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 Improving Treatment Adherence in HIV-Infected Individuals (NCT00247611), a randomized phase 2 study. The primary outcome is AIDS Clinical Trials Group (ACTG) 3-day Recall Measure of Doses Taken.
| Feature | Registry value |
|---|---|
| ClinicalTrials.gov identifier | NCT00247611 |
| Phase | Phase 2 |
| Condition | HIV Antiretroviral Therapy (ART) Adherence, Health Behavior |
| Allocation | Randomized |
| Design model | Parallel |
| Masking | None (open label) |
| Enrollment | 594 |
| Arms in the primary outcome | n/a |
| Lead sponsor | University of Connecticut |
| Results posted | Yes |
2. Clinical Question
Population
HIV-infected individuals receiving antiretroviral therapy (ART). The primary-outcome description refers to participants with adherence assessments and defines analysis populations using the number of assessments and treatment interruptions.
Intervention
LifeWindows Intervention Sessions: an individualized, interactive computer software program delivered in conjunction with clinical care, intended to increase and support ART adherence.
Comparator
A control condition. The registry record does not describe the control condition in enough detail to identify its specific content.
Primary question
Whether the intervention was associated with a difference in the proportion of participants meeting the study’s definition of perfect adherence, based on the AIDS Clinical Trials Group (ACTG) 3-day Recall Measure of Doses Taken.
The study addresses a behavioral and measurement question: can an individualized software-based intervention, delivered alongside clinical care, change reported adherence to prescribed ART? The primary outcome reduces adherence to a perfect-versus-imperfect classification. It is therefore not a direct measure of every aspect of medication-taking behavior, nor does the registry describe it as an objective measure of doses taken.
Adherence was assessed at clinical care visits over 18 months of participation. Participants reported doses taken during the past 3 days relative to the number prescribed. The description says adherence was established separately for each medication and then averaged over the full regimen before perfect versus imperfect adherence was evaluated. Viral Load Count is listed as a secondary outcome, but comparative results for that outcome are not reported in the registry record.
The record identifies the clinical question and the outcome definition, but does not report an estimated difference between arms, an uncertainty interval, or a statistical test result for the primary outcome. The registry information therefore describes what the trial set out to evaluate, but does not establish the direction or precision of an intervention effect.
3. Trial Design
This was a phase 2, randomized, parallel-group trial with no masking recorded. The registry lists an enrollment of 594 and a completed status. The study began in 2005-10, and primary completion was recorded in 2008-03. The sponsor was the University of Connecticut.
Participants were assigned to a LifeWindows Intervention Sessions arm or a control arm. The intervention was behavioral and was described as an individualized, interactive computer software program delivered with clinical care. The control arm is identified simply as receiving the control condition; its activities are not described in detail in the registry record. Because no masking was recorded, the trial should be understood as unmasked in its recorded design. The registry does not specify who knew the assignments or describe procedures intended to limit awareness.
The primary outcome was assessed at each clinical care visit over 18 months of participation. Its stated measure was the ACTG 3-day Recall Measure of Doses Taken, analyzed as the percentage with 100% adherence. The record describes both an intent-to-treat population and an OP sample for the primary outcome. The intent-to-treat population included participants with at least two adherence assessments; the OP sample was further restricted to those with at least six assessments and no treatment interruptions over 18 months.
Viral Load Count is named as a secondary outcome. The registry lists four outcome measures and seven analyses, but does not report a complete analysis plan explaining how these relate to each other. The statistical model, covariate adjustment, handling of repeated measurements beyond the stated HLM analyses, and planned multiplicity procedure are not reported in the registry record. Nor is a sample-size justification or the intended detectable effect reported. These omissions limit what can be inferred about the design choices and analysis strategy from the registry entry alone.
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.
Baselined
| Stage | Control | Intervention |
|---|---|---|
| STARTED | 304 | 290 |
| COMPLETED | 287 | 277 |
| NOT COMPLETED | 17 | 13 |
Reasons for not completing this period:
| Reason | Control | Intervention |
|---|---|---|
| Completed baseline, never placed on ARVs | 8 | 5 |
| Completed baseline only | 9 | 8 |
Intention to Treat Sample
| Stage | Control | Intervention |
|---|---|---|
| STARTED | 304 | 290 |
| COMPLETED | 287 | 277 |
| NOT COMPLETED | 17 | 13 |
On Protocol Sample
| Stage | Control | Intervention |
|---|---|---|
| STARTED | 287 | 277 |
| COMPLETED | 176 | 152 |
| NOT COMPLETED | 111 | 125 |
Analysis population for the primary outcome, as defined in the registry: Intent to treat included all participants with \>=2 adherence assessments; OP sample further restricted this sample to those with \>=6 assessments and no treatment interruptions over 18-months of participation. Multiple pairwise comparison replaced missing values. HLM was used for over-time analyses on proportion reporting perfect adherence.
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: Baseline characteristics provided for randomized sample. Analyses were conducted on those with \>=2 adherence assessments for ITT.
| Characteristic | Unit | Control | Intervention | Total |
|---|---|---|---|---|
| Participants analysed | 304 | 290 | 594 | |
| Age, Categorical: <=18 years | Participants | 0 | 0 | 0 |
| Age, Categorical: Between 18 and 65 years | Participants | 301 | 284 | 585 |
| Age, Categorical: >=65 years | Participants | 3 | 6 | 9 |
| Age, Categorical: <=18 years | participants | 0 | 0 | 0 |
| Age, Categorical: Between 18 and 65 years | participants | 174 | 149 | 323 |
| Age, Categorical: >=65 years | participants | 2 | 3 | 5 |
| Age, Continuous | years | 47.28 (SD 8.01) | 46.76 (SD 8.00) | 47 (SD 8) |
| Sex: Female, Male: Female | Participants | 119 | 110 | 229 |
| Sex: Female, Male: Male | Participants | 185 | 180 | 365 |
| Gender: Female | participants | 72 | 66 | 138 |
| Gender: Male | participants | 104 | 86 | 190 |
| Region of Enrollment: United States | participants | 304 | 290 | 594 |
6. Primary Endpoint
The registry lists the primary outcome as AIDS Clinical Trials Group (ACTG) 3-day Recall Measure of Doses Taken, measured over Measured at each clinical care visit over 18 months of participation.
| Element | Registry entry |
|---|---|
| Result type | Number |
| Measure of spread | not applicable |
| Unit | percent with 100% adherence |
| Endpoint classification | not reduced to a single comparison |
The primary endpoint is the percentage with 100% adherence, using the AIDS Clinical Trials Group (ACTG) 3-day Recall Measure of Doses Taken. Participants reported the number of doses taken on each of the past 3 days relative to the number prescribed. For this study, the description says adherence was established separately for each medication and averaged over the full regimen. The main outcome classified adherence as perfect or imperfect.
This measure focuses on recent, self-reported medication-taking. In general, a short recall period can make questions concrete and relate reports to a specific recent interval. It also means the measure describes reported behavior over that interval, rather than directly verifying doses or representing every day across the full follow-up. The registry does not describe an objective adherence measure or provide validation results for use in this trial.
Perfect-versus-imperfect classification makes the primary endpoint easy to interpret as a threshold outcome, but it does not retain all distinctions among participants below the perfect-adherence threshold. The record says sensitivity tests were to examine whether findings on perfect/imperfect adherence were maintained at thresholds of 90%, 80%, and 70%. Those analyses are described as sensitivity tests, not as the primary endpoint. Their results are not reported in the registry record.
The endpoint is clinically relevant to the stated aim of supporting ART adherence, while remaining a self-report-based outcome. Interpretation depends on the precise scoring and analysis, including how repeated assessments and missing reports were handled. The registry gives some population and model information but does not supply comparative endpoint estimates or their precision.
7. Why No Re-analysis Is Shown
8. Statistical Interpretation
The registry record describes the trial’s objective, design, and primary adherence measure, but it does not report comparative efficacy results for that endpoint. No arm-specific primary-outcome values, treatment-effect estimate, confidence interval, or p-value are reported in the registry record. Consequently, the record does not establish whether the intervention arm had a higher or lower proportion meeting the perfect-adherence definition, or how precisely any difference was estimated.
The enrollment of 594 is a design and conduct descriptor, not a result. It does not reveal how many participants contributed to each analysis or how much information was available at each visit. The primary-outcome description gives eligibility rules for the intent-to-treat and OP samples, but arm-specific analysis counts are not reported in the registry record.
In general, an estimated effect describes the magnitude and direction of a comparison, while a confidence interval describes the range of values compatible with the estimate under the analysis assumptions. A p-value addresses compatibility with a specified null hypothesis under those assumptions; it does not measure the probability that the intervention works. None of these comparative statistics for the primary outcome is reported in the registry record, so the magnitude, direction, and precision of a possible difference cannot be assessed from the posted information.
The outcome description notes that perfect versus imperfect adherence was the main classification and that threshold-based sensitivity tests were to follow significant findings. However, the registry does not report whether the primary comparison was significant, what the sensitivity tests found, or how the planned analyses were distributed across the four outcome measures and seven analyses. Viral Load Count is listed as secondary, but its results are also not reported in the registry record.
The posted adverse-event figures list zero participants affected by serious adverse events and zero affected by other adverse events in each arm. These counts are reassuring only as a description of the adverse-event entries in this registry record; they do not establish that no events occurred outside the recorded categories or reporting period. Death counts are not reported in the registry record.
Any sample-size illustrations based on observed effects must be treated as planning illustrations, not as evidence of trial power or as observed outcomes. The registry record does not provide comparative effect estimates needed to interpret such illustrations. Overall, the available information supports a description of the research question and measurement approach, but not a conclusion about an intervention effect.
9. Statistical Methods Explained
The statistical methods described in the registry record include a binary primary-outcome classification, multiple pairwise comparison for missing values, and HLM for analyses over time. The entry gives some detail about the analysis populations, but does not describe enough about the model specification or missing-data procedure to reproduce the analysis. The explanations below distinguish the recorded methods from general statistical considerations.
What does the perfect-adherence endpoint represent?
The endpoint classifies participants as having perfect or imperfect adherence according to the ACTG 3-day Recall Measure of Doses Taken. In general, converting a continuous or graded adherence score into categories can make a threshold easier to communicate, but it discards information about differences among scores within each category. A participant just below the perfect threshold and a participant with a much lower score are both categorized as imperfect for the main outcome. The registry also describes sensitivity tests at thresholds of 90%, 80%, and 70%; their purpose is to check whether interpretation depends on the selected cutoff. The results of those tests are not reported in the registry record.
Why do repeated assessments matter?
Adherence was measured at each clinical care visit over 18 months of participation, so the outcome can vary within a participant over time. In general, repeated observations from the same participant are related rather than statistically independent. An analysis that treats every observation as independent can misrepresent uncertainty if it fails to account for this correlation. The registry says HLM was used for over-time analyses. HLM generally refers to a modeling approach that represents data at more than one level, such as observations grouped within participants. The specific model, assumptions, covariates, outcome distribution, and link function are not reported in the registry record, so the exact analysis cannot be reconstructed.
What does the analysis-population description mean?
The registry defines the intent-to-treat population as including all participants with at least two adherence assessments. It describes the OP sample as further restricted to this sample’s participants with at least six assessments and no treatment interruptions over 18 months of participation. These definitions determine which participants contribute to the respective analyses; they are not equivalent to a statement that every randomized participant was included in every primary-outcome calculation. The reason for requiring the assessment counts, and the consequences of those restrictions for the comparison, are not reported in the registry record.
In general, analysis populations can answer different questions. A population closer to all randomized participants may preserve the benefits of random assignment, while a restricted population can describe results among participants who meet specified follow-up or treatment criteria. Restriction can also affect comparability if eligibility for the analysis is related to adherence, follow-up, or treatment experience. The registry provides the population definitions but does not report enough participant-flow information or arm-specific counts to assess those issues.
How is missing information handled?
The primary-outcome description states that multiple pairwise comparison replaced missing values. The registry does not define the procedure further, including the information used to replace values, the assumptions about why assessments were missing, or whether uncertainty from replacement was incorporated into estimates. The phrase alone is insufficient to determine how the procedure was implemented.
In general, missing-data methods rely on assumptions about whether missingness is related to observed or unobserved outcomes. Different assumptions can lead to different estimates. A useful evaluation considers the amount and pattern of missingness, the method used, and sensitivity analyses under alternative assumptions. These details, including the number of missing assessments and any missing-data sensitivity results, are not reported in the registry record.
What can be said about uncertainty and model adequacy?
For a binary outcome, estimates can be expressed in different ways, and each expression has its own interpretation. In general, an interval estimate helps show how much uncertainty remains around an estimated comparison. Model adequacy also depends on whether the statistical assumptions fit the data structure and outcome. The registry does not report primary-outcome estimates, interval estimates, test results, model diagnostics, or the complete statistical analysis plan. It is therefore not possible to judge the precision or robustness of the primary comparison from the record.
Over-dispersion is a general issue that can arise when observed variability exceeds what a model assumes. If unaddressed, uncertainty estimates may be too small. The registry describes a count other outcome type and identifies HLM for over-time analyses, but does not report a specific over-dispersion assessment or a count-model specification. No conclusion about over-dispersion or its handling can be drawn from the record.
What is not specified about planning?
The registry gives a general description of sample-size illustrations based on a stated fraction of an observed effect, a two-sided significance level of 0.05, equal allocation, and a normal approximation. These are planning assumptions for an illustration, not evidence of observed power. A sample-size justification for the trial, the assumptions used when it was designed, and the effect targeted at planning are not reported in the registry record. Comparative results needed to interpret any illustration are also not reported in the registry record.
10. Multiplicity, Missing Data and Analysis Populations
The registry lists four outcome measures and seven analyses. The primary outcome also includes a main perfect-versus-imperfect classification and describes sensitivity tests at thresholds of 90%, 80%, and 70%, to be used following significant findings. This creates several potential comparisons or analytic results to interpret. The record does not explain the relationship between the seven analyses and the outcomes, identify a hierarchy of hypotheses, or state whether any multiplicity adjustment was planned or used. Whether the reported analyses control the chance of false-positive findings is therefore not reported in the registry record.
In general, making multiple statistical comparisons can increase the chance that at least one appears statistically significant by chance. Prespecifying a primary comparison, distinguishing confirmatory from exploratory analyses, and explaining any adjustment can help readers interpret a group of results. The description of threshold sensitivity tests suggests a robustness question, but the record does not state how those tests were formally interpreted or report their results.
Missing adherence assessments are addressed in the registry description by multiple pairwise comparison replacing missing values. The procedure is not explained in enough detail to evaluate its assumptions. The number and timing of missing assessments, reasons for missingness, and sensitivity analyses for missing-data assumptions are not reported in the registry record.
The intent-to-treat population included participants with at least two adherence assessments. The OP sample further required at least six assessments and no treatment interruptions over 18 months of participation. These analysis populations differ by eligibility criteria, but arm-specific counts and a participant-flow account are not reported in the registry record. General interpretation should recognize that restricting analyses to participants meeting follow-up criteria can change the population being compared and may affect the protection against bias that randomization is intended to provide.
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 |
|---|---|---|---|
| Control | 0 | 304 | 0.0% |
| Intervention | 0 | 290 | 0.0% |
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 |
|---|---|
| Control | 0 of 304 |
| Intervention | 0 of 290 |
The registry lists the number of participants at risk and the number affected for serious and other adverse events in each arm. For control, 304 participants were listed at risk, with 0 affected by serious adverse events and 0 affected by other adverse events. For intervention, 290 participants were listed at risk, with 0 affected in each of those categories.
These entries describe the registry’s recorded counts. They do not provide event descriptions, severity grades beyond the serious-versus-other categories, timing, duration, or a detailed account of how events were collected. No serious adverse-event terms are listed. Death counts are not reported in the registry record.
In general, zero recorded affected participants does not establish that an event was impossible or that no unrecorded event occurred. Safety interpretation depends on the completeness of monitoring, definitions, follow-up, and reporting. Those details are not reported in the registry record. The listed counts should therefore be read as the posted adverse-event summary, not as a complete account of all possible safety outcomes or a comparative safety analysis.
12. What the Result Does, and Does Not, Mean
The registry record identifies a randomized phase 2 study of LifeWindows Intervention Sessions for ART adherence and defines a self-reported primary endpoint. It shows that the planned question concerned perfect adherence based on the ACTG 3-day Recall Measure of Doses Taken, with assessments over 18 months of participation. It also describes an intent-to-treat population, a more restricted OP sample, and HLM for over-time analyses.
The record does not report the comparative primary-outcome results. It therefore does not show whether the proportion meeting the perfect-adherence definition differed between intervention and control, how large any difference was, or how precisely it was estimated. Viral Load Count is listed as a secondary outcome, but its results are not reported in the registry record. The intervention’s effect on either outcome cannot be determined from the posted information.
The measure concerns self-reported adherence over a recent 3-day period, combined across medications as described in the record. In general, this is not interchangeable with directly observed medication-taking or with a measure of adherence over every day of follow-up. The endpoint’s perfect-versus-imperfect classification also does not describe the full range of adherence values in the main comparison. The planned threshold sensitivity tests could inform how conclusions depend on the cutoff, but their results are not reported in the registry record.
The adverse-event summary records zero affected participants in each arm for serious and other adverse events, with deaths not reported in the registry record. Those entries describe what is listed, not every possible safety outcome. The record also does not give enough information to evaluate missing-data assumptions, multiplicity, or all model details. Thus, the registry entry supports a careful account of the study design and recorded outcomes, but it does not support a conclusion that the intervention did or did not change adherence or viral load.
13. Why This Trial Matters Statistically
This trial is statistically notable because it frames adherence as a repeated behavioral outcome rather than a single measurement. Assessments were planned at clinical care visits over 18 months, and the registry identifies HLM for over-time analyses. In general, repeated measurements can describe change and variation over time, but analysis must account for the relationship among observations from the same participant. The record names an approach without providing its full specification, illustrating why a method label alone does not settle how an analysis was conducted.
The primary endpoint also highlights a common measurement choice: turning a reported adherence score into a perfect-versus-imperfect outcome. Such a threshold can make a target behavior easy to communicate, while losing information about differences among imperfect scores. The record’s planned sensitivity thresholds recognize that the selected definition may affect interpretation. Without the corresponding results, however, the robustness of findings cannot be evaluated.
The stated missing-data approach and distinct analysis populations make participant inclusion important to interpretation. The intent-to-treat population requires at least two assessments, while the OP sample requires additional assessments and no treatment interruptions. In general, restrictions related to follow-up or treatment experience can change the group represented by an analysis. Transparent participant counts and missing-data details are needed to assess that possibility, and those details are not reported in the registry record.
Finally, the entry illustrates the distinction between registering an outcome and reporting a result. It names the primary and secondary outcomes and reports design information, yet does not provide comparative efficacy estimates or their uncertainty. Statistical interpretation depends not only on the outcome definition and model label, but also on effect estimates, precision, missing-data handling, and the relationship among multiple analyses.
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 registry record does not report comparative results for the primary adherence endpoint or the secondary Viral Load Count outcome. Arm-specific adherence values, treatment-effect estimates, confidence intervals, and p-values are not reported in the registry record. As a result, the direction, size, and precision of a possible effect cannot be assessed.
The primary measure is based on participant reports of doses taken during the past 3 days. The registry describes how medication-specific adherence was averaged over the full regimen and classified as perfect or imperfect, but does not report objective adherence verification or trial-specific validation information. The main classification may conceal variation among participants who fall below the perfect-adherence cutoff. Results for the stated sensitivity thresholds are not reported in the registry record.
Important statistical details are also absent. The full HLM specification, missing-value procedure, model assumptions, covariate use, diagnostic checks, and multiplicity approach are not reported in the registry record. The record states that multiple pairwise comparison replaced missing values, but does not explain the method or its assumptions. It does not give the extent or pattern of missing adherence assessments.
The registry defines an intent-to-treat population and an OP sample, but does not report the number of participants contributing to each analysis or a detailed account of participant flow. The control condition is not described beyond its label. A trial-specific sample-size justification and complete statistical analysis plan are not reported in the registry record. These limitations constrain both reproducibility and the assessment of how well the posted analyses address the stated clinical question.
15. Related Tutorials
16. Related Statistical Calculators
17. Sources
- ClinicalTrials.gov: Improving Treatment Adherence in HIV-Infected Individuals — NCT00247611.
- Registry record last updated 2013-12-20; status last verified 2013-11.
- Results first posted 2013-12-20.
- 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
The University of Connecticut sponsored this completed phase 2 randomized, parallel-group study, with no masking recorded. Enrollment was 594; the study start was 2005-10, and primary completion was 2008-03. The intervention consisted of LifeWindows sessions, described as an individualized, interactive computer software program delivered alongside clinical care. The comparator was a control condition whose content is not detailed in the registry record.
The primary outcome was the percentage with 100% adherence measured using the ACTG 3-day Recall Measure of Doses Taken, with assessment at clinical care visits over 18 months of participation. The record describes intent-to-treat and OP analysis populations, multiple pairwise comparison for missing values, and HLM for over-time analyses. Viral Load Count is listed as a secondary outcome.
Comparative results for the primary and secondary outcomes, statistical estimates, and uncertainty measures are not reported in the registry record. The adverse-event entries list zero affected participants in both arms for serious and other adverse events; deaths are not reported in the registry record.