This page provides an independent statistical analysis and educational interpretation of publicly reported results. ClinicalTrials.gov provides the official trial registry record. Numerical trial results on this page are restricted to the results posted in the ClinicalTrials.gov record; calculations added by Clinical Biostats are labelled as such.
1. Trial at a Glance
This page covers the registry results for Mobile Technology to Extend Clinic-based Counseling for HIV+s in Uganda (NCT03928418), a randomized na study. The primary outcome is Change in Alcohol Use Measured by Self Report.
| Feature | Registry value |
|---|---|
| ClinicalTrials.gov identifier | NCT03928418 |
| Phase | NA |
| Condition | Alcohol Use, Unspecified, HIV/AIDS |
| Allocation | Randomized |
| Design model | Parallel |
| Masking | Single |
| Enrollment | 272 |
| Arms in the primary outcome | n/a |
| Lead sponsor | University of California, San Francisco |
| Results posted | Yes |
2. Clinical Question
Population
HIV-positive persons in Uganda with unhealthy alcohol use were the population of interest. The registry lists alcohol use and HIV/AIDS as conditions. The primary outcome analysis population is described as participants with self-reported alcohol-use data at the 6-month or 9-month visit; the record does not give the number analyzed for that outcome.
Intervention
One intervention combined in-person counseling during 2 quarterly clinic visits with live booster phone calls every three weeks in the interim. The other combined the same in-person counseling schedule with technology boosters, using the participant’s choice of SMS or IVR once to twice weekly in the interim.
Comparator
The comparison group received standard of care: brief, unstructured advice, with a wait-listed intervention. The registry describes the study as randomized and parallel-group, with 3 arms.
Primary question
The primary measure was change in self-reported drinking days, measured as the number of days drinking in the prior 21 days using the alcohol use timeline follow-back. It was assessed at the 6-month and 9-month visits, described as 3 and 6 months post intervention. The registry record does not report the arm-specific estimates or the statistical comparisons, so the size and precision of any between-group differences cannot be described from this record.
In practical terms, the study set out to compare two ways of extending clinic-based counseling between visits with standard care. The primary endpoint focused on reported drinking days rather than on booster use or satisfaction. Secondary outcomes included unhealthy alcohol use measured by AUDIT-C, heavy drinking days, CD4 cell count, ART adherence, booster completion, counselor call-back requests, and booster satisfaction and usability. The registry record names these measures but does not report numerical results for them.
3. Trial Design
The EXTEND study is recorded as a randomized controlled trial with a parallel-group model and single masking. It enrolled 272 participants, and its status is completed. The record lists the study start as 2019-09-19 and primary completion as 2021-08-30. The registry does not identify who was masked, so the masking designation cannot be interpreted more specifically from this record.
There were 3 study arms. Both active approaches included in-person counseling during 2 quarterly clinic visits. One added live booster phone calls every three weeks between those visits. The other added technology boosters, by SMS or IVR, once to twice weekly in the interim. The standard-of-care arm received brief unstructured advice and a wait-listed intervention. The record does not provide further detail about the counseling content, delivery procedures, or the timing and conditions of the wait-list intervention.
The primary measure was assessed at the 6-month and 9-month visits. The registry describes these as 3 and 6 months post intervention. Its description of the analysis population refers to participants with self-reported alcohol-use data at either visit. It does not provide the number analyzed at each visit or explain how data from the two visits were combined or compared.
The record lists 13 outcome measures and 12 analyses. It does not describe the allocation procedure beyond stating that the trial was randomized, nor does it report details such as allocation concealment, stratification, or the identity of masked parties. These design details are not reported in the registry record. The study is listed with phase NA.
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: The first 3 individuals enrolled in the study were considered "pilot" participants. They were randomized to one of the three study arms separately from the randomization for the main study participants. They are excluded from all analyses.
Overall Study
| Stage | Live Phone Call Booster Arm | Technology Booster Arm | Standard of Care (SOC) Arm |
|---|---|---|---|
| STARTED | 91 | 91 | 90 |
| Pilot Participant | 1 | 1 | 1 |
| 6-month Visit | 83 | 86 | 85 |
| COMPLETED | 85 | 87 | 84 |
| NOT COMPLETED | 6 | 4 | 6 |
Reasons for not completing this period:
| Reason | Live Phone Call Booster Arm | Technology Booster Arm | Standard of Care (SOC) Arm |
|---|---|---|---|
| Lost to Follow-up | 5 | 3 | 5 |
| Pilot participant - excluded from analyses | 1 | 1 | 1 |
Analysis population for the primary outcome, as defined in the registry: The overall number of participants analyzed is the number of participants with self-reported alcohol use data at 6 months or 9 months.
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: (1 person was found to have been enrolled in the SOC arm twice \[protocol violation\]; they are included only once.)
| Characteristic | Unit | Live Phone Call Booster Arm | Technology Booster Arm | Standard of Care (SOC) Arm | Total |
|---|---|---|---|---|---|
| Participants analysed | 90 | 90 | 89 | 269 | |
| Age, Categorical: <=18 years | Participants | 0 | 0 | 0 | 0 |
| Age, Categorical: Between 18 and 65 years | Participants | 89 | 90 | 85 | 264 |
| Age, Categorical: >=65 years | Participants | 1 | 0 | 3 | 4 |
| Age, Continuous | years | 40.0 (34.0 to 47.0) | 39.5 (32.0 to 46.0) | 38.5 (33.0 to 46.0) | 39.5 (33.0 to 46.0) |
| Sex: Female, Male: Female | Participants | 33 | 30 | 30 | 93 |
| Sex: Female, Male: Male | Participants | 57 | 60 | 59 | 176 |
| Race and Ethnicity Not Collected | Participants | — | — | — | 0 |
| Region of Enrollment: Uganda | participants | 90 | 90 | 89 | 269 |
6. Primary Endpoint
The registry lists the primary outcome as Change in Alcohol Use Measured by Self Report, measured over At 6 and 9 month visits (3 and 6 months post intervention)..
| Element | Registry entry |
|---|---|
| Result type | Mean |
| Measure of spread | 95% Confidence Interval |
| Unit | drinking days |
| Endpoint classification | Continuous Endpoint |
The primary endpoint was change in alcohol use measured by self-report. Specifically, it was the number of days drinking in the prior 21 days, reported using the alcohol use timeline follow-back. The measure was assessed at the 6-month and 9-month visits, which the registry describes as 3 and 6 months post intervention. The stated summary measure is the mean, with a 95% confidence interval.
General explanation: a count of drinking days gives a direct measure of how often alcohol use occurred during a defined recall window. A mean summarizes the average number of reported drinking days in a group, while a confidence interval describes statistical uncertainty around an estimated quantity. Those concepts do not establish that self-report is error-free or that a particular change is clinically important.
The registry record does not report the observed means, changes, confidence intervals, or between-arm contrasts for this endpoint. It also does not explain how the two follow-up visits were represented in the primary analysis. Consequently, the measure and its timing can be described, but the direction, magnitude, and precision of the primary outcome findings cannot be assessed from the registry record.
7. Why No Re-analysis Is Shown
8. Statistical Interpretation
The registry record identifies the primary endpoint and its follow-up schedule, but does not report the numerical primary outcome results. There are no arm-specific means, changes from baseline, confidence intervals, p-values, or stated comparisons in the record details. As a result, the record does not establish whether the reported drinking-day measure differed between either counseling-plus-booster approach and standard of care, or between the two active approaches.
The endpoint was measured at both the 6-month and 9-month visits. The record does not say whether each visit was analyzed separately, whether a combined estimate was used, or how any repeated measurements were handled. It also does not provide visit-specific analysis counts. The analysis-population description says that participants with self-reported alcohol-use data at 6 months or 9 months were analyzed, but it does not give the number of such participants or further define the population.
General explanation: an estimated difference describes the observed contrast between groups, while its confidence interval conveys how uncertain that estimate is under the statistical model. A confidence interval should be considered together with the estimate and the design; a p-value alone would not show the magnitude or practical importance of a difference. None of those primary numerical quantities is reported in the registry record, so no such interpretation is possible.
The posted safety counts are separate from evidence about the primary alcohol-use endpoint. They record one serious adverse event in the live phone call booster arm, none in the technology booster arm, and none in the standard-of-care arm. The listed serious event term is suicidal ideation at 1.1%. These counts should not be treated as a comparison of efficacy or as proof that an arm caused or prevented an event.
The record includes a planning illustration based on the observed effect and specified assumptions, but planning calculations are not observed power and do not replace trial results. A sample-size illustration cannot show what happened to participants or establish that an outcome difference was detected. The underlying primary estimates and the sponsor’s analysis are not reported in the registry record.
9. Statistical Methods Explained
Randomization is a method for assigning participants to study arms by chance. General explanation: when conducted as intended, random allocation can reduce systematic differences between groups at entry, although it does not guarantee that every characteristic will be balanced. The registry identifies the trial as randomized, but it does not describe the allocation sequence, concealment, or any stratification procedure.
What does a parallel-group design mean?
In a parallel-group design, participants are assigned to an arm and receive that arm’s assigned approach. Here, the registry describes 3 arms: live phone call boosters, technology boosters, and standard of care. The two active approaches both included in-person counseling during 2 quarterly clinic visits; their booster delivery differed. The technology approach allowed SMS or IVR boosters once to twice weekly in the interim, whereas the live-call approach used calls every three weeks. These are descriptions of assigned approaches, not evidence that participants completed every planned contact.
How should a mean and confidence interval be read?
The primary measure is summarized as a mean number of drinking days, with a 95% confidence interval. General explanation: the mean is an average, and the interval conveys uncertainty about an estimate under the analysis assumptions. A narrower interval generally indicates greater precision than a wider interval, but interpretation also depends on the estimated effect and the clinical context. The record does not report the endpoint’s means or confidence intervals, so their precision cannot be evaluated.
The outcome counts drinking days during the prior 21 days, but the registry labels the primary measure as continuous and specifies a mean. General explanation: a statistical model should be appropriate for the outcome’s scale and the way observations were collected. The record does not state which model was used, whether the analysis adjusted for baseline values, or whether it accounted for repeated measurements at the 6-month and 9-month visits. Those analysis details are not reported in the registry record and should not be inferred from the endpoint description.
What does the planning illustration mean?
The record describes a planning illustration based on the observed effect, with a two-sided significance level of 0.05, equal allocation between two arms, and a normal approximation; the effect to detect is described as a stated fraction of the observed effect. For non-time-to-event calculations, the illustration is expressed as participants per arm for 80% or 90% power. These are planning quantities, not observed trial outcomes and not observed power. The record does not provide the trial’s own sample-size justification, so the illustration should not be presented as the study’s original design calculation.
General explanation: power is a design-stage probability of detecting a specified effect under stated assumptions if that effect is real. It is not the probability that a result is true, and a post hoc calculation based on an observed effect does not reveal whether the trial was well designed. The posted illustration should therefore be read only as a conditional planning exercise. The numerical illustration values themselves are not reported in the registry record.
What can the registered analyses establish?
The record lists 13 outcome measures and 12 analyses, but does not describe a complete statistical analysis plan. It does not state which contrasts were primary beyond naming the primary endpoint, how the two follow-up time points were handled, or whether adjustments were made for multiple comparisons. It also does not report model diagnostics, missing-data methods, or adjusted and unadjusted estimates. These omissions limit any attempt to reproduce or assess the statistical analysis.
General explanation: estimates should be interpreted in light of the outcome definition, analysis population, model, uncertainty, and how many comparisons were considered. Without the relevant estimates and methods, a reader cannot verify the precision of the findings or distinguish a prespecified analysis from exploratory work. The registry record supports a description of the design and measures, but not a statistical conclusion about the primary endpoint.
10. Multiplicity, Missing Data and Analysis Populations
The registry lists 13 outcome measures and 12 analyses. The primary outcome has more than one follow-up time point, at the 6-month and 9-month visits, and the record notes that it was not reduced to a single comparison. It does not report how multiple outcomes, time points, or pairwise group comparisons were handled. A multiplicity adjustment plan, if any, is not reported in the registry record.
General explanation: when several outcomes or comparisons are examined, the chance of at least one apparently unusual result can increase. Adjustment procedures can address some forms of this issue, but their appropriateness depends on which comparisons were planned and how they relate to one another. Without a stated analysis plan or results, the extent to which multiplicity affects interpretation cannot be assessed.
For the primary endpoint, the registry describes the analysis population as participants with self-reported alcohol-use data at 6 months or 9 months. It does not report the number analyzed, the amount or pattern of missing outcome data, reasons for missingness, or how missing values were handled. It also does not say whether the same participants contributed data at both visits.
General explanation: analyses based on available observations may be affected if missingness is related to participants’ outcomes or characteristics. Methods such as imputation or likelihood-based approaches require assumptions, and the assumptions and methods should be stated to assess their plausibility. Those details are not reported in the registry record. The registry’s description therefore identifies an outcome-data-based analysis population but does not support a judgment about missing-data bias.
The record does not provide a fuller definition of the analysis population, such as whether all randomized participants were included regardless of adherence. That information is not reported in the registry record.
11. Safety Results
The registry reports serious adverse events as affected participants divided by participants at risk.
| Arm | Participants with serious adverse events | Participants at risk | Percentage |
|---|---|---|---|
| Live Phone Call Booster Arm | 1 | 90 | 1.1% |
| Technology Booster Arm | 0 | 90 | 0.0% |
| Standard of Care (SOC) Arm | 0 | 89 | 0.0% |
Deaths
| Arm | Deaths (participants affected of at risk) |
|---|---|
| Live Phone Call Booster Arm | 0 of 90 |
| Technology Booster Arm | 0 of 90 |
| Standard of Care (SOC) Arm | 0 of 89 |
Other (non-serious) adverse events
Participants with at least one non-serious event at or above the registry's 0% reporting threshold.
| Arm | Participants affected of at risk |
|---|---|
| Live Phone Call Booster Arm | 0 of 90 |
| Technology Booster Arm | 0 of 90 |
| Standard of Care (SOC) Arm | 0 of 89 |
Most frequent serious adverse event terms
| Event term | Organ system | Live Phone Call Booster Arm | Technology Booster Arm | Standard of Care (SOC) Arm |
|---|---|---|---|---|
| Suicidal ideation | Psychiatric disorders | 1 of 90 | 0 of 90 | 0 of 89 |
Serious events with the highest arm-level rate: suicidal ideation (1.1%).
The registry reports serious adverse event counts by arm. In the live phone call booster arm, 1 of 90 participants at risk had a serious adverse event; the listed serious term was suicidal ideation at 1.1%. In the technology booster arm, 0 of 90 participants at risk had a serious adverse event. In the standard-of-care arm, 0 of 89 participants at risk had a serious adverse event. Deaths were recorded as 0 in each arm.
These figures describe reported counts in the registry. They do not by themselves show whether an event was caused by an intervention, whether the risk differs between arms, or how precisely any such difference could be estimated. The registry record does not give additional event narratives, timing, relatedness assessments, or a comparative statistical analysis. It also does not report non-serious adverse-event results in the registry record for this summary.
General explanation: safety counts should be read with their definitions, observation periods, and participant denominators in mind. Small numbers of reported events can provide limited information about uncommon outcomes, and a count of zero does not establish that an event cannot occur. These data are a descriptive safety account, not a basis for medical advice or a conclusion about treatment effect.
12. What the Result Does, and Does Not, Mean
The registry record establishes that EXTEND was a completed randomized, parallel-group study of counseling delivery approaches for HIV-positive persons in Uganda. It describes the active approaches, standard-of-care comparator, primary self-reported drinking-day measure, follow-up visits, secondary outcome topics, and serious adverse event counts. Those details allow readers to understand what the study was designed to compare.
The record does not include numerical results for the primary outcome or the listed secondary outcomes. It therefore cannot show the amount or direction of any change in drinking days, whether the groups differed, or how precise an estimate might be. It also does not report an outcome-based conclusion about booster uptake, ART adherence, CD4 cell count, or satisfaction. The existence of an outcome measure is not evidence of its result.
The safety record gives descriptive serious adverse event counts, including one serious event in the live phone call booster arm and none in the other arms. Without further event details or a comparative analysis, these counts cannot establish causation or comparative safety. Similarly, the registry’s planning illustration is not an observed efficacy finding and should not be interpreted as evidence that a treatment worked or failed.
General explanation: a randomized design can support comparisons when outcomes, analysis populations, and statistical methods are reported clearly. The design label alone does not supply the missing estimates, confidence intervals, or analysis details. The registry record should therefore be read as a description of the trial and selected safety counts, rather than as a complete report of the primary statistical findings.
13. Why This Trial Matters Statistically
This trial is statistically notable because it compares two ways to deliver boosters alongside the same described clinic-counseling schedule, as well as a standard-of-care control. That structure distinguishes the delivery approach of interest: live phone calls every three weeks versus SMS or IVR boosters once to twice weekly in the interim. The record also makes clear that the primary endpoint was self-reported drinking days, not booster engagement or satisfaction.
The endpoint was planned for two follow-up visits, the 6-month and 9-month visits. Multiple time points can provide different views of an outcome over follow-up, but interpretation depends on how those measurements are analyzed and whether the comparisons were prespecified. The registry record does not report that analysis strategy or the primary estimates, so readers cannot determine whether the study found a difference or how uncertain it was.
The trial also lists outcomes spanning alcohol use, HIV-related measures, adherence, and booster experience. General explanation: a broad outcome set can describe several dimensions of an intervention, but it makes transparent reporting of the primary question, analysis population, and multiplicity approach especially important. The record lists 13 outcome measures and 12 analyses but does not describe the handling of multiple comparisons.
For biostatistical interpretation, the record illustrates the distinction between a completed randomized design and a reportable result. Enrollment and design information establish the trial framework; they do not substitute for estimated effects, confidence intervals, and analysis methods. The missing outcome results leave the principal statistical question unresolved in the registry record.
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.
The principal limitation for interpreting the trial statistically is that the registry record does not report numerical results for the primary endpoint. Arm-specific means, changes, confidence intervals, and p-values are not reported. The record also does not give numerical results for the secondary outcomes listed. Accordingly, effect size, direction, and precision cannot be evaluated.
The primary measure was assessed at two visits, but the analysis of those time points is not described. The record does not state whether results were estimated separately by visit or combined, how repeated observations were handled, or which group contrasts were tested. The masking designation is single, but the identity of the masked parties is not reported. Details of randomization implementation are also not reported.
The primary analysis population is described in terms of available self-reported alcohol-use data at the 6-month or 9-month visit. The record does not give the number analyzed, missing-data amounts or reasons, or the methods used to address missing values. It does not define an intention-to-treat population or explain adherence-based exclusions, if any. Those details are not reported in the registry record.
The record lists multiple outcome measures and analyses without describing a multiplicity strategy or full analysis plan. It also does not provide the trial’s own sample-size justification. The planning illustration is explicitly based on the observed effect and stated assumptions, so it should not be treated as an original design rationale or as observed power. Safety information is limited to the listed serious-event counts and term; further event context is not reported.
15. Related Tutorials
16. Related Statistical Calculators
17. Sources
- ClinicalTrials.gov: Mobile Technology to Extend Clinic-based Counseling for HIV+s in Uganda — NCT03928418.
- Registry record last updated 2023-09-14; status last verified 2023-09.
- Results first posted 2023-09-14.
- 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.
Continue through the Clinical Biostats statistical pathway
Use this trial's endpoint as a starting point for deeper study of effect measures, confidence intervals, and sample-size planning.
18. Record Summary
EXTEND is listed as a completed randomized, single-masked, parallel-group trial with enrollment of 272. It compared live phone call boosters, technology boosters by SMS or IVR, and standard of care among HIV-positive persons in Uganda. Both active approaches included in-person counseling during 2 quarterly clinic visits; the booster schedules differed by arm.
The primary outcome was change in self-reported drinking days during the prior 21 days, assessed at the 6-month and 9-month visits. The registry record does not report numerical primary or secondary outcome results, the detailed statistical analysis plan, or methods for missing data and multiplicity. Serious adverse event counts were 1 in the live phone call booster arm and 0 in each of the other arms; deaths were recorded as 0 in all arms. These safety counts are descriptive and do not establish causation.