Traditional Healers as Adherence Partners for Persons Living With HIV in Rural Mozambique: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Traditional Healers as Adherence Partners for Persons Living With HIV in Rural Mozambique (NCT03076359), a non-randomized na study. The primary outcome is Retention in Care.

2. Clinical Question

The clinical question is about a support program added to HIV care, rather than a comparison of ART medications. The registry describes the intervention as behavioral and the comparison as standard care. The study was non-randomized, so group membership may have reflected factors other than the intervention. Differences between groups therefore cannot automatically be attributed to healer support. The registry record does not report how participants entered each group, the baseline characteristics needed to assess their comparability, or adjusted effect estimates. It also does not report numerical results for adherence, retention, or viral load.

3. Trial Design

This was a completed, parallel-group, non-randomized study sponsored by Vanderbilt University Medical Center. The registry lists enrollment as 320, the start date as 2015-06-10, and primary completion as 2019-05-30. The phase is listed as NA. Masking is listed as none. The condition is HIV/AIDS.

The two registered groups were Standard of Care and the Traditional Healer Support Program. The latter was described as standard care plus a partnership with a trained traditional healer, including community- and clinic-based support. The Standard of Care description includes HIV treatment and ART medications. Some arm-description text is incomplete in the registry record, so the exact activities and treatment details cannot be fully reconstructed from it.

Non-randomized allocation matters for interpretation. Random assignment can, in general, help balance both measured and unmeasured characteristics across groups; without it, group differences may reflect participant characteristics, site or community circumstances, or other factors as well as the intervention. The registry itself cautions that the arms are not a like-for-like control comparison. It does not report enough detail to determine what methods, if any, were used to address confounding.

The stated project goal concerned adherence, retention, and viral load among people newly initiated on ART in rural Mozambique. Retention in care was the primary outcome, described as the percentage of medication coverage over 12 months. Viral load appears as a secondary outcome. Although the record lists 2 outcome measures, it reports no numerical result for either outcome. The statistical analysis plan, any interim analyses, and the sample-size justification for the original study are not reported in the registry record.

4. Participant Flow and Accounting

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

Overall Study

Analysis population for the primary outcome, as defined in the registry: Several patient ID numbers were not able to be matched to online database of patient records at this point give discrepancies in name or community. Given new paper records were introduced in 2019 we were not able to go back to the original records for cross-reference of medication pick-up.

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 Retention in Care, measured over Retention in care over 12 months.

The primary endpoint is retention in care over 12 months, expressed as the percentage of days on which a participant had medication. The registry gives an example: a participant with a 30 day supply collected on August 1st, who did not collect the next supply until October 1st, would be considered retained for 50% of the time. This illustrates that the measure reflects medication availability across follow-up, rather than simply whether a participant attended a visit at a particular moment.

Medication coverage is a way to operationalize continuity of access to treatment. It can capture gaps between medication collections that a simple visit count might not describe. It does not, by itself, establish that medication was taken as prescribed. It is also dependent on having medication-pick-up records that can be linked to the participant and follow-up period.

The registry describes the primary outcome using a median and an inter-quartile range. A median gives the middle value in an ordered set of participant measurements, while the inter-quartile range describes the spread of the middle half of those values. These summaries can be useful when participant measurements are skewed or include unusually high or low values. The numerical median and inter-quartile range for the study groups are not reported in the registry record, so the size and precision of any group difference cannot be assessed from the registered results.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The statistical methods that were used to analyze the clinical outcomes are not reported in the registry record. The endpoint description does indicate that retention was summarized using a median and an inter-quartile range. The registry also lists a secondary outcome of viral load, but does not provide its measurement details, summary statistics, or analysis method. The concepts below explain how such methods are generally interpreted; they do not imply that a particular analysis was performed for this study.

What does the median and inter-quartile range describe?

For a set of participant-level measurements, the median is the middle value after ordering the measurements. The inter-quartile range describes the span between the lower and upper quartiles, giving a sense of the spread among the middle portion of the observations. These summaries are often useful when a distribution is asymmetric or contains extreme values, because the median is less influenced by extremes than the arithmetic mean. They describe the distribution within a group; by themselves, they do not quantify a treatment effect or establish a between-group difference.

For this endpoint, each participant’s measure was intended to reflect the percentage of days with medication coverage during 12 months. A group median would describe a typical participant measurement in that group, while the inter-quartile range would show how varied participant measurements were. The numerical summaries are not reported in the registry record, and no group contrast is posted. As a result, neither the typical level nor the spread can be compared between groups from the registry results.

Why does non-randomized allocation matter?

In general, randomization aims to make comparison groups similar on average before treatment, including with respect to characteristics that may not have been measured. Non-randomized studies do not obtain that protection from allocation. If factors related to both group membership and retention also differ across groups, a simple comparison can mix the intervention contrast with those pre-existing differences. Statistical adjustment can sometimes address measured differences, but it cannot automatically remove bias from unmeasured factors. The registry record does not report a model, covariate adjustment, matching strategy, or other method for addressing confounding.

The record explicitly warns that the arms are not a like-for-like control comparison. This means that any numerical contrast, if reported elsewhere, would need to be interpreted with attention to how participants came to receive each approach, the groups’ starting characteristics, and how outcomes were recorded. Those details are not reported in the registry record. The parallel model indicates that the study compared groups during the same general study rather than describing a crossover sequence, but the registry does not give further analysis details.

What would a confidence interval contribute?

As a general statistical concept, a confidence interval presents a range of values compatible with the data and the specified statistical method. A narrow interval indicates greater precision than a wide interval, while an interval spanning values in different directions may leave the direction of an effect uncertain. Its interpretation depends on the design and assumptions used to calculate it. No confidence interval for retention or viral load is reported in the registry record. Thus, the precision of any outcome contrast cannot be assessed.

What can be said about significance testing?

A p-value is a statistical summary that, under a specified model and null hypothesis, measures how unusual the observed data would be if that hypothesis were true. It is not the probability that the null hypothesis is true, and it does not measure clinical importance. Any test also depends on the analysis method and assumptions. The registry record does not report a test, p-value, or statistical model for the primary outcome. It is therefore not possible to characterize a result as statistically significant or not based on the registry record.

What do the planning illustrations represent?

The sample-size assumptions describe planning illustrations based on the observed effect, not the trial’s own design. They specify a two-sided significance level of 0.05, equal allocation between the two arms, a normal approximation, and an effect to detect expressed as a stated fraction of the observed effect. For non-time-to-event outcomes, the illustration is expressed as participants per arm; time-to-event illustrations use a total number of events. These are hypothetical planning calculations and are not observed power. They do not establish what sample-size justification the investigators used before conducting the study. That original justification and the underlying outcome effect are not reported in the registry record.

The registry lists no analyses. That listing does not explain whether analyses were conducted but not posted, or whether other reports contain them. It only means that an analysis result is not recorded in this registry entry. The analysis plan, handling of repeated or missing measurements, and any model assumptions are not reported in the registry record.

10. Multiplicity, Missing Data and Analysis Populations

The record lists 2 outcome measures, identifies retention in care as the primary outcome, and lists viral load as a secondary outcome. Multiple outcomes can create a multiplicity issue when many statistical tests are interpreted together: the more tests considered, the greater the chance that at least one appears notable by chance. Statistical plans may address this in different ways, for example by defining a primary outcome in advance or adjusting inference across tests. The multiplicity strategy for this study is not reported in the registry record. No numerical test results are reported for either outcome.

The registry notes that several patient ID numbers could not be matched to the online patient-record database because of discrepancies in names or communities. It also states that new paper records introduced in 2019 could not be used to cross-reference original records. This is specifically relevant to the medication-pick-up information used to construct the retention measure. The number of participants affected, the group distribution of unmatched records, and the approach to missing or unmatchable data are not reported in the registry record.

Without those details, it is not possible to determine whether the reported outcome population, if one exists outside the registry, included all enrolled participants or only those with matchable records. Nor can the potential effect of missing record information on the retention estimates be assessed. The registry does not report an intention-to-treat analysis population, a per-protocol population, or other analysis-population definitions. Methods for handling missing viral-load measurements are also not reported in the registry record.

The study is described as completed, but its statistical analysis plan, any interim analyses, and the handling of multiplicity and missing outcomes are not reported in the registry record. These omissions limit assessment of whether the outcome definitions and analyses followed a prespecified plan.

11. 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.

The registry lists serious adverse event counts as 0 among 170 participants at risk in the Standard of Care group and 0 among 150 participants at risk in the Traditional Healer Support Program group. It also lists other adverse event counts as 0 among 170 and 0 among 150, respectively. Deaths are recorded as 14 in Standard of Care and 15 in the Traditional Healer Support Program group.

These counts should be read as separate registry fields. The record does not define the serious or other adverse-event categories in the information available, describe event ascertainment, or state whether the deaths were included in either event category. It also does not report causes, timing, attribution to the intervention, or participant-level details. The death figures therefore should not be interpreted as evidence that the program caused or prevented deaths, and the zero event counts should not be taken to establish that no harms occurred outside the registry’s recorded categories.

The comparison is non-randomized, and the registry cautions that the arms are not a like-for-like control comparison. Differences in group counts cannot be interpreted as causal safety effects without additional design and analysis information. No further serious adverse event terms are listed in the registry record.

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

The registry record establishes that a study of traditional healer support alongside standard HIV care was conducted and marked completed. Its stated aim was to assess adherence, retention, and viral load in people newly initiated on ART in rural Mozambique. It identifies retention in care over 12 months, measured through medication coverage, as the primary outcome.

What the record does not establish is the size or direction of any outcome difference. Numerical results for retention and viral load are not reported in the registry record. There are no posted group estimates, effect estimates, confidence intervals, or p-values for those outcomes. Consequently, the recorded information does not show whether the groups had similar or different outcomes, or how precisely any estimate was determined.

Even if a simple outcome contrast were available, the non-randomized allocation would limit causal interpretation. The record itself cautions that the arms are not a like-for-like control comparison. Without information about baseline group characteristics and the analysis used to account for differences, a contrast could reflect factors beyond the support program. The registry also reports challenges matching some participant IDs to medication records, while leaving the number affected and the handling of missing data unspecified.

The sample-size calculations associated with the study are planning illustrations based on the observed effect, not evidence of the trial’s original design or observed power. They cannot substitute for outcome estimates or a clear analysis plan. Likewise, the recorded adverse-event and death counts should not be interpreted as causal safety comparisons in this non-randomized study.

Accordingly, the defensible interpretation is limited: the registry describes the intended intervention, comparison, and primary outcome, but does not contain the numerical clinical results needed to quantify the findings. The absence of results in this registry entry is not evidence that the intervention had no effect, and it is not evidence that an effect occurred. Other reports or records may contain information, but they are not described in this registry record.

13. Why This Trial Matters Statistically

This study raises a statistical question about how to evaluate a community-based support program alongside ongoing HIV treatment. Its primary measure focuses on medication coverage over time, a dimension of continuity that differs from a one-time clinic attendance measure. Describing the operational definition matters because conclusions about retention depend on what the endpoint actually captures and on the quality of the records used to construct it.

The non-randomized design also highlights a central challenge in evaluating real-world support programs. Groups can differ for reasons tied to both participation and outcomes, making an unadjusted comparison hard to interpret causally. The registry’s warning that the arms are not a like-for-like control comparison is important context for any future reading of results. Clear reporting of group characteristics, the allocation process, and adjustment methods would help readers judge the strength of a comparison.

Record linkage is another statistically important feature of this study. The primary outcome relies on medication-pick-up information, while the registry notes that some patient IDs could not be matched and that later paper records could not be cross-referenced. Missing or unlinked records can affect both the number of participants contributing outcome data and the representativeness of those participants. Reporting the extent and handling of these data gaps would be necessary to assess potential bias.

Finally, the distinction between a completed study and a study with posted results is consequential. The registry describes the design and outcomes, but numerical retention and viral-load results are not reported in the registry record. Without those results, readers cannot evaluate the magnitude, variability, or precision of the outcomes. This makes transparent reporting of estimates, uncertainty, analysis populations, and missing-data methods especially important for a non-randomized study.

14. 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.

Several limitations are specific to the information recorded for this study. The primary outcome depends on linking patient identifiers to medication-pick-up records, yet the registry states that several IDs could not be matched because of name or community discrepancies. It also notes that paper records introduced in 2019 could not be used to cross-reference the original records. The number of participants affected, the reasons for each unmatched record, and the handling of missing information are not reported in the registry record.

The descriptions of the intervention and standard-care arm are incomplete. The support program is described as a partnership with a trained traditional healer and community- and clinic-based support, but the full description of activities is not present. The Standard of Care description is also truncated. This limits the ability to understand exactly what was delivered in each group or how consistently the program was implemented.

Non-randomized allocation and the registry’s explicit warning that the groups are not a like-for-like control comparison constrain interpretation. Baseline characteristics and any adjustment for group differences are not reported in the registry record. The primary outcome and viral-load estimates, their uncertainty, and the analyses used to produce them are also not reported. The statistical analysis plan, multiplicity approach, analysis populations, and interim analyses are not reported in the registry record.

Safety information is limited to event counts and deaths. Definitions, timing, attribution, and causes are not reported in the registry record. These omissions prevent a detailed assessment of either the outcome comparison or safety reporting from this entry alone.

15. Related Tutorials

16. Related Statistical Calculators

17. Sources

• ClinicalTrials.gov: Traditional Healers as Adherence Partners for Persons Living With HIV in Rural Mozambique — NCT03076359.
• Registry record last updated 2020-09-29; status last verified 2020-09.
• Results first posted 2020-09-29.
• 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.

18. Record Summary

NCT03076359 is a completed, non-randomized, parallel-group study sponsored by Vanderbilt University Medical Center. The registry lists enrollment of 320, a start date of 2015-06-10, and primary completion on 2019-05-30. Masking is listed as none, and the phase is listed as NA.

The study compared Standard of Care with the Traditional Healer Support Program, which added support from a trained traditional healer. Retention in care over 12 months, measured as percentage of medication coverage, was the primary outcome; viral load was listed as a secondary outcome. Numerical results for these outcomes are not reported in the registry record.

The record notes challenges matching some patient IDs to medication records and cautions that the study arms are not a like-for-like control comparison. It lists serious and other adverse event counts of 0 in each group, alongside recorded deaths of 14 in Standard of Care and 15 in the Traditional Healer Support Program group. The registry does not provide additional detail needed to interpret those safety counts causally.