The S.M.A.R.T. Project: Stress Management and Resilience Training for Teens: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for The S.M.A.R.T. Project: Stress Management and Resilience Training for Teens (NCT03270943), a randomized na study. The primary outcome is Time to Incident Depression.

2. Clinical Question

3. Trial Design

The S.M.A.R.T. Project was a randomized, parallel-group behavioral study sponsored by the University of North Carolina, Chapel Hill. Its registry enrollment was 77. The study began on 2017-09-05, reached primary completion on 2020-06-03, and is listed as completed. The phase is recorded as NA. The design is listed as single-masked, but who was masked is not reported in the registry record.

Participants were assigned to Mindful Self-Compassion (MFY) or Healthy Lifestyles (HLG). Both groups involved an 8-week course and 6 monthly continuation sessions. The healthy-lifestyles group was intended to provide an attention-control comparison, so the contrast concerns the two programs rather than a program versus no intervention.

The primary outcome was time to incident depression, assessed weekly for up to 36 weeks using the PROMIS Pediatric depression measure. The primary analysis excluded 3 participants with high baseline scores. The posted arm summaries report analysis counts of 28 for MFY and 31 for HLG. These analysis counts should not be assumed to account for every randomized participant; the reasons for any other differences between enrollment and analysis are not reported in the registry record.

Secondary outcomes included changes in PROMIS Pediatric Depression scores, SMFQ scores, and Brief Resiliency Scale scores. The registry lists 7 outcome measures and 7 analyses, although only 3 secondary outcomes are identified in the posted outcome information. Detailed assignment procedures, the planned analysis protocol, and the rationale for the enrollment target 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: \*Note: the first two groups (Period 1: Intervention Refinement) were not randomized. Participants enrolled in those groups received the intervention, but the intervention and logistics were adapted based on participant feedback. Hence, these participants are not included in statistical analyses.

Intervention Refinement

Randomized (Willingness to be)

Analysis population for the primary outcome, as defined in the registry: 3 participants with high baseline scores were excluded from the analysis procedure.

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: Includes only the randomized population (Period 2: Randomized (Willingness to be)) who attended at least one intervention class.

6. Primary Endpoint

The registry lists the primary outcome as Time to Incident Depression, measured over weekly for up to 36 weeks.

The primary endpoint is time to incident depression, expressed in weeks. Participants were assessed weekly for up to 36 weeks. The measure was the PROMIS Pediatric short-form depression 8a. Its raw scores range from 0 to 32, with higher scores indicating greater depression. Analysis used a T-score transformation with a population mean of 50 and a standard deviation of 10; a score of 65 was used to indicate incident depression.

Time-to-event outcomes capture when a defined event occurs, rather than only a score at a single follow-up visit. This can make use of the timing of events and can accommodate participants whose event is not observed during follow-up, provided the analysis handles that situation appropriately. The registry record does not describe censoring rules or how incomplete follow-up was handled.

The registry lists the endpoint summary as a median with inter-quartile range. The posted median time is 36.0 weeks for MFY and 18.86 weeks for HLG. The accompanying interval endpoints are 0.71 to 36.0 weeks for MFY and 2.86 to 36.0 weeks for HLG. A median time describes the middle of the event-time distribution under the analysis; it is not an estimate of every participant’s experience. The PROMIS measure was used instead of the SMFQ for this endpoint because the registry states it allowed more frequent measurement.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

What does a median time summarize?

For a time-to-event endpoint, the median is the time by which half of the analyzed group’s event-time distribution has reached the defined event, according to the analysis. It offers a time-based summary, but it does not describe all features of the distribution. Two groups can have similar medians while differing at other times, or different medians while their event patterns vary in more complex ways. The posted median summaries are 36.0 weeks for MFY and 18.86 weeks for HLG, with the registry identifying inter-quartile range as the dispersion measure.

The reported interval endpoints are 0.71 and 36.0 weeks for MFY, and 2.86 and 36.0 weeks for HLG. These give information about the spread of observed event times as summarized in the registry, but do not replace a confidence interval for a between-group effect. An inter-quartile range describes variability within a group; a confidence interval describes uncertainty in an estimated quantity. They answer different questions.

What does the log-rank test compare?

The log-rank test is a general method for comparing time-to-event patterns between groups. It uses information about events and the participants still under observation over follow-up, rather than comparing only the two median values. Its interpretation depends on the event definition, follow-up, and analysis handling of participants who do not have an observed event. Those implementation details are not reported in the registry record.

The sponsor reports a log-rank p-value of 0.02 and characterizes the result as superiority. That test result does not itself quantify the magnitude of a group difference. A hazard ratio, confidence interval, or other effect estimate would help express the relative contrast and its precision, but no such estimate is reported for this comparison in the registry record.

How should the p-value be read?

General explanation: a p-value is calculated under a null hypothesis and test assumptions. It reflects how unusual the observed test statistic, or one more extreme, would be if that null model held. A smaller p-value can indicate greater incompatibility with that null model, but it does not state the probability that a treatment effect is real. It also does not establish that an observed difference is large, important, or free from bias.

The posted p-value is a result of the sponsor’s reported analysis. No separate unadjusted recalculation is available for this median-time endpoint: values were missing or not estimable for a two-arm comparison. Thus, there is no recalculated comparison estimate to set against the sponsor’s test result.

Why can time-to-event data be difficult to summarize?

General explanation: follow-up can end before every participant experiences the event. Such observations are commonly treated as censored, meaning that the event time is known only to be beyond the observed follow-up. Appropriate time-to-event methods can use the information available up to censoring, but the validity of the result depends on assumptions about how observation ends and how the analysis is conducted. The registry record does not report the censoring rules or the amount and pattern of incomplete follow-up.

Repeated weekly measurement also raises questions about when an event is considered to have occurred. The registry defines the threshold used to indicate incident depression and states that measurement was weekly, but does not detail the event-date convention or handling of missed assessments. This limits independent assessment of the endpoint implementation.

What do the planning illustrations represent?

General explanation: sample-size planning uses assumptions about the effect to be detected, the significance level, allocation, and statistical method. The registry record’s planning description specifies a two-sided significance level of 0.05, equal allocation, and a normal approximation, with illustrations based on fractions of the observed effect. Such values are planning illustrations, not observed power and not proof that the original trial was designed with those assumptions. The registry does not report the original sample-size justification or enough detail to reconstruct the trial’s design planning.

For time-to-event planning, the stated approach uses a total number of events. That distinction matters because the amount of information in a time-to-event comparison depends on events as well as enrollment. Planning illustrations should therefore be read as hypothetical design scenarios rather than as a description of what the completed study achieved.

10. Analyses Reported by the Sponsor

The sponsor’s reported analysis is a log-rank test comparing MFY with HLG. The registry records p = 0.02 and classifies the comparison as superiority. No hazard ratio, confidence interval, or other sponsor-reported relative effect estimate is included for this outcome.

A separate recalculated two-arm comparison is not available for this median-time endpoint because the necessary values are missing or not estimable. The posted medians and their interval endpoints are descriptive group summaries; they are not a replacement for the sponsor’s log-rank analysis. Accordingly, the sponsor’s p-value is the only reported inferential comparison in the registry record, while the medians provide a separate description of each group’s event-time distribution.

The registry record does not report the detailed statistical analysis plan, including whether the log-rank test was prespecified or whether adjustments were made. It also does not provide a confidence interval around the comparison. The sponsor’s result can therefore be described as reported, but the full analysis choices and the precision of a relative effect cannot be independently evaluated from the registry information.

11. Multiplicity, Missing Data and Analysis Populations

The registry lists 7 outcome measures and 7 analyses, and identifies 3 secondary outcomes: change in PROMIS Pediatric Depression scores, change in SMFQ scores, and change in Brief Resiliency Scale scores. It does not report how many hypotheses were treated as confirmatory, whether the primary endpoint had priority over secondary outcomes, or whether any adjustment for multiple comparisons was planned or applied. Without that information, the reported p-value should be understood as the result for the stated comparison, not as evidence that multiplicity was addressed.

Missing-data methods are not reported in the registry record. The primary endpoint was assessed weekly for up to 36 weeks, but the record does not describe missed assessments, loss to follow-up, censoring rules, imputation, or sensitivity analyses. It is therefore not possible to judge whether incomplete outcome data could have affected the event-time comparison or how robust the result is to alternative assumptions.

The primary analysis excluded 3 participants with high baseline scores. The criteria used to identify those scores and the rationale for exclusion are not reported. The registry gives analysis counts of 28 for MFY and 31 for HLG, but does not fully define the analysis population or report a complete account of participants included in each analysis. Whether the analysis followed the randomized assignment for all eligible participants is not reported in the registry record.

These omissions matter because interpretation depends not only on the test statistic but also on which outcomes, participants, and follow-up observations were included. The registry record does not provide enough detail to assess multiplicity control, missing-data assumptions, or the relationship between randomized enrollment and the final primary analysis population.

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 0% reporting threshold.

Most frequent serious adverse event terms

Serious events with the highest arm-level rate: hospitalization for unknown reason (2.8%).

The registry reports serious adverse events among 36 participants at risk in MFY and 41 participants at risk in HLG. One serious adverse event is recorded for MFY and none for HLG. Other adverse events are recorded as zero in both groups, and deaths are recorded as zero in both groups.

The most frequently listed serious adverse-event term is “Hospitalization for Unknown Reason,” reported as 2.8%. The record does not identify the group associated with that term or provide further details about its circumstances. The event counts and percentages should be read as the registry’s safety reporting, not as a complete characterization of risk.

These figures do not establish that an intervention caused or prevented an event. The record does not report event timing, severity details beyond the serious-event classification, relatedness assessments, or the method used to collect and classify adverse events. Differences in the numbers at risk and the small event counts also limit what can be inferred from a direct numerical comparison. Safety interpretation should therefore remain descriptive and tied to the information recorded.

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

The posted results show that the median time to incident depression was 36.0 weeks in MFY and 18.86 weeks in HLG. The sponsor’s log-rank test reports p = 0.02 and labels the comparison as superiority. These are the central descriptive and inferential results available for the primary outcome in the registry record.

They do not show the size of a relative effect, because the registry does not report a hazard ratio or a confidence interval for the comparison. Nor does the p-value express the probability that either program is effective. It is a test result under a statistical model, not a measure of clinical importance. The reported medians describe the groups’ event-time distributions but do not show how individual participants’ outcomes varied or whether the group patterns differed consistently throughout follow-up.

The study was a randomized comparison of two group programs, with an attention-control comparator. Its findings therefore concern the contrast between MFY and HLG as implemented in this study, rather than a contrast with no program. The registry describes the intended population as adolescents with subsyndromal depression, but provides limited eligibility and analysis-population detail. The primary analysis also excluded 3 participants with high baseline scores.

The completed study’s registry record does not provide a full statistical analysis plan, missing-data methods, multiplicity procedures, or detailed participant flow. Those gaps prevent a complete assessment of how the analysis was selected and how incomplete follow-up may have influenced the result. The results should be interpreted as the posted trial findings, with these reporting limits in view, rather than as a complete account of the study’s evidence or its implications for an individual.

14. Why This Trial Matters Statistically

This trial illustrates why a time-to-event outcome needs more than a single group summary. Median times provide a direct description in weeks, while a log-rank test compares the event-time patterns over follow-up. The distinction matters: group medians do not convey the full pattern, and a test result does not quantify the size or precision of the difference.

The registry’s reported medians and log-rank p-value answer different statistical questions. The medians describe each group’s event-time distribution; p = 0.02 is the sponsor’s inferential comparison. Without a reported hazard ratio and confidence interval, readers cannot assess the relative effect estimate and its uncertainty from the registry entry. This is a useful example of why effect estimates and interval estimates are important companions to hypothesis tests.

The trial also highlights the importance of clear reporting for weekly outcomes. Interpretation of time to incident depression depends on how the threshold was applied, how events were dated, and how missed assessments or incomplete follow-up were handled. The registry supplies the score threshold and assessment frequency, but not those operational details. Transparency about these decisions helps readers assess whether the endpoint and statistical method align.

Finally, a randomized design does not remove the need to understand the analysis population, missing data, or multiple outcomes. The record identifies exclusions and several secondary outcomes, but does not provide the analysis plan or multiplicity and missing-data procedures. These are not minor details: they determine the scope and precision of claims that can be drawn from a reported test. The statistical lesson is to read the estimate, method, population, and reporting limits together.

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.

Several trial-specific details needed for a fuller assessment are not reported in the registry record. Although the study is listed as single-masked, the identity of the masked party is not stated. The record also does not describe the randomization procedure, allocation concealment, group-session delivery, adherence, or fidelity to the intervention programs.

The primary analysis excluded 3 participants with high baseline scores, but the threshold defining high baseline scores and the reason for exclusion are not reported. The posted analysis counts are 28 for MFY and 31 for HLG, while enrollment is 77. The registry does not provide enough participant-flow information to explain all differences between enrollment and the analysis counts.

For the time-to-event endpoint, censoring rules, missed weekly assessments, event dating, and incomplete follow-up are not described. No hazard ratio or confidence interval is reported for the log-rank comparison. The record also does not state whether the primary analysis was prespecified, how secondary outcomes were handled, or whether multiplicity adjustments were used.

The available safety information is limited to reported counts and a serious-event term; event timing, relatedness, and further clinical details are not given. The original sample-size justification and detailed statistical plan are also not reported in the registry record. These limitations restrict assessment of precision, robustness, analysis selection, and generalizability beyond the analyzed participants and study programs.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: The S.M.A.R.T. Project: Stress Management and Resilience Training for Teens — NCT03270943.
• Registry record last updated 2021-05-18; status last verified 2021-03.
• Results first posted 2021-05-18.
• 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

NCT03270943, titled “The S.M.A.R.T. Project: Stress Management and Resilience Training for Teens,” is a completed randomized, parallel-group behavioral study sponsored by the University of North Carolina, Chapel Hill. The registry lists enrollment of 77, a start date of 2017-09-05, and primary completion on 2020-06-03. The phase is recorded as NA, and masking is listed as single.

The study compared an 8-week Mindful Self-Compassion course with an 8-week Healthy Lifestyles course; each included 6 monthly continuation sessions. The primary endpoint was time to incident depression, assessed weekly for up to 36 weeks using a PROMIS Pediatric depression threshold of 65. The posted medians were 36.0 weeks for MFY and 18.86 weeks for HLG. The sponsor reported a log-rank p-value of 0.02 and classified the comparison as superiority.

The primary analysis excluded 3 participants with high baseline scores. The registry does not report a hazard ratio, confidence interval, full analysis plan, or detailed missing-data methods.