This page provides an independent statistical analysis and educational interpretation of publicly reported results. ClinicalTrials.gov provides the official trial registry record. The ClinicalTrials.gov record indicates that results were posted, but no formal statistical analyses were included in the statistical-analysis field used for this page.
1. Trial at a Glance
iAdhere was a randomized, parallel, open-label phase 3 prevention trial in latent tuberculosis infection. The registered question concerned treatment completion under directly observed therapy versus self-administered therapy without reminders and versus self-administered therapy with weekly SMS reminders.
| Feature | iAdhere |
|---|---|
| Trial name | iAdhere |
| Brief title | Study 33: Adherence to Latent Tuberculosis Infection Treatment 3HP SAT Versus 3HP DOT |
| Phase | Phase 3 |
| Condition | Latent Tuberculosis Infection |
| Allocation | Randomized |
| Design model | Parallel |
| Masking | None |
| Primary purpose | Prevention |
| Enrollment | 1002 |
| Arms | 3 |
| Primary endpoint type | Binary |
| Results posted | Yes |
| Outcome measures posted | 8 |
| Lead sponsor | Centers for Disease Control and Prevention |
| Sponsor type | FED |
| ClinicalTrials.gov | NCT01582711 |
2. Clinical Question
The central statistical question was whether treatment completion differed between participants randomized to directly observed therapy and those randomized to self-administered therapy, both without and with weekly SMS reminders.
Population
Participants in a phase 3 prevention trial for latent tuberculosis infection.
Intervention strategies
Self Administered Therapy (SAT), with SMS reminders as an additional behavioral intervention, alongside isoniazid and rifapentine.
Comparator
3HP Directly Observed Therapy (DOT).
Primary question
How do treatment completion rates compare for DOT versus SAT without reminders and for DOT versus SAT with weekly SMS reminders?
3. Trial Design
3HP Directly Observed Therapy
- Directly observed therapy (DOT)
- Isoniazid and rifapentine
- Serious adverse events: 11/337
3HP Self Administered Therapy
- Self Administered Therapy (SAT)
- Isoniazid and rifapentine
- No SMS-reminder strategy identified for this arm
- Serious adverse events: 5/337
3HP SAT With SMS Reminders
- Self Administered Therapy
- Weekly SMS reminders
- Isoniazid and rifapentine
- Serious adverse events: 6/328
The serious-adverse-event denominators of 337, 337, and 328 sum to the registered enrollment of 1002.0. They are presented here as the affected/at-risk counts reported in the ClinicalTrials.gov record, rather than as independently reconstructed randomized-arm counts.
4. Randomization and What It Contributes Statistically
Randomization is the central design feature identified in the ClinicalTrials.gov record. In a randomized comparison, assignment to treatment strategy is determined by the study design rather than by the participant's subsequent treatment-completion behavior. This creates the basis for comparing completion outcomes across treatment strategies without making the comparison depend solely on observed differences in participants who chose or were able to complete therapy.
For this trial, the prespecified primary outcome is binary: whether treatment completion, as defined by the registry, occurred within the specified time frame.
Randomization does not make the groups numerically identical in every characteristic, and it does not eliminate uncertainty. Its principal statistical value is that, under the randomized design, systematic differences in treatment assignment are controlled by the allocation process rather than by participants selecting their treatment strategy.
5. Primary Endpoint
| Endpoint | Definition | Time frame | Type |
|---|---|---|---|
| Treatment Completion Rate. | Compare treatment completion rates between participants randomized to DOT vs SAT without reminders and DOT versus SAT with weekly SMS reminders. Treatment completion is defined as taking at least 90% of the doses (11/12 doses of each drug) within 16 weeks of treatment initiation. | Up to 16 weeks from start of treatment. | Binary |
The endpoint is fundamentally a patient-level binary outcome. Each participant can be classified according to whether the registry-defined completion criterion was achieved within the specified 16-week period.
6. Results
ClinicalTrials.gov indicates that results were posted and that 8 outcome measures were posted. However, the ClinicalTrials.gov record contains no statistical analyses for the primary endpoint: no formal statistical analyses were posted. Consequently, there is no registry-reported treatment-effect estimate, confidence interval, or p-value that can be reported as a formal comparison.
What is and is not reported in the ClinicalTrials.gov record
Results posted: Yes · Statistical analyses posted: 0
The ClinicalTrials.gov record does not provide a formal DOT-versus-SAT effect estimate for the primary endpoint.
What the posted result status tells us
The trial record says that results were posted and that the primary endpoint has a posted result. But the ClinicalTrials.gov recordset separates posted outcome measures from formal statistical analyses. That distinction is important: an outcome can be reported descriptively without the registry extract used here containing the corresponding inferential comparison.
A binary completion endpoint can be summarized as a proportion in each randomized group and compared using an appropriate binary-outcome method. However, the registry-reported iAdhere data do not give those treatment-completion proportions or a formal comparison. Therefore, no effect size should be constructed from the available serious-adverse-event counts, and no p-value or confidence interval should be inferred.
The missing estimate also means that there is no confidence interval to interpret for the primary endpoint. A confidence interval would ordinarily describe the statistical precision of a prespecified treatment contrast; it would not describe the range of individual participants' treatment responses.
Likewise, a p-value would ordinarily address evidence against a prespecified null hypothesis under a particular analysis model. A p-value is not itself a measure of the magnitude or clinical importance of an effect. Because no p-value is reported here, the page does not substitute one.
7. How the Primary Endpoint Would Typically Be Analysed
The registered endpoint is binary, so a natural analysis begins by estimating the treatment-completion proportion in each randomized arm. Because the primary question contains two comparisons involving DOT, the analysis plan would need to identify the estimand, comparison measure, confidence-interval procedure, and any approach to controlling type I error across the planned comparisons.
| Statistical question | Common approach for a binary endpoint | What it answers |
|---|---|---|
| What proportion completes treatment? | Binomial proportion estimate | Describes completion within each treatment group |
| How large is the absolute difference? | Risk difference | Difference in completion probabilities between groups |
| How large is the relative difference? | Risk ratio | Ratio of completion probabilities between groups |
| How do the odds compare? | Odds ratio or logistic regression | Relative comparison on the odds scale |
| How should uncertainty be quantified? | Confidence interval for the selected effect measure | Precision of the estimated treatment contrast |
These are educational descriptions of methods commonly used for binary outcomes, not claims about which model was prespecified or used in iAdhere. The ClinicalTrials.gov record does not identify a formal statistical-analysis method.
8. Statistical Methodology
Binary endpoint analysis
The primary endpoint is classified as binary in the ClinicalTrials.gov record. That classification matters because the outcome is based on whether a participant satisfies a defined completion criterion by the end of the specified assessment period.
For a simple two-group comparison, a completion proportion can be written as:
The choice of denominator and the exact inferential method must follow the prespecified analysis population and statistical analysis plan. Those details are not reported in the statistical-analysis field for this record.
Risk difference
One clinically direct effect measure for a binary completion endpoint is the risk difference:
A positive value would indicate a higher estimated completion probability in the SAT group under this coding. This is a general interpretive framework, not a reported iAdhere estimate.
Risk ratio
A relative measure is the risk ratio:
A value of 1 would represent equal estimated completion probabilities. Values above or below 1 indicate relative differences according to the direction of the comparison.
Odds ratio
Another possible binary-outcome measure is the odds ratio:
The odds ratio is not the same as the risk ratio. When the outcome is common, the numerical difference between the two can be substantial. That distinction is particularly important when communicating adherence or completion outcomes.
Multiple randomized comparisons
The registered definition explicitly describes two treatment comparisons: DOT versus SAT without reminders and DOT versus SAT with weekly SMS reminders. When a trial makes multiple confirmatory comparisons, the prespecified statistical plan may need to address the familywise type I error associated with those comparisons.
Confidence intervals
For a binary endpoint, a confidence interval can accompany an absolute or relative treatment effect. The interval is useful because it communicates precision rather than merely indicating whether a hypothesis test crossed a chosen threshold.
For example, a confidence interval around a risk difference describes statistical uncertainty around the estimated difference in completion probabilities under the specified analysis framework. It does not describe the probability that the true effect lies inside the particular interval after the data have been observed.
P-values
A p-value is tied to a null hypothesis and an analysis procedure. It is not an effect-size measure. A small p-value can occur with a modest effect in a sufficiently large sample, while a clinically meaningful effect can fail to reach a prespecified threshold when information is limited.
No p-value is posted on ClinicalTrials.gov for the iAdhere primary endpoint in the ClinicalTrials.gov record.
9. Statistical Methods Explained
Why is randomization important for this trial?
Randomization creates treatment groups through an allocation process rather than allowing participants or investigators to choose the treatment strategy. That makes the randomized comparison the foundation for assessing whether differences in treatment completion are associated with the assigned strategy.
Why is treatment completion a binary endpoint?
The registry defines completion using a threshold: taking at least 90% of the doses, specified as 11/12 doses of each drug, within 16 weeks of treatment initiation. Each participant can therefore be classified according to whether that criterion was met.
Why might a risk difference be useful?
The risk difference expresses the absolute difference in completion probability between treatment strategies. For adherence outcomes, an absolute measure can be especially intuitive because it describes how much the estimated completion probability changes rather than only expressing the change as a ratio.
Why might a risk ratio also be reported?
A risk ratio describes the relative completion probability between groups. It answers a different question from the risk difference. For example, the same absolute difference can correspond to different relative effects depending on the underlying completion probabilities.
Why should an odds ratio not automatically be called a risk ratio?
Odds and probabilities are related but not interchangeable. The odds of completion are calculated as the probability of completion divided by the probability of non-completion. An odds ratio therefore measures a ratio of odds, not a ratio of probabilities.
Why does the two-comparison structure matter?
The registered primary definition specifies DOT versus SAT without reminders and DOT versus SAT with weekly SMS reminders. If both comparisons are treated as confirmatory hypotheses, the statistical plan needs to specify how the overall type I error is controlled. The ClinicalTrials.gov record does not state the method used for that purpose.
10. Safety Results
The ClinicalTrials.gov record reports serious adverse events by arm as affected participants over the corresponding at-risk population. These are the only safety results included in the ClinicalTrials.gov recordset.
| Safety measure | Affected / at risk |
|---|---|
| 3HP Directly Observed Therapy (DOT) | 11/337 |
| 3HP Self Administered Therapy (SAT) | 5/337 |
| 3HP SAT With SMS Reminders | 6/328 |
These counts should be interpreted as serious-adverse-event counts, not as treatment-completion results. The safety endpoint and the primary adherence endpoint answer different questions and should not be combined into a single efficacy or benefit-risk statistic.
11. What the Missing Formal Analysis Means
The distinction between a posted outcome and a posted statistical analysis is important in registry-based statistical review. A registry can contain outcome data without the particular statistical-analysis metadata needed to reproduce the formal comparison.
What is supported
The trial was randomized, had 3 arms, enrolled 1002.0 participants, registered a binary treatment-completion endpoint, and indicates that results were posted.
What is not supported
The registry-reported statistical-analysis field does not provide a treatment effect estimate, confidence interval, p-value, formal model, or multiplicity procedure for the primary endpoint.
Why this matters
A numerical treatment effect cannot be responsibly reconstructed merely from the trial design or from the serious-adverse-event data.
Educational implication
The trial is still useful for understanding randomization, binary endpoints, treatment-completion definitions, and the distinction between descriptive results and inferential analysis.
12. Analysis Population and Missing Data
The ClinicalTrials.gov record identifies the trial as randomized and provide the registered primary endpoint, but they do not specify the formal efficacy analysis population, the treatment of participants who discontinued, the denominator used for the completion analysis, or the statistical handling of missing treatment-completion assessments.
Those details can materially affect a binary adherence endpoint. For example, if some participants have no evaluable completion assessment, an analysis must specify whether they are included in the denominator, classified as non-completers under a prespecified rule, or handled through another method. The appropriate approach depends on the protocol and statistical analysis plan.
13. Crossover, Interim Analysis, and Bayesian Methods
The ClinicalTrials.gov record does not report a crossover strategy, an interim-analysis procedure, an alpha-spending method, a non-inferiority margin, or a Bayesian analysis for the primary endpoint.
| Design topic | Supported by registry-reported iAdhere data? | Interpretation |
|---|---|---|
| Crossover | Not reported | No crossover analysis is described. |
| Interim analysis | Not reported | No interim-monitoring method is reported. |
| Non-inferiority margin | Not reported | No non-inferiority margin is identified. |
| Bayesian methods | Not reported | No Bayesian analysis is identified. |
| Multiplicity adjustment | Not reported | The two registered comparisons are identified, but no adjustment procedure is reported. |
| Missing-data method | Not reported | No imputation or missing-outcome strategy is reported. |
14. Limitations
- No formal statistical estimates in the registry-reported analysis field: although results are marked as posted, no formal statistical analyses were posted.
- Primary endpoint results cannot be reconstructed: the available data do not provide treatment-completion counts or proportions by randomized comparison.
- No confidence intervals or p-values:
- Analysis population is not specified: the ClinicalTrials.gov record does not identify the formal denominator or analysis population used for treatment completion.
- Missing-data handling is not specified: the registry extract does not identify how participants without an evaluable completion assessment were handled.
- Multiplicity is unresolved: the primary endpoint definition contains two comparisons involving DOT, but the ClinicalTrials.gov record does not identify an error-control strategy.
- Safety and efficacy are separate outcomes: serious-adverse-event counts cannot be used as a proxy for treatment completion.
- Registry-level interpretation: the page is deliberately limited to the ClinicalTrials.gov record and does not import numerical findings from the linked publications.
15. Why This Trial Matters Statistically
iAdhere is a useful teaching case because it demonstrates that an apparently straightforward adherence question contains several distinct statistical decisions. The primary endpoint is binary, but the scientific interpretation depends on how completion is defined, how randomized groups are compared, which effect measure is selected, how uncertainty is quantified, and how the two registered comparisons are handled statistically.
| Concept | How it appears in iAdhere |
|---|---|
| Randomization | Participants were randomized in a parallel phase 3 design. |
| Binary endpoint | Treatment completion is defined as meeting a prespecified dose threshold within a specified time frame. |
| Absolute effect | A risk difference can describe the absolute difference in completion probability between groups. |
| Relative effect | A risk ratio can describe the relative completion probability between groups. |
| Odds ratio | An odds-scale comparison is another possible model-based summary for a binary outcome. |
| Confidence interval | Would quantify statistical precision around a selected treatment-effect estimate. |
| P-value | Would evaluate evidence against a specified null hypothesis under a specified analysis. |
| Multiplicity | The primary definition identifies two comparisons involving DOT, making error-control planning potentially relevant. |
| Analysis population | The ClinicalTrials.gov record does not specify the formal efficacy analysis denominator. |
| Missing data | The ClinicalTrials.gov record does not specify the handling of participants without an evaluable completion outcome. |
16. Statistical Interpretation vs Clinical Interpretation
Statistical interpretation
The trial uses randomization to create treatment-strategy comparisons, and the registered primary outcome is a binary treatment-completion endpoint assessed within 16 weeks. The ClinicalTrials.gov record does not provide the formal inferential comparison.
Clinical interpretation
The clinical question concerns whether treatment delivery strategies are associated with different rates of completing the defined 3HP regimen. A clinical interpretation requires the actual completion results and their prespecified analysis, neither of which is numerically reported here.
This distinction prevents a common analytical error: treating a trial design as though it were itself an observed treatment effect. Randomization establishes the framework for inference; the endpoint data and prespecified statistical analysis provide the numerical evidence.
17. A Practical Framework for Reading This Trial
Start with the completion rule
The registered definition requires at least 90% of doses, specified as 11/12 doses of each drug, within 16 weeks of treatment initiation.
Respect the randomized groups
The registered primary definition specifies DOT versus SAT without reminders and DOT versus SAT with weekly SMS reminders.
Decide how the effect is expressed
A binary endpoint can be summarized using an absolute risk difference, risk ratio, odds ratio, or another prespecified contrast.
Use confidence intervals
The treatment effect should be accompanied by an uncertainty measure appropriate to the selected estimand.
Separate effect size from statistical significance
The size and precision of an estimated treatment effect provide information that a p-value alone cannot provide.
18. Primary Endpoint: What Would a Complete Statistical Result Look Like?
The ClinicalTrials.gov record does not provide the numerical result, but a complete statistical presentation for this binary endpoint would ordinarily make several elements explicit.
| Reporting element | Purpose |
|---|---|
| Completion proportion by arm | Shows the observed frequency of meeting the registered completion criterion. |
| Specified treatment contrast | Defines exactly which groups are being compared. |
| Effect estimate | Quantifies the magnitude of the difference or ratio between groups. |
| 95% confidence interval | Describes statistical precision around the estimated effect. |
| P-value, if prespecified | Quantifies compatibility with the specified null hypothesis under the analysis model. |
| Analysis population | Defines which participants contribute to the denominator and inference. |
| Multiplicity procedure | Clarifies how the two primary comparisons are interpreted within the overall type I error framework. |
The important point is not simply to obtain a p-value. A statistically complete report should identify the estimand, the analysis population, the comparison, the effect measure, and its uncertainty. Without those elements, a statement such as "SAT was better" or "the strategies were equivalent" would not be a statistically complete description of the evidence.
19. Related Tutorials
Learn more about the statistical concepts most relevant to this trial:
20. Related Statistical Calculators
The primary endpoint is binary, so the relevant calculator pathway focuses on comparing proportions and quantifying uncertainty around treatment effects.
21. Sources
- ClinicalTrials.gov: iAdhere (NCT01582711).
- Linked PubMed record: PubMed 36815322.
- Linked PubMed record: PubMed 29393655.
- Linked PubMed record: PubMed 29114781.
The linked publications are provided as registry-associated sources. This page deliberately does not import numerical results from those publications because the ClinicalTrials.gov record specifies that all trial numbers on this page must come from the provided ClinicalTrials.gov dataset.
Continue with Clinical Biostats
Use the trial framework to study randomized comparisons, binary endpoints, effect measures, and confidence intervals in greater depth.
22. Record Summary
iAdhere is a randomized phase 3 prevention trial with 1002.0 enrolled participants and 3 parallel treatment arms. Its registered primary endpoint is treatment completion, defined as taking at least 90% of the doses, specifically 11/12 doses of each drug, within 16 weeks of treatment initiation. The primary definition specifies comparisons of DOT with SAT without reminders and DOT with SAT with weekly SMS reminders.
The statistical teaching value of the record lies in the relationship between randomized design and binary-outcome analysis. A rigorous analysis would distinguish the observed completion proportions from the treatment-effect measure, quantify uncertainty with confidence intervals, identify the analysis population, and explain how the two registered comparisons are handled within the statistical testing framework.