Maternal Depression Treatment in HIV: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Maternal Depression Treatment in HIV (NCT03892915), a randomized phase 2, phase 3 study. The primary outcome is Rate of Maternal HIV Viral Suppression.

2. Clinical Question

3. Trial Design

The registry describes a randomized, parallel-group, cluster-randomized controlled trial comparing depression care with usual care. The trial took place at 8 antenatal care clinics in Uganda, with clinics as the stated setting for the cluster-randomized design. The registry enrollment is 391. It lists the phase as Phase 2 and Phase 3, and the status as completed. The recorded start date is 2019-07-08, and the primary completion date is 2023-08-31.

In the depression-care arm, the described approach included screening and psychoeducation, diagnosis, and evidence-based problem solving therapy or antidepressant therapy in specified circumstances. Usual care was described as referral to mental health specialists and access to a Family Support Group program. The intervention details are truncated in the registry record, so the full delivery protocol and any differences in service availability or uptake cannot be established from that description.

Masking is recorded as none. The record therefore does not describe a masked trial, but it does not specify which people involved in treatment delivery, outcome measurement, or analysis knew assignment. Those details are not reported in the registry record. The statistical method used for the posted primary comparison is also not reported. Because assignment was clustered by clinic, the sponsor's own analysis is the relevant reported analysis; a simple individual-level re-analysis would be misleading, as noted in the record.

The registry lists Depression Status as a secondary outcome, 6 outcome measures, and 5 analyses. It does not provide enough detail in the registry record to reconstruct the full analysis plan or the way each outcome measure and analysis relates to the primary question. The trial's stated primary endpoint was maternal HIV viral suppression two months post pregnancy.

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: Once enrolled, participants were automatically assigned to their group, given the cluster randomization (e.g., they received what their site was assigned to)

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 Rate of Maternal HIV Viral Suppression, measured over Two months post pregnancy.

The primary endpoint is the rate of maternal HIV viral suppression two months post pregnancy. The registry describes it as the percentage of participants with an undetectable HIV viral load measured by blood assay. It is therefore a binary outcome at the participant level: a participant either meets the stated suppression definition or does not. The registry identifies the unit as participants and the parameter as a count of participants, while the reported comparison is an odds ratio.

In general, an endpoint should represent an outcome that is relevant to the trial's question and can be measured in a consistent way. Viral suppression is relevant to the stated focus on adherence to the prevention of mother-to-child-transmission care cascade and pregnancy outcomes. A blood assay offers an objective measurement of viral load, although details of the assay, threshold for undetectability, testing procedures, and handling of unavailable measurements are not reported in the registry record.

The timing matters for interpretation: the recorded endpoint is assessed two months post pregnancy, not at an unspecified point during antenatal care or at delivery. The registry does not clarify the exact visit window or whether all participants were assessed at the same elapsed time. Nor does it report how the endpoint was handled for participants without a measurement. The posted odds ratio summarizes a comparison between groups, but the outcome description alone does not establish the direction of the group ordering used for that ratio.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

What does the odds ratio describe?

The reported effect measure is an odds ratio. In general, odds compare the chance of an event with the chance of its non-occurrence; an odds ratio compares those odds between groups. It is not the same as a risk ratio, and it should not be read directly as a percentage-point change or as a percentage change in probability. The registry reports an odds ratio of 1.19 for viral suppression, but the analysis group codes are not mapped to the named arms in the registry record. Without that mapping, the numerical direction cannot be assigned to depression care or usual care.

The outcome is described as the percentage of participants who achieved undetectable HIV viral load on a blood assay. However, the primary outcome event counts, group-specific proportions, and exact measurement threshold are not reported in the registry record. Those omissions prevent a reader from reconstructing the odds ratio from raw outcome counts or from translating it into absolute probabilities.

How should the confidence interval be read?

The 95% confidence interval is 0.73 to 1.93. In general, a confidence interval indicates the precision of an estimate under the assumptions of the analysis: a narrower interval indicates more precision, while a wider interval indicates less. A 95% interval is not a statement that there is a 95% probability that the true effect lies in this particular interval. Its formal interpretation depends on the statistical procedure and assumptions used.

For an odds ratio, 1 is the null value representing equal odds. The interval includes 1, so it spans values compatible with odds below, equal to, and above the null. This is why the estimate should not be described as establishing a clear direction. The registry does not report a p-value or the exact inferential procedure used to construct the interval. The analysis method is not reported in the registry record.

Why does the cluster design matter?

The trial is described as cluster randomized at 8 antenatal care clinics. General statistical explanation: participants within the same clinic can share providers, care processes, and local circumstances, so their outcomes may be correlated. Treating every participant as statistically independent can understate uncertainty when such correlation is present. Analysis should account for the unit of randomization and any within-clinic dependence. The registry specifically says that a simple individual-level re-analysis would be misleading and that only the sponsor's own analysis is shown.

The registry entries do not identify the sponsor's model, the way clustering was handled, any covariates, or the method for estimating uncertainty. Those methods are not reported in the registry record. It is therefore not possible to independently assess whether the interval accounts for all features of the design from the reported summary alone.

What can be said about re-calculation?

No recalculated individual-level estimate is presented. In general, re-calculations from summary statistics may be useful as checks when the design and available data support them, but they do not replace an analysis that respects cluster randomization. Here, the registry's stated reason for omitting a simple individual-level re-analysis is that it would be misleading. The posted sponsor estimate is the relevant result described in this account.

The record also includes planning assumptions for sample-size illustrations based on the observed effect, with a two-sided significance level of 0.05, equal allocation, and a normal approximation. Such planning illustrations concern hypothetical future study sizes; they are not observed power and do not change the uncertainty in the current trial's estimate. The numerical outputs of those illustrations are not included in the registry record, so no sample-size values can be stated.

What remains unknown about the analysis?

The registry records 6 outcome measures and 5 analyses, and names Depression Status as a secondary outcome. It does not specify the full set of models, contrasts, covariate adjustments, analysis populations, missing-data methods, or multiplicity procedures. The analysis method and p-value for the primary outcome are also not reported in the registry record. These details are important when assessing how the posted interval was produced and how strongly to interpret the estimate.

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("Depression Care", "Usual Care")
n      <- c(191, 200)
events <- c(128, 132)
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  Depression Care  191  128  67.0  59.9  73.6  60.1  73.3  60.1  73.3
ARM  Usual Care  200  132  66.0  59.0  72.5  59.2  72.2  59.2  72.2

11. Analyses Reported by the Sponsor

The registry reports the sponsor's own odds ratio for the primary outcome: 1.19, with a 95% confidence interval from 0.73 to 1.93. It does not report a recalculated estimate alongside it. The reason recorded is that the trial was cluster randomized and a simple individual-level re-analysis would be misleading, so only the sponsor's own analysis is shown.

This distinction matters because a calculation that treats participants as independent could fail to reflect the clinic-level assignment and within-clinic dependence. In general, a valid analysis should respect the design used to allocate treatment. The registry does not provide the statistical method or enough details to reproduce the sponsor's analysis independently. As a result, the sponsor's estimate is the reported basis for interpretation, while its specific modeling assumptions cannot be assessed from the registry record.

The registry entries do not include a second estimate against which to compare the sponsor's result, so no numerical disagreement or agreement can be described. Nor are the group codes in the reported analysis explicitly connected to the arm names in the registry record. The result should therefore be quoted as reported, without assigning the odds-ratio direction to a named arm or replacing it with an individual-level calculation.

12. Multiplicity, Missing Data and Analysis Populations

The registry lists 6 outcome measures and 5 analyses, and identifies Depression Status as a secondary outcome. These figures indicate that the record encompasses more than the single primary endpoint, but they do not establish how many statistical hypotheses were tested, which outcomes were analyzed together, or which comparisons were prespecified. The multiplicity procedure, if any, is not reported in the registry record.

In general, when investigators examine multiple outcomes or comparisons, the chance of a false-positive finding across the full set can increase if each is interpreted separately without adjustment. The appropriate approach depends on the scientific priorities and analysis plan. Because no multiplicity strategy or hierarchy of outcomes is described in the registry record, the primary result should be distinguished from secondary findings, and the record does not support reconstructing how the study handled multiple testing.

Missing-data information is also not reported in the registry record. The primary endpoint description identifies a blood assay but does not state how unavailable assays, missed visits, or incomplete outcome records were treated. The number of participants included in the primary analysis and the analysis population definition are not reported. It is not possible to determine from the posted summary whether the analysis followed an intention-to-treat approach, used another population, or applied imputation or other missing-data methods.

The registry does not report the statistical method, covariate adjustment, clinic-level variance approach, or the p-value for the primary comparison. Masking is recorded as none, but the record does not specify which roles were aware of treatment assignment. These omissions do not invalidate the reported estimate, but they limit independent assessment of the assumptions and analytic choices underlying it. More detail on the prespecified analysis plan and participant flow would be needed to evaluate these issues fully.

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 0% reporting threshold.

Most frequent serious adverse event terms

Serious events with the highest arm-level rate: death to fetus/newborn (9.0%).

The serious adverse event figures are reported separately for the Depression Care and Usual Care groups. The registry lists 13 serious events among 191 at risk in Depression Care and 18 among 200 at risk in Usual Care. It reports zero other events in each group and records deaths as 0 in Depression Care and 2 in Usual Care. These are descriptive counts from the registry; they do not establish that an event was caused by an intervention or that the groups differ in safety.

The listed top serious-event term is death to fetus/newborn, with 9.0% shown. The registry entries do not state the denominator or reporting basis for that percentage, so it should be quoted as listed rather than converted into a count or interpreted as a rate for a particular arm. Definitions of serious adverse events, ascertainment procedures, event timing, attribution, and whether the listed term overlaps with the death figures are not reported in the registry record.

In general, adverse-event counts are difficult to compare without information about exposure, observation time, definitions, and how events were collected. The listed at-risk figures describe the denominators in the adverse-event record, not necessarily the analysis population for the primary endpoint. These safety figures should therefore be read as registry-reported descriptive information, not as a causal comparison or a complete assessment of safety.

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

The posted result is an odds ratio of 1.19 for maternal HIV viral suppression two months post pregnancy, with a 95% confidence interval from 0.73 to 1.93. The interval includes the odds-ratio null value of 1 and extends in both directions. The data as reported therefore leave uncertainty about the direction and size of the comparison. The point estimate alone should not be treated as a definitive finding.

The estimate does not show the absolute proportion of participants who achieved suppression in either arm. The registry entries do not include group-specific primary outcome counts or percentages, and the analysis group codes are not explicitly linked to the arm names. The odds ratio cannot be translated into an absolute difference or a probability change from the information. Such a translation would require information that is not reported in the registry record.

The finding also does not establish that depression care caused a change in viral suppression. Although the study is described as randomized at the clinic level, causal interpretation depends on the analysis appropriately accounting for that design, as well as on outcome ascertainment, missing data, and other analytic details. The registry does not report the model or these supporting details. The sponsor's own result is reported because a simple individual-level re-analysis would be misleading.

Nor does this result establish a conclusion about every step of the PMTCT care cascade, pregnancy outcomes generally, or depression status. The primary endpoint concerns viral suppression at a specified post-pregnancy time. Depression Status is listed as a secondary outcome, but its result is not included in the registry record. The interpretation should remain limited to the posted primary comparison and its uncertainty, without extending it to outcomes that are not reported.

15. Why This Trial Matters Statistically

Statistically, this trial illustrates why the unit of randomization matters. The study is described as cluster randomized at antenatal care clinics, rather than as a simple allocation of each participant independently. In general, participants receiving care in the same clinic may have correlated outcomes. An analysis that ignores that structure can give misleading estimates of uncertainty. The registry explicitly warns against a simple individual-level re-analysis and presents the sponsor's own analysis.

The case also shows the value of reading an effect estimate together with its confidence interval. An odds ratio of 1.19 is not enough by itself to characterize the evidence. The interval from 0.73 to 1.93 includes the null value and spans effects in either direction, making uncertainty central to interpretation. A point estimate should not be mistaken for a precise or conclusive effect.

Outcome definitions and timing are equally important. The primary endpoint is not simply a general measure of health or adherence: it is undetectable HIV viral load measured by blood assay two months post pregnancy. That specificity clarifies what the reported comparison addresses, while the unreported assay threshold, missing-data handling, and analysis population limit independent interpretation.

Finally, the record distinguishes the primary outcome from a wider set of outcome measures and analyses, but does not report a multiplicity strategy. Together, these features make the trial a useful example of how design, endpoint definition, uncertainty, and incomplete analysis details shape what can responsibly be concluded from a registry result.

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 summary leaves important details unavailable for independent evaluation. The statistical method for the primary comparison is not reported, and the registry entries do not specify how the cluster-randomized design was modeled, whether covariates were included, or how uncertainty was estimated. The record notes that an individual-level re-analysis would be misleading, but it does not describe the sponsor's model in enough detail to reproduce it.

The primary outcome is defined as undetectable viral load by blood assay, but the assay threshold, testing procedures, visit window, and handling of missing or unavailable measurements are not reported in the registry record. Group-specific primary outcome counts and proportions are also absent. The analysis group codes are not connected to the named treatment arms in the registry record, which prevents assigning the direction of the reported odds ratio to either arm.

Participant eligibility criteria, baseline characteristics, clinic-level characteristics, and participant flow are not reported. The intervention description is truncated, so the complete delivery and implementation details cannot be assessed. Masking is recorded as none, but it is not clear which trial roles were aware of assignment.

The registry lists multiple outcome measures and analyses, but does not report the full prespecified analysis plan or a multiplicity approach. It also does not state the analysis population or missing-data method. These limits do not alter the reported estimate, but they constrain judgments about its precision, robustness, and scope. The serious adverse event summary likewise lacks definitions, attribution, observation-time detail, and the denominator for the listed 9.0% serious-event term.

17. Related Tutorials

18. Related Statistical Calculators

19. Sources

• ClinicalTrials.gov: Maternal Depression Treatment in HIV — NCT03892915.
• Registry record last updated 2024-11-15; status last verified 2024-09.
• Results first posted 2024-11-15.
• 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

NCT03892915, titled Maternal Depression Treatment in HIV, is recorded as a completed randomized, parallel-group, cluster-randomized trial. It lists Phase 2 and Phase 3, an enrollment of 391, and 8 antenatal care clinics in Uganda. The recorded start date is 2019-07-08, and primary completion is 2023-08-31. Masking is listed as none.

The comparison was task-shifted depression care versus usual care. The primary endpoint was maternal HIV viral suppression, defined as undetectable viral load by blood assay two months post pregnancy. The sponsor-reported odds ratio is 1.19, with a 95% confidence interval from 0.73 to 1.93; a p-value is not reported. The registry notes that a simple individual-level re-analysis would be misleading because of the cluster-randomized design.

Depression Status is listed as a secondary outcome. The record also reports serious adverse event counts and deaths, but does not include the details needed to establish attribution or reconstruct the primary analysis independently.