Preventing Drug Abuse Among Sexual Minority Youth: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Preventing Drug Abuse Among Sexual Minority Youth (NCT03954535), a randomized na study. The primary outcome is Average Number of Times Youth Drank Alcohol in the Past Month.

2. Clinical Question

3. Trial Design

The registry record describes a randomized, single-masked, parallel-group behavioral study. The two groups were Free2b and an attention-placebo control. Overall enrollment was 1216, and the record lists the study status as completed. The start date was 2020-01-15, and the primary completion date was 2024-03-30. The study phase is listed as NA. The identity of the person or people masked is not reported in the registry record.

Both groups received a website with entertainment features and resources. The intervention group additionally had 9 sessions. This active website comparator can help frame the contrast: the posted comparison concerns the intervention package relative to the attention-placebo website, rather than relative to no website or no study contact. The exact content and delivery of the sessions, as well as use or adherence measures, are not reported in the registry record.

The primary assessment is described as occurring an average of 5 months from study onset, and the analyzed population includes only participants who completed that assessment at approximately 5 months after baseline. The arm-level primary outcome counts are 568 in the intervention group and 605 in the control group. Those outcome counts need not equal total enrollment because the primary analysis population is assessment completers. The record does not report the original design assumptions, the randomization procedure, allocation concealment, or a detailed statistical analysis plan.

The study was sponsored by Columbia University and concerned substance abuse prevention among sexual minority youth. Its brief description characterizes the planned trial as longitudinal and nationwide. The registry record does not report the achieved geographic distribution or provide enough detail to assess how participants were recruited or whether the analyzed group represents all randomized participants.

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.

Overall Study

Analysis population for the primary outcome, as defined in the registry: The analyzed population reflects only those participants who completed the assessment at approx. 5 months after baseline.

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: 8 participants did not complete baseline assessments. Baseline data was not collected from these participants.

6. Primary Endpoint

The registry lists the primary outcome as Average Number of Times Youth Drank Alcohol in the Past Month, measured over an average of 5 months from study onset.

The primary endpoint is the average number of times a participant drank alcohol in the past month. Participants were asked to report how many times they had used the substance in the past month, and the endpoint is recorded as a mean with a standard deviation. It is assessed at an average of 5 months from study onset; this wording does not identify a single identical assessment date for every participant.

A count of drinking occasions is one way to describe the frequency of alcohol use over a specified recall period. A group mean gives a concise summary across participants, while the standard deviation describes how dispersed individual reports are around that mean. The endpoint addresses reported drinking frequency, not the quantity consumed on an occasion, alcohol-related consequences, or other substance use. Those are distinct outcomes and should not be inferred from this measure.

Using a past-month recall period makes the question specific to recent behavior at the assessment. The registry record does not report the exact wording of the instrument beyond the description, whether reports were verified by another source, or how the assessment handled unusual or incomplete responses. The analyzed population is limited to participants who completed the assessment at approximately 5 months after baseline. Consequently, the endpoint describes those completers and should not automatically be treated as a complete account of every randomized participant.

7. Results Posted in the Registry

• Intervention: mean 0.83 versus 1.03 on control; difference -0.20 (95% CI -0.45 to 0.05), standardized effect d = -0.09; z-test p = 0.118.

8. Secondary Efficacy Endpoints

The registry also posts results for secondary outcomes. For each outcome below, the endpoint type, the test applied to the posted summary statistics, the result and its interpretation are given in turn. Secondary endpoints are usually not adjusted for multiple testing, so they support rather than confirm the primary finding.

Mean Scores for Goal Setting

• Endpoint type: Continuous: a mean value per participant, compared as a difference in means.
• Test: Unadjusted two-sample z-test for the difference in means, using the posted standard deviations.
• Result:
  • Intervention: mean 2.38 versus 2.34 on control; difference 0.04 (95% CI -0.03 to 0.11), standardized effect d = 0.07; z-test p = 0.262.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Intervention and the control. The interval gives the range of effects that remain compatible with the data.

Mean Scores for Problem Solving

• Endpoint type: Continuous: a mean value per participant, compared as a difference in means.
• Test: Unadjusted two-sample z-test for the difference in means, using the posted standard deviations.
• Result:
  • Intervention: mean 2.06 versus 2.12 on control; difference -0.06 (95% CI -0.11 to -0.01), standardized effect d = -0.14; z-test p = 0.014.
• Interpretation:
  • The 95% confidence interval excludes no difference, so the mean is statistically lower for Intervention than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.

Mean Scores for Self-Esteem

• Endpoint type: Continuous: a mean value per participant, compared as a difference in means.
• Test: Unadjusted two-sample z-test for the difference in means, using the posted standard deviations.
• Result:
  • Intervention: mean 2.63 versus 2.73 on control; difference -0.10 (95% CI -0.17 to -0.03), standardized effect d = -0.18; z-test p = 0.003.
• Interpretation:
  • The 95% confidence interval excludes no difference, so the mean is statistically lower for Intervention than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.

The posted secondary outcomes present a mixed overall pattern rather than a uniform shift across all measured domains. Goal-setting scores are close between groups, whereas the posted comparisons for problem-solving and self-esteem point in directions that differ from one another. This range of patterns does not mirror a consistent broad improvement across every reported measure, and secondary scores should be considered separately from the primary alcohol-frequency outcome.

These outcomes can add context about measures related to the intervention's stated aims, but they do not substitute for the primary endpoint. A difference in a secondary score does not establish a corresponding difference in drinking frequency, and a result for the primary alcohol measure does not imply a result for every psychosocial measure. The registry record does not report a framework that links changes in these scores to the primary outcome or a mediation analysis.

Unadjusted secondary findings carry less weight when many outcomes are examined. In general, the more comparisons considered, the greater the chance that some may appear noteworthy through random variation alone. The record lists 64 outcome measures and has 40 secondary outcomes posted, but it does not report a multiple-comparison adjustment, a hierarchy among the secondary endpoints, or which results were prespecified for confirmatory interpretation. The registry record also lists repeated descriptions of goal-setting and problem-solving outcomes; it does not explain how those entries correspond to distinct measures or analyses.

For these reasons, secondary patterns are best treated as descriptive and hypothesis-generating unless supported by a prespecified analysis plan and appropriate control of multiplicity. The record does not report such details. The posted unadjusted comparisons can identify areas for further study, but they should not be interpreted as a single confirmatory body of evidence or as proof that the intervention changed each underlying construct.

Secondary outcomes registered for this trial, for reference (only those above were recalculated):

• Mean Scores for Goal Setting
• Mean Scores for Goal Setting
• Mean Scores for Goal Setting
• Mean Scores for Goal Setting
• Mean Scores for Problem Solving
• Mean Scores for Problem Solving
• Mean Scores for Problem Solving
• Mean Scores for Problem Solving

9. Statistical Interpretation

10. Statistical Methods Explained

The registry record posts group means and standard deviations for a continuous primary outcome, along with a comparison of the group means. The primary measure is the average number of times participants reported drinking alcohol in the past month. For this type of outcome, the mean describes the average reported count in each group. The standard deviation describes how much individual observations tend to vary around that average; it is not the uncertainty of the mean itself.

What does a mean difference describe?

The comparison is expressed as the intervention mean minus the attention-placebo control mean. A negative value therefore indicates a lower observed mean in the intervention group, while a positive value indicates a higher observed mean. The estimated difference is in the original outcome units, or number of times. This makes the estimate directly interpretable against the measure's scale. The difference is a summary of group averages; it does not show that every participant had an outcome matching the group average or that all participants changed in the same direction.

The posted mean difference is an unadjusted recalculation from the two group summaries. In general, an unadjusted comparison does not account for additional variables or modeling choices beyond the information used in that summary calculation. The registry record does not report a separate sponsor analysis for this endpoint, a model specification, or covariate adjustment. Thus, the recalculated contrast should be read as a summary-statistic comparison rather than as evidence of a particular regression or longitudinal model.

What does a confidence interval convey?

In general, a confidence interval expresses uncertainty around an estimate under the method used to calculate it. It gives a range of values compatible with the observed data and the calculation's assumptions. A narrower interval indicates greater precision than a wider interval for the same kind of estimate. An interval that includes zero for a difference means the data do not distinguish the estimate from no difference at the confidence level represented by that interval. It does not prove that the true difference is zero, nor does it assign a probability to a particular fixed value being true.

The primary interval runs from -0.451 to 0.051. It includes zero and includes values on either side of zero, so the direction and size of the underlying mean difference are not pinned down precisely by this estimate. The registry record does not specify the confidence level or calculation method for the posted interval. No additional methodological detail should be inferred from its appearance alone.

How should the p-value be read?

In general, a p-value measures how unusual data at least as incompatible with a null model would be if that model and the analysis assumptions were correct. It is not the probability that the null hypothesis is true, and it does not measure the size or practical importance of an effect. The primary unadjusted p-value is 0.118. It is consistent with the interval's inclusion of zero. The registry record does not report whether the comparison was prespecified as the sole primary inferential test or whether a more complex analysis was intended.

The record also gives standardized summaries for comparisons and a planning illustration based on observed effects. Standardized effect measures express a difference relative to variability, rather than in the original units. They can help describe scale, but their interpretation depends on the outcome and context. They do not replace the mean difference or its interval, and they do not imply a clinical threshold. The planning illustrations use effect fractions and assumptions described with those figures; they are not analyses of observed power.

Why does variation matter?

In general, larger variability among individuals makes a group mean harder to estimate precisely when other features are held similar. For count outcomes, the data may also have distributional features that a simple mean-and-standard-deviation summary does not reveal. Over-dispersion is a general statistical concept in which observed variability exceeds what a chosen count model expects. When present, it can make uncertainty estimates from an unsuitable model too small. The registry record does not report a count-model assessment or an analysis that evaluates over-dispersion, so it cannot be determined whether that issue was examined.

Finally, the analyzed population matters. The primary analysis includes participants who completed the assessment at approximately 5 months after baseline. In general, restricting analysis to completers can affect interpretation if completion is related to group assignment or outcomes. The record does not report a missing-data method or a comparison of completers with non-completers. The posted summaries therefore describe observed completers; they do not by themselves resolve possible missing-outcome bias.

11. Confidence Intervals and Precision

The primary mean difference is -0.2 times, with a confidence interval from -0.451 to 0.051. This interval includes zero and extends into both negative and positive values. It therefore leaves uncertainty not only about the size of the difference, but also about its direction. The posted point estimate is below zero, but the interval does not support treating that direction as settled.

In general, a confidence interval is a way to express the precision of an estimate under a specified statistical method. Its width reflects sampling uncertainty and depends on the data and assumptions used in the calculation. The interval is not a range containing every participant's result, and it does not describe the spread of individual drinking reports. The standard deviations, 2.12 in the intervention group and 2.26 in the control group, describe individual-level variation; the confidence interval describes uncertainty around the group contrast.

The unadjusted p-value of 0.118 is consistent with an interval that includes zero. Neither measure should be read as proof of no difference. Rather, the posted primary estimate is not precise enough to rule out zero or a difference in either direction under the reported calculation. The registry record does not report the interval's confidence level or method, nor does it report an adjusted primary analysis that could yield a different estimate or precision.

The number of outcome observations also matters for interpreting precision. The primary summaries include 568 intervention participants and 605 control participants, drawn from those who completed the assessment. They are not a count of every participant enrolled. Without information about outcomes for those who did not complete the assessment, the interval quantifies uncertainty for the posted comparison but does not account for possible selection related to missing assessments.

12. Planning a Similar Trial: Sample Size

This is a planning illustration, not a re-evaluation of the original trial. It asks: if a new study wanted to detect a difference like the one observed, or a smaller one, how large would it need to be? It is deliberately not “observed power” of the finished trial, which is not a meaningful quantity.

The planning figures are participants per arm, not a total across both groups and not total events. They are illustrations based on the observed effect, not the trial's original design calculation. The stated assumptions are a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. Each planning row considers an effect equal to the stated fraction of the observed effect: 100%, 75%, or 50%.

The 80% power column describes the number of participants per arm that the illustration estimates would be needed to have an 80% chance of detecting the specified effect under those assumptions, if that effect were real. The 90% power column gives the corresponding planning target for a higher chance of detection. Power is conditional on the assumptions and effect size used; it is not a guarantee that an analysis will find a difference. These figures are not observed power.

Smaller effects are more difficult to distinguish from random variation, so the planning illustration requires more participants per arm as the assumed effect becomes a smaller fraction of the observed effect. The figures are not a retrospective judgment that the trial was or was not adequately designed. The original sample-size justification and the assumptions used for the trial's own design are not reported in the registry record. The illustration should be read as a way to explore how hypothetical enrollment targets depend on an assumed effect, not as evidence that the observed result has a particular probability of being true.

13. How the Sample Size Is Derived

Power is the probability that a study of a given size rejects the null hypothesis when a specific true effect exists. It depends on four things: the effect to detect, the variability of the outcome, the significance level (here a two-sided α of 0.05, so z_1−α/2 = 1.960) and the desired power (z_1−β = 0.842 for 80% and 1.282 for 90%).

Formula

n is the number of participants in each arm, Δ is the difference in means to detect and σ is the common standard deviation.

Worked calculation for this trial

1. Observed difference Δ = -0.200; pooled standard deviation σ = 2.193.
2. n = 2 × (1.96 + 0.842)² × 2.193² / -0.200² = 1,887.9, rounded up to 1,888 participants per arm for 80% power.

The smaller-effect rows in the table above repeat the same calculation with the effect multiplied by 0.75 and 0.5 (for ratios, applied to the logarithm of the ratio).

R code

alpha <- 0.05
z_a   <- qnorm(1 - alpha / 2)          # two-sided
delta <- -0.2                      # observed difference in means
sigma <- 2.19332797762236                     # pooled standard deviation
n_per_arm <- function(power, frac = 1) {
  ceiling(round(2 * (z_a + qnorm(power))^2 * sigma^2 / (delta * frac)^2, 6))
}
for (f in c(1, 0.75, 0.5))
  cat(sprintf("%3.0f%% of observed effect: n per arm = %s (80%% power), %s (90%% power)\n", 100 * f,
              format(n_per_arm(0.80, f), big.mark = ","), format(n_per_arm(0.90, f), big.mark = ",")))

# Check against R's built-in function (t-based, so very slightly larger)
cat("power.t.test:", round(power.t.test(delta = delta, sd = sigma, power = 0.80)$n, 1), "\n")

# Simulation: exact two-sample t-test using simulated means and variances
set.seed(3954535)
n    <- n_per_arm(0.80)
reps <- 20000
m_diff <- rnorm(reps, delta, sigma * sqrt(2 / n))             # difference in sample means
s2     <- sigma^2 * rchisq(reps, 2 * (n - 1)) / (2 * (n - 1))   # pooled sample variance
tstat  <- m_diff / sqrt(s2 * 2 / n)
cat(sprintf("Simulated power at n = %s per arm: %.3f (%d replicates)\n", format(n, big.mark = ","), mean(abs(tstat) > qt(0.975, 2 * (n - 1))), reps))

Output from running the code

100% of observed effect: n per arm = 1,888 (80% power), 2,528 (90% power)
 75% of observed effect: n per arm = 3,357 (80% power), 4,494 (90% power)
 50% of observed effect: n per arm = 7,552 (80% power), 10,110 (90% power)
power.t.test: 1888.9 
Simulated power at n = 1,888 per arm: 0.801 (20000 replicates)

The simulation draws many sets of group means and variances at the planned sample size and counts how often the two-sample t-test rejects the null. It should land close to the target power; at very small required sample sizes it can fall below it, because the normal approximation is rough there.

14. Multiplicity, Missing Data and Analysis Populations

The registry record lists 64 outcome measures and 40 posted secondary outcomes. It does not report how the outcomes were prioritized, whether all were prespecified before data collection, or whether a hierarchy or multiplicity adjustment was used. When many outcomes are tested, the chance of obtaining one or more apparently notable results by chance can increase. In general, multiplicity procedures address that issue by adjusting how evidence is evaluated across a family of comparisons. The record does not identify any such procedure.

No registry-reported analyses are listed. The numerical comparisons accompanying the posted group summaries are unadjusted recalculations rather than a record of a sponsor-reported statistical model. The trial's original statistical plan, including its primary analysis method, handling of multiple outcomes, and any planned interim analyses, is not reported in the registry record. Whether any interim analyses were conducted is also not reported in the registry record.

The primary analysis population is explicitly limited to participants who completed the assessment at approximately 5 months after baseline. The primary outcome counts are 568 and 605, while the secondary outcome counts vary by measure. The registry record does not report why participants lacked an assessment, whether missingness differed by group, or whether analyses used imputation, weighting, or another missing-data method. It also does not report an intention-to-treat analysis or a sensitivity analysis for missing outcomes.

In general, a complete-assessment analysis describes participants with observed outcomes and can be vulnerable to bias if outcome availability is related to treatment assignment or the outcome itself. The posted results should therefore be understood as completer-based summaries. Without a reported missing-data strategy and analysis population details beyond the stated assessment-completer definition, it is not possible to determine how the results would change under other approaches.

15. 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.

The adverse-event record lists 608 participants at risk for serious adverse events in each group, with 0 affected in the intervention group and 0 affected in the attention-placebo control group. It also lists 608 at risk for other adverse events in each group, with 0 affected in either group. Deaths are listed as 0 in both groups, and no serious adverse-event terms are listed.

These figures describe what the registry record reports for the stated at-risk counts and categories. They do not establish that no participant experienced any problem outside the events captured and reported in this record. The record does not provide further detail on event definitions, collection procedures, duration of safety monitoring, or whether all events were actively solicited. The adverse-event counts should be read as reported observations, not as proof that risk was absent.

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

The primary posted result estimates the difference in average reported past-month drinking frequency between participants assigned to Free2b and those assigned to an attention-placebo website, among participants who completed the assessment at approximately 5 months after baseline. The point estimate is -0.2 times, and its interval includes zero. The result is therefore uncertain in direction and magnitude under the unadjusted comparison. It does not establish a clear difference in this primary endpoint, but it also does not prove that the groups had exactly the same mean.

The result is specific to the measured outcome, the assessment period, the comparator, and the analyzed population. It does not answer whether the groups differed in other substances, amounts consumed, harms, or outcomes at other times. It also does not compare the intervention with no access to a website, because both groups received websites with entertainment features and resources. The intervention package included 9 sessions, while the control description does not list those sessions.

Statistical uncertainty and practical importance are distinct. The registry record does not define a threshold for an important change in drinking frequency, so the point estimate cannot be labeled meaningful or unimportant on that basis. The interval communicates uncertainty about plausible values under its calculation, rather than a treatment recommendation or a guarantee about what an individual participant would experience.

The trial's randomized design is relevant to comparing groups, but the posted result is a completer-based summary and not a reported adjusted analysis. The record does not provide enough detail to assess missing-data assumptions, adherence, or the full analysis plan. Results from the secondary measures are also not equivalent to the primary endpoint and should be interpreted cautiously given the number of outcomes and lack of a reported multiplicity strategy.

Accordingly, the safest statistical reading is narrow: the posted summaries show a modest negative point estimate with an interval spanning zero for the primary mean contrast at the stated assessment among completers. They do not settle whether an effect exists, establish a causal mechanism, or show what would happen in populations or settings not represented by the analyzed participants.

17. Why This Trial Matters Statistically

This trial raises a statistical question that is common in behavioral prevention research: how to distinguish a group difference from the substantial variation in participants' reported behavior. The primary outcome is a count of drinking occasions summarized as a mean, and the posted standard deviations indicate broad variation around those group averages. Reporting an estimate alongside an interval is important because the point estimate alone would not show how much uncertainty remains.

The attention-placebo website also gives the comparison a specific interpretation. The contrast is between a multi-session intervention package and a website that also provides entertainment features and resources. It is not a comparison with no contact. This distinction matters when describing what the estimated contrast represents and when considering which aspects of the intervention package it may reflect.

The record illustrates why enrollment, outcome completion, and analysis population should not be treated as interchangeable. Overall enrollment was 1216, while the primary summaries include only participants who completed the assessment at approximately 5 months after baseline. The difference in what those counts represent is central to interpreting precision and potential missing-data concerns.

Finally, the record contains many outcome measures but no reported analysis details for multiplicity, missing-data handling, or the original design assumptions. Those omissions limit how strongly unadjusted comparisons can be interpreted, particularly for secondary outcomes. Clear reporting of analysis plans and outcome availability would help readers distinguish confirmatory evidence from exploratory patterns and assess how robust the estimates are to analytical choices.

18. 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 primary outcome is based on participants' reports of drinking in the past month. Self-report can be informative about behavior, but the registry record does not report validation against another source, the instrument's full wording, or procedures for checking responses. The outcome also captures frequency, not drinking quantity or consequences, and should not be treated as a measure of all substance use.

The primary endpoint is assessed at an average of 5 months from study onset, and the analysis includes only assessment completers. The record does not give a complete account of missing assessments, the reasons for noncompletion, or any method used to address missing outcomes. This limits interpretation of how well the posted group summaries represent all randomized participants.

The control group received a website with entertainment features and resources, so the study estimates a contrast against that active digital comparison rather than against no program. The record does not report actual engagement, adherence, or exposure to the sites, making it difficult to assess how participants used the assigned materials. The description of the intervention includes 9 sessions, but details about their delivery and content are not reported in the registry record.

The brief description refers to a nationwide sample, but the achieved geographic distribution and recruitment details are not reported in the registry record. Generalizability beyond the analyzed participants cannot be assessed from the posted summaries alone. The identity of the person or people masked, the original sample-size justification, the statistical analysis plan, and any missing-data or multiplicity procedures are also not reported. These gaps make it difficult to evaluate how the posted unadjusted comparisons relate to the intended confirmatory analysis.

19. Related Tutorials

20. Related Statistical Calculators

21. Sources

• ClinicalTrials.gov: Preventing Drug Abuse Among Sexual Minority Youth — NCT03954535.
• Registry record last updated 2025-08-26; status last verified 2025-08.
• Results first posted 2025-08-26.
• 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.

22. Record Summary

The registry record describes a completed randomized, single-masked, parallel-group behavioral trial sponsored by Columbia University. It enrolled 1216 participants and compared Free2b, described as 9 sessions plus a website with entertainment features and resources, with an attention-placebo website offering entertainment features and resources. The intended population was sexual minority youth aged 15 to 17.

The primary outcome was the average number of times participants drank alcohol in the past month at an assessment averaging 5 months from study onset. The analyzed population consisted of participants who completed that assessment. The posted unadjusted mean difference was -0.2, with a confidence interval from -0.451 to 0.051 and an unadjusted p-value of 0.118. The registry record does not list sponsor-reported analyses or report the original analysis plan, missing-data approach, or multiplicity strategy.