored during treatment as usual before receiving Siyakhana training: approximately 1 Month of TAU in Sequence 1, approximately 2 Months in Sequence 2, and approximately 3 Months in Sequence 3. The registry describes the training as multi-day and specifies that the primary attendance endpoint concerns all three training days.

A crossover design means participants or groups move through more than one study condition or period, rather than remaining in only one condition throughout. Here, the sequence descriptions indicate that TAU preceded the training in each sequence, but the registry record does not report enough detail to reconstruct the full schedule, timing of outcome assessments across periods, or whether and how any carryover was addressed. Because the study is identified as cluster-randomized, assignment was not simply at the individual level; the registry record does not report the cluster unit or the number of clusters.

The study status is completed. The recorded start date is 2022-06-08, and the primary completion date is 2023-08-01. The sponsor is the University of Maryland, College Park. The registry lists seven outcome measures, but it reports no analysis results. The record does not report the statistical analysis plan, randomization procedure, allocation concealment, or the planned analysis population. These omissions limit what can be concluded about how outcomes were intended to be compared across sequences and periods.

4. Participant Flow and Accounting

The registry reports how participants moved through each stage of the study. All counts below are exactly as posted; missing counts are not inferred from other stages.

Overall Study

Reasons for not completing this period:

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 Training Feasibility, measured over 3-months post-training.

The primary endpoint is the number of CHWs who attended all three days of the Siyakhana training. It is measured at 3-months post-training and is expressed as a count of participants. The registry defines a feasibility benchmark: training is considered feasible if > 75% of CHWs attend the full training. This endpoint therefore addresses whether CHWs complete the training, rather than whether their attitudes or patient-care outcomes change.

Attendance is useful as an implementation measure because a training cannot deliver its intended content to participants who do not complete it. A full-attendance count can describe uptake and completion in the enrolled study population. The registry record does not report the observed count, the percentage attending all three days, or whether the benchmark was reached. It also does not report how attendance was handled for participants with incomplete or unavailable records.

General statistical context: a count communicates the number of observed participants meeting a defined criterion. Interpretation depends on the denominator and on consistent ascertainment of the criterion. The feasibility benchmark is a decision rule described in the registry; it is not itself evidence of effectiveness. Attendance alone cannot establish that stigma changed or that patient engagement in HIV/TB care improved. Those questions require their own outcome data and analyses, which are not reported in the registry record.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The record identifies a randomized, cluster-randomized crossover design, but does not report the statistical analysis plan or analysis results. The concepts below explain why those design features matter in general; they do not describe analyses confirmed to have been used in this trial.

Why does cluster randomization matter?

In cluster randomization, groups rather than individual participants are assigned to study conditions. Participants in the same group may have more similar outcomes than participants in different groups. General statistical context: this similarity is called within-cluster correlation. If an analysis treats every participant as independent when outcomes are correlated within clusters, it may understate uncertainty. An analysis generally needs to reflect the assignment unit and the dependence structure. The registry record does not identify the cluster unit, report the number of clusters, or describe any adjustment for clustering.

What does the crossover structure add?

A crossover study organizes conditions across periods, so the order or timing of conditions can be relevant to interpretation. Here, the sequence descriptions place an approximate period of TAU before Siyakhana training, with the sequences differing in the approximate duration of TAU. General statistical context: period effects can arise when outcomes change over time for reasons apart from the intervention, while carryover means that experience in an earlier period may affect outcomes in a later period. Analyses of crossover studies may need to account for period and sequence and to consider carryover. The registry record does not report whether these features were included in an analysis or how outcomes were timed across periods.

What does the primary endpoint measure?

The endpoint is a participant count: CHWs who attended all three days of the training, assessed at 3-months post-training. The feasibility rule is attendance by > 75% of CHWs. General statistical context: a count can be summarized directly, while a percentage relates the count to a clearly defined denominator. The registry record does not report the observed count, the endpoint denominator, or a confidence interval. It also does not describe the handling of missing attendance data. Without those details, the feasibility result cannot be reconstructed.

How should the posted analysis status be understood?

The registry record lists seven outcome measures and no analysis results. This means that the record does not present inferential estimates for the primary or secondary outcomes. General statistical context: a point estimate describes the estimated magnitude of an outcome or contrast, while its uncertainty interval conveys precision under the analysis model. A p-value, when reported, relates the observed data to a specified null hypothesis and model assumptions. None of these should be inferred from the trial's enrollment or safety counts.

The record also indicates that a simple individual-level re-analysis would be misleading for this cluster-randomized design, and that only the sponsor's own analysis should be used. No sponsor analysis result is reported in the available record. Accordingly, there is no basis for a recalculated unadjusted comparison, and no analysis should be substituted for the absent sponsor result.

What do the planning assumptions mean?

The record describes planning illustrations based on an observed effect, using a two-sided significance level of 0.05, equal allocation between two arms and a normal approximation; the effect to detect is expressed as a stated fraction of the observed effect. For time-to-event illustrations, the stated quantity is total events; for other illustrations, it is participants per arm. General statistical context: such calculations explore how sample size relates to a chosen effect and assumptions about variability and error rates. They are planning illustrations, not observed outcomes or evidence that the study had a particular probability of detecting an effect. The record does not report the numerical illustration results or the trial's own sample-size justification. Because no observed effect estimate is reported, no interpretation of effect-based planning values can be made.

More broadly, the absence of analysis details matters because statistical conclusions depend on the model, unit of analysis, outcome definition, missing-data handling and population analyzed. Those details are not reported in the registry record. The design description alone does not establish which methods were used or whether their assumptions were met.

10. Exact Estimates for Each Arm

No two-arm comparison is shown for this trial, but each arm's result can still be summarized with exact interval estimates computed from the posted counts. The Clopper-Pearson interval is exact and conservative; the Wilson interval is a common closer-to-nominal alternative; the Bayesian credible interval uses a uniform Beta(1, 1) prior. For small groups the three can differ noticeably.

R code and output

arm    <- c("Sequence 1: Approximately 1 Month Treatment as Usu", "Sequence 2: Approximately 2 Months TAU, Siyakhana ", "Sequence 3: Approximately 3 Months TAU, Siyakhana ")
n      <- c(29, 25, 28)
events <- c(26, 25, 27)
for (i in seq_along(n)) {
  ci <- binom.test(events[i], n[i])$conf.int          # Clopper-Pearson exact interval
  cr <- qbeta(c(0.025, 0.975), 1 + events[i], 1 + n[i] - events[i])   # Bayesian, uniform Beta(1,1) prior
  wi <- (function(x, n) { p <- x / n; z <- qnorm(0.975); d <- 1 + z^2 / n
    c((p + z^2 / (2 * n) - z * sqrt(p * (1 - p) / n + z^2 / (4 * n^2))) / d, (p + z^2 / (2 * n) + z * sqrt(p * (1 - p) / n + z^2 / (4 * n^2))) / d) })(events[i], n[i])
  cat(sprintf("ARM|%s|%d|%d|%.1f|%.1f|%.1f|%.1f|%.1f|%.1f|%.1f\n", arm[i], n[i], events[i], 100 * events[i] / n[i],
              100 * ci[1], 100 * ci[2], 100 * wi[1], 100 * wi[2], 100 * cr[1], 100 * cr[2]))
}

ARM  Sequence 1: Approximately 1 Month Treatment as Usu  29  26  89.7  72.6  97.8  73.6  96.4  73.5  96.2
ARM  Sequence 2: Approximately 2 Months TAU, Siyakhana   25  25  100.0  86.3  100.0  86.7  100.0  86.8  99.9
ARM  Sequence 3: Approximately 3 Months TAU, Siyakhana   28  27  96.4  81.7  99.9  82.3  99.4  82.2  99.2

11. Multiplicity, Missing Data and Analysis Populations

The registry record lists seven outcome measures, including training feasibility as the primary outcome and CHW stigma towards substance use and CHW stigma towards depression as secondary outcomes. It does not report a multiplicity strategy, such as whether and how the interpretation of multiple outcome tests was adjusted. General statistical context: when several hypotheses are tested, the chance of at least one apparently notable result can increase, depending on the testing plan and relationships among outcomes. A prespecified strategy helps readers distinguish primary questions from exploratory ones. No such strategy is reported, and no outcome test results are posted.

Missing-data methods and analysis populations are also not reported in the registry record. The record does not say how missing attendance information or missing stigma assessments were handled, whether participants were analyzed according to their assigned sequence, or which participants contributed to each outcome. Enrollment of 82 does not establish the denominator for any particular endpoint. General statistical context: missingness can affect both the representativeness of an observed result and its precision, especially when missingness is related to participant characteristics or outcomes.

The cluster-randomized crossover design makes these omissions especially relevant to interpretation. The record does not report the analysis unit, the cluster structure used in analysis, how sequence and period were handled, or whether any adjustment for clustering was made. These details are needed to understand the validity and precision of any comparison. Since the registry record reports no analyses or results, neither multiplicity adjustments nor missing-data procedures can be evaluated.

12. 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.

For Sequence 1, the registry reports 0 serious adverse events among 29 participants at risk, 0 other adverse events among 29, and 0 deaths. For Sequence 2, it reports 0 serious adverse events among 25 participants at risk, 0 other adverse events among 25, and 0 deaths. For Sequence 3, it reports 0 serious adverse events among 28 participants at risk, 0 other adverse events among 28, and 0 deaths. These are the posted counts for the sequence groups.

General statistical context: a count of zero means no events of that category were recorded in the stated group; it does not prove that the underlying risk is zero. The precision of safety estimates depends on the number of participants observed, duration and completeness of follow-up, and how events were collected and defined. The registry record does not report event definitions, follow-up completeness, event-level descriptions, or a comparative safety analysis. The approximate TAU durations differ by sequence, and the record does not provide enough detail to determine exposure time to each condition for a safety comparison. These counts should therefore be read descriptively, without interpreting them as evidence that the training is free of risk or as proof of a difference between sequences.

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

The registry record establishes that the study evaluated a multi-day CHW training intended to address stigma around mental health and substance use, in the context of TB/HIV care. Its primary endpoint concerns completion of all three training days, with a stated feasibility benchmark of > 75% attendance. The record also identifies stigma towards substance use and stigma towards depression as secondary outcomes. Those details define the questions the study was designed to address; they do not establish the answers.

The posted record does not report the primary attendance count or results for the stigma outcomes. It therefore does not show whether the feasibility threshold was met, whether stigma changed, or how precisely any such changes were estimated. Nor does the information reported establish whether patient re-engagement in care changed. Enrollment, design descriptions and safety counts cannot substitute for outcome estimates.

General statistical context: a result can be statistically uncertain even when its point estimate appears favorable, and a lack of a reported result is not the same as evidence of no effect. Interpretation requires outcome estimates, appropriate uncertainty measures and sufficient information about the analysis. For this cluster-randomized crossover study, readers would also need to know how clustering, sequence and period were handled. The registry record does not report those analysis details.

The zero adverse-event counts are limited to the stated event categories and sequence groups. They describe events recorded in the registry, but cannot establish that no events occurred outside the reported observation or that the intervention has no safety risks. Overall, the record supports a description of the research question and design, not a conclusion about effectiveness, feasibility achievement or the absence of harm.

14. Why This Trial Matters Statistically

This trial is statistically notable because the registry describes a cluster-randomized crossover design, while the primary endpoint is a straightforward participant count. The apparent simplicity of a count does not remove the need to account for how the study assigned conditions and organized observations. If outcomes are related within clusters, an analysis that ignores that structure may misstate uncertainty. In a crossover design, period and sequence can also matter. The record does not report how those features were handled.

The feasibility question is distinct from the study's broader motivation. Completion of all three training days measures participation in the training; stigma outcomes measure attitudes; patient re-engagement would be a separate outcome. Keeping these questions separate prevents an implementation measure from being interpreted as evidence of an attitude or care outcome. The registry record does not report results for the primary or secondary outcomes, so no such interpretation is supported.

The stated > 75% benchmark makes the primary question understandable, but a benchmark still requires an observed count and a defined denominator. The registry record reports neither the attendance result nor the analysis population. The enrollment figure alone cannot answer whether the benchmark was reached.

Finally, the record's safety counts are descriptive and its analysis results are absent. Transparent reporting of estimates, uncertainty, analysis populations, missing-data handling and methods for clustering and crossover structure would be needed for readers to assess the strength and precision of the findings. The statistical importance of this trial record is therefore partly in identifying the design and outcome questions, and partly in recognizing that the results needed to answer them are not reported.

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 main limitation for interpretation is that the registry record does not report the primary attendance result or any statistical analysis results. It does not give the number completing all three training days, the endpoint denominator, or results for the listed stigma outcomes. The study's stated feasibility criterion can therefore be described, but whether it was achieved cannot be assessed.

The design description is also incomplete for statistical interpretation. The record identifies cluster randomization and a crossover model but does not identify the cluster unit or number of clusters. It does not describe the randomization procedure, allocation concealment, sequence assignment details, outcome timing across periods, or methods for handling clustering, period effects or carryover. These are specific gaps for a study with this design.

The record lists seven outcome measures but does not report a multiplicity plan, missing-data methods, or analysis populations. It does not explain how incomplete attendance or stigma assessments were handled, and enrollment does not establish the denominator for any endpoint. The statistical analysis plan and the trial's own sample-size justification are not reported in the registry record.

For safety, the posted counts include no reported serious or other adverse events and no deaths in the listed sequences, but event definitions, follow-up completeness and comparative analysis are not reported. The approximate TAU periods differ across sequences, and the record does not specify enough about exposure timing to make a comparative safety interpretation. These limitations mean the registry description should not be treated as a complete account of outcome magnitude, precision or safety.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: Community Health Worker Training to Reduce Depression and Substance Use Stigma in TB/HIV Care in South Africa — NCT05282173.
• Registry record last updated 2025-03-06; status last verified 2025-02.
• Results first posted 2025-03-06.
• 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

NCT05282173 is a completed randomized, cluster-randomized crossover study of Siyakhana, a multi-day training intended to support CHWs working with patients in TB/HIV care who may experience depression or substance use. Enrollment was 82. The three sequences received approximately 1 Month, 2 Months or 3 Months of TAU before training. The primary endpoint is the number attending all three training days at 3-months post-training, with feasibility defined as > 75% attendance. Secondary outcomes concern CHW stigma towards substance use and depression.

The registry record does not report the primary endpoint result, secondary outcome results, or statistical analyses. It reports zero serious adverse events, zero other adverse events and zero deaths in each sequence group. Analysis methods, missing-data handling, analysis populations and the study's own sample-size justification are not reported.