Intervention to Improve Adherence in Teen Kidney Transplant: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Intervention to Improve Adherence in Teen Kidney Transplant (NCT01356277), a randomized na study. The primary outcome is Taking Adherence.

2. Clinical Question

The trial also aimed to assess graft outcomes and to identify healthcare-system characteristics associated with adherence. The registry record lists secondary measures including self-reported taking and timing adherence, acute rejection, tacrolimus trough-level variability, and change in estimated glomerular filtration rate. The primary comparison concerns adherence as measured by electronic monitoring, not a direct measure of graft function or a general assessment of every aspect of medication use.

The intervention combined behavioral support and electronic monitoring rather than isolating one component. Consequently, a comparison between the randomized groups addresses the multi-component package against attention control; it cannot identify which individual component accounts for any group difference. Details of eligibility criteria, such as an age range or transplant-related requirements, are not reported in the registry record.

3. Trial Design

TAKE-IT was a randomized, parallel-group trial with no masking. The registry record lists an enrollment of 170 and describes the study as completed. The recorded start was 2012-02, and primary completion was 2016-06. The listed phase is NA. The allocation ratio, concealment procedures, and any stratification factors are not reported in the registry record.

The intervention was a multi-component adherence program. Its description includes a patient-parent-Coach Adherence Support Team, standardized education about immunosuppressive medications, identification of adherence barriers, and electronic adherence monitoring with feedback. The record also lists action-focused problem-solving, dose reminders, and feedback. The attention-control group received visits at matching intervals, with active listening and non-specific support from the Coach; adherence was not discussed with control participants. The record does not describe masking, which is consistent with the trial being listed as having none.

The primary outcome summary includes 64 participants in the Behavioral Intervention arm and 74 in the Attention Control arm. Other endpoint summaries use different arm counts, including 76 and 88 for several measures. The reasons for these differences are not reported in the registry record, so these endpoint-specific counts should not be treated as interchangeable. They also should not be used to infer why participants were absent from a particular outcome analysis.

The registry record identifies repeated daily adherence observations and describes generalized estimating equations for the primary comparison. Further details about the analysis population, missing-data methods, protocol deviations, or prespecified statistical analysis plan are not reported in the registry record. The record mentions an additional as-treated analysis, but its methods and results 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.

Before assignment: 1 patient withdrew after completing the enrollment visit, but before randomization.

Run-in

Intervention Interval

Analysis population for the primary outcome, as defined in the registry: We used unadjusted ordinal logistic regression with generalized estimating equations to account for repeated measures (i.e. score on each day) to compare taking and timing adherence between the intervention and control groups during the intervention interval. An additional 'as-treated' analysis was also conducted. All available data were included.

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 Taking Adherence, measured over 12 months.

The primary endpoint was taking adherence over 12 months, measured with electronic monitoring. For each observed day, a participant received a score based on the percentage of prescribed doses taken. The daily score could be 0%, 50%, or 100% for a participant prescribed medication twice daily, and 0% or 100% for a participant prescribed medication once daily. The reported arm-level summary was the percentage of days of observation with 100% taking adherence.

This endpoint focuses on whether all prescribed doses were taken on a given observed day. It is not the same as a count of doses taken, and it does not describe whether doses were taken at the correct time; timing adherence was a separate secondary outcome. In general, electronic monitoring can capture patterns across repeated daily observations, whereas a single self-report summary depends on participant reporting. The record does not establish that the electronic measure captures every dose taken under all circumstances.

The arm summary pooled observed days across participants: days without a score because of technical problems or participant non-use, including travel, were not assigned a taking-adherence score. The record does not report how many days were unscored or how missing observation days were handled analytically beyond this endpoint definition. The reported summaries were 78% for Behavioral Intervention and 68% for Attention Control, based on endpoint counts of 64 and 74, respectively.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The primary analysis is described as unadjusted ordinal logistic regression with generalized estimating equations. Each daily adherence score had an ordered scale: lower scores represented less complete taking adherence, and the highest score represented 100% taking adherence. An ordinal model uses that ordering rather than treating the score simply as a continuous measurement. The registry reports the result as an odds ratio from logistic regression, but the precise category coding and model specification are not reported in the registry record.

Why account for repeated observations?

Each participant could contribute a daily score across the observation interval. Scores from the same participant may be correlated because they come from the same person and medication-taking pattern. In general, generalized estimating equations account for this within-participant dependence when estimating a group comparison. Treating all daily observations as independent would fail to recognize that repeated measurements from one participant are not equivalent to observations from unrelated participants. The registry says this approach was used, but details such as the working correlation structure and variance estimator are not reported in the registry record.

What does the odds ratio communicate?

An odds ratio compares odds under a model; it is not a percentage-point difference and is not a ratio of the arm-level adherence percentages. The reported OR is 1.66, with a 95% confidence interval from 1.15 to 2.39 and p = 0.006. Because the outcome is ordinal, interpretation depends on how the model defines the ordered contrast. The record does not give that coding detail, so the safest reading is to retain the reported contrast and relate its direction to the arm summaries without assigning a more specific odds interpretation.

In general, the confidence interval expresses uncertainty around an estimate under the model and its assumptions. It is not a range containing a specified proportion of individual participant outcomes. A p-value describes how compatible the observed result is with a null model, assuming the analysis model is appropriate; it does not quantify clinical importance. Neither statistic addresses possible bias from missing observations or other design limitations by itself.

How should the endpoint summaries be read?

The primary arm summaries are 78% and 68% of observed days with 100% taking adherence. They are calculated from days pooled across participants within each arm, rather than by averaging a participant-level adherence percentage with equal weight for every participant. Participants contributing different amounts of observed time can therefore contribute different numbers of days to the pooled summary. The registry does not report the distribution of each participant’s observation time or individual adherence values.

Days without a score because of technical problems or participant non-use were excluded from the endpoint numerator and denominator as described. In general, excluding unobserved days can make an arm summary harder to interpret if the likelihood of a missing score is related to adherence or differs between groups. Whether such a relationship occurred is not reported in the registry record. No imputation procedure or missingness sensitivity analysis is reported.

What do the secondary recalculations represent?

For some secondary outcomes, recalculated estimates use unadjusted methods applied to posted summary statistics. These calculations do not reproduce a participant-level analysis and may differ from a sponsor-reported analysis that uses another method or additional data. For the self-reported outcomes, the recalculated mean differences are 1.2 percentage points for taking adherence and 2.1 percentage points for timing adherence. Their recalculated confidence intervals include values on either side of no difference. These simple comparisons should be read as descriptive checks, not replacements for the sponsor’s reported analyses.

Acute rejection is reported as an event rate. In general, event rates describe events relative to observed time, while a rate ratio compares rates between groups. A simple recalculation from summary rates can be unstable when event counts are small; a wide interval signals substantial uncertainty. The recalculated rate ratio is 0.627, with a confidence interval from 0.042 to 9.439, and an unadjusted p-value of 0.736. The sponsor-reported comparison uses a chi-squared test and reports p = 0.26. The registry does not provide enough detail to reconcile those different analyses.

How do planning illustrations differ from power?

The sample-size illustrations use the observed effect as a planning input and show the number of participants per arm for specified power targets under stated assumptions. Those assumptions are a two-sided significance level of 0.05, equal allocation, and a normal approximation; the effect to detect is a stated fraction of the observed effect. In general, planning calculations ask how large a future study might need to be under an assumed effect and model. They do not measure the probability that this completed trial detected its effect, and they are not observed power. The registry’s own design justification and sample-size assumptions are not reported in the registry record.

The illustrations are conditional on their assumptions and on the observed effect, which is itself uncertain. Planning for a smaller fraction of an observed effect requires more participants per arm in the posted calculations. These values should be treated as illustrations, not as a recommendation or as evidence that the original trial was designed to a particular target. The registry record does not report an interim analysis plan, multiplicity adjustment, or a protocol-based power calculation.

10. Analyses Reported by the Sponsor

The registry reports a sponsor analysis for the primary endpoint: logistic regression with an odds ratio of 1.66, a 95% confidence interval from 1.15 to 2.39, and p = 0.006. The model description also identifies ordinal logistic regression with generalized estimating equations to account for repeated daily scores. That analysis uses the repeated observations and is the central inferential result reported for the primary comparison.

The arm-level summaries, 78% and 68%, describe the percentage of observed days with 100% taking adherence. They are useful for understanding the scale and direction of the outcome, but they are not an independent recalculation of the sponsor’s model-based odds ratio. Reconstructing the primary model would require participant-level repeated data and model details that are not reported in the registry record.

For secondary endpoints, posted sponsor analyses and simple recalculations may use different methods. For example, the registry reports a two-sided t-test for the self-reported outcomes and a chi-squared test for acute rejection, while the recalculations use unadjusted comparisons based on posted summaries. Differences between those p-values do not necessarily mean that the underlying values conflict: the methods and inputs differ. The sponsor-reported results remain the trial’s reported analyses; recalculations are supplementary checks and should not be treated as substitutes for analyses using the full trial data.

11. Multiplicity, Missing Data and Analysis Populations

The registry record does not report whether the statistical analysis plan adjusted for multiple testing, specified a hierarchy among outcomes, or designated particular secondary analyses as confirmatory. Several secondary measures and analyses are listed. In general, testing multiple outcomes without adjustment can increase the chance of at least one apparently positive result even when no corresponding effect exists. This makes unadjusted secondary findings exploratory unless a prespecified plan supports a stronger interpretation. The posted primary analysis remains the principal result identified in the record.

The primary endpoint definition says that no score was assigned on days when the pillbox was not used because of technical problems or participant non-use, including travel. Such days were not part of the observed-day summary as described. The number of unscored days, reasons by group, and any sensitivity analyses for missing observations are not reported in the registry record. In general, missingness matters when the chance of an unobserved score is associated with adherence or differs between study groups; the record does not establish whether either situation occurred.

The registry describes an additional as-treated analysis, but its definition, results, and relationship to the randomized comparison are not reported in the registry record. The analysis population for the reported primary result, including whether all randomized participants were analyzed in their assigned groups, is also not reported. It is therefore not possible to determine from the record whether the primary estimate is an intention-to-treat estimate or how exclusions and treatment adherence were handled.

Different endpoint summaries use different arm counts. Reasons for those differences, participant flow, withdrawals, and outcome-specific missingness are not reported in the registry record. Those omissions limit assessment of how representative each endpoint analysis is of the randomized groups. The record also does not report interim analyses or stopping rules.

12. 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 5% reporting threshold.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: hospitalizations (33.0%), ptld (1.1%).

The serious adverse event summary lists 76 participants at risk in the Multi-component Intervention group and 88 in the Attention Control group. Serious events affected 24 and 30 participants, respectively. Other adverse events affected 42 and 52 participants, respectively. The record reports zero deaths in each group. These are counts; the registry summary does not provide exposure-adjusted rates or a statistical comparison for these adverse-event figures.

Hospitalizations are listed as the most common serious-event term at 33.0%, followed by PTLD at 1.1%. The registry record does not specify the denominator or further details for these term percentages in the information reported. It also does not provide event-level descriptions, timing, severity grades, or an assessment of whether events were related to the intervention. The affected-participant counts should not be assumed to represent mutually exclusive categories, because that relationship is not described.

These figures describe reported adverse events during the trial; they do not, on their own, show that the intervention caused or prevented any event. Differences in group counts should be interpreted cautiously without exposure time, event detail, and an analysis accounting for the trial design. The registry record does not report a formal comparative safety analysis.

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

The main statistical finding is a higher posted summary of electronically monitored taking adherence in the intervention group, together with a reported model-based odds ratio above the null and a p-value of 0.006. The confidence interval gives a range of plausible model-based effect values, but it does not remove uncertainty or establish the size of any practical benefit. The result applies to the measured adherence outcome and trial observation interval.

It does not establish that every participant in the intervention group had better adherence than every participant in the attention-control group. The arm summaries pool observed days, and participant-level distributions are not reported. It also does not establish that every intervention component contributed, because the trial compared a package rather than separately randomized components.

The result does not by itself establish improved graft outcomes, changes in kidney function, or a reduction in acute rejection. Those are separate outcomes, with limited or uncertain numerical evidence in the posted record. Nor does the result establish that the same estimate would apply to populations or settings not described by the trial. The registry’s eligibility criteria and detailed participant characteristics are not reported.

In general, statistical evidence and clinical importance are related but distinct questions. A model-based result can indicate evidence of a group difference without specifying whether its magnitude is meaningful to patients or clinicians. The record does not report a threshold defining a clinically important adherence difference. There is also no basis in the registry record to infer an effect beyond the measured intervention period or to make treatment recommendations.

14. Why This Trial Matters Statistically

This trial is statistically informative because its primary outcome uses repeated daily measurements rather than relying only on a single retrospective report. The analysis recognizes that observations from the same participant are related. That design feature matters: a repeated-measures analysis can use the temporal detail in adherence monitoring while accounting for within-person dependence, rather than treating every day as an independent source of information.

The endpoint also illustrates a distinction between a pooled day-level summary and a participant-level summary. The percentage of observed days with complete taking adherence gives weight to observed days, so its interpretation depends partly on how much monitoring time participants contribute. The record explains the pooled denominator and identifies situations in which a day received no score. It does not report observation completeness, which leaves an important part of endpoint interpretation unresolved.

The primary result is complemented by self-reported adherence, timing adherence, and a clinical event rate, but the measures do not all answer the same question. Their different scales and levels of precision illustrate why a coherent primary endpoint and a prespecified analysis plan are important. The registry does not report its multiplicity strategy, so the secondary findings have a more limited confirmatory role.

Finally, the trial separates the effect of a multi-component program from the effect of attention and non-specific support, but it does not isolate individual program ingredients. The statistical interpretation should therefore remain tied to the randomized comparison and the electronically measured outcome. The difference between the primary endpoint summaries and the model-based analysis also shows why descriptive figures and inferential estimates are complementary rather than interchangeable.

15. 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 leaves important details unavailable for evaluating generalizability and analysis conduct. Participant eligibility criteria and detailed baseline characteristics are not reported, so the applicability of the findings to particular adolescent kidney transplant populations cannot be assessed from this record alone. The record also does not explain why endpoint-specific arm counts differ or provide participant flow through each outcome analysis.

The primary endpoint depends on electronic monitoring and excludes days with technical problems or participant non-use. The record gives examples of those circumstances but does not report the amount or group distribution of unscored time. If missing monitoring days were related to adherence, the pooled summaries could be affected; whether that occurred is not reported. Individual-level adherence distributions and observation-time distributions are also not reported.

Further primary-model details, including the ordinal coding and correlation structure, are not reported. The analysis population and handling of protocol deviations are not reported either. Although an as-treated analysis is mentioned, its results and methods are unavailable. Multiplicity procedures, interim analyses, and the trial’s own sample-size justification are not reported in the registry record.

The intervention is a package of behavioral support and monitoring components, so the randomized comparison cannot separate their contributions. Masking is listed as none, and the record does not describe safeguards against expectation effects for the self-reported outcomes. Numerical results are not reported for all listed secondary outcomes. Together, these limits counsel against extending the primary adherence finding to unreported outcomes, longer-term effects, or intervention components considered individually.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: Intervention to Improve Adherence in Teen Kidney Transplant — NCT01356277.
• Registry record last updated 2018-07-18; status last verified 2018-06.
• Results first posted 2018-07-18.
• 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.

19. Record Summary

The registry record describes TAKE-IT as a completed, randomized, parallel-group trial with no masking and an enrollment of 170. It compares a multi-component adherence intervention with attention control in adolescent kidney transplant recipients. The primary outcome is electronically monitored taking adherence over 12 months, summarized as the percentage of observed days with 100% taking adherence.

The posted summaries are 78% for Behavioral Intervention and 68% for Attention Control. The reported primary model gives an odds ratio of 1.66, with a 95% confidence interval from 1.15 to 2.39 and p = 0.006. This is evidence of a group difference in the measured adherence outcome under the reported analysis, not proof of effects on every secondary or clinical outcome.

Interpretation is limited by unreported details about missing observations, analysis populations, multiplicity, and model specification. The registry’s additional secondary outcomes and safety summaries provide context, but they do not resolve those limitations or establish which intervention components account for the primary result.