Brief Electronic Intervention for Heavy Drinking and Sex Risk Among MSM Seeking HIV Testing: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Brief Electronic Intervention for Heavy Drinking and Sex Risk Among MSM Seeking HIV Testing (NCT03435783), a randomized na study. The primary outcome is Number of New Sex Partners, Past 30 Days.

2. Clinical Question

3. Trial Design

The record describes a randomized pilot test of a behavioral intervention for men who have sex with men seeking HIV testing. Allocation is listed as randomized, and masking is listed as single. The record does not say who was masked. It lists the model as single-group, despite also listing randomized allocation and two behavioral interventions; this classification is not explained in the registry record.

The intervention is called Brief Electronic Intervention in the intervention listing and Brief Web-based Intervention in the outcome results. The comparator is called Attention-Matched Control in the intervention listing and Standard Post-test Counseling in the outcome results. Arm descriptions are blank, so details about what participants received are not reported in the registry record.

The recorded enrollment is 40. The primary outcome results list 20 in each arm. The study is marked completed; its recorded start date is 2017-10-13 and its primary completion date is 2018-03-03. Phase is listed as not applicable. Brown University is listed as sponsor. The brief summary frames the work as a pilot intended to explore potential effects on alcohol- and HIV-related behavioral outcomes; it does not establish an efficacy conclusion.

The registry lists five outcome measures and no posted analyses. No further details of the randomization procedure, concealment, follow-up schedule, or the original design assumptions are reported in the registry record. The planning calculations associated with the posted result are illustrations based on the observed effect, not a description of the trial's original design.

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

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 Number of New Sex Partners, Past 30 Days, measured over Past 30 days.

The primary endpoint is the number of new sex partners reported for the past 30 days. It is recorded as a continuous outcome summarized by a mean and standard deviation. The registry says it was assessed by self-report using Timeline Follow Back. Results are available for standard post-test counseling and the brief web-based intervention.

A count of new partners is relevant to the stated research focus on sexual risk behavior because it describes one aspect of recent sexual behavior. It is not, by itself, a measure of infection, transmission, condom use, or other sexual practices. Nor does this endpoint measure heavy drinking, although alcohol is part of the broader research objective. Those distinctions matter when interpreting the result: the posted primary outcome addresses the reported number of partners over a defined recall period, not every behavior or health outcome named in the study rationale.

General statistical context: a mean gives the average outcome in an arm, while the standard deviation describes how spread out individual observations are around that average. For a count outcome, the mean is a compact summary, but it does not show the distribution of counts or whether a small number of high values influenced the average. Distributional details are not reported in the registry record.

7. Results Posted in the Registry

• Brief Web-based Intervention: mean 1.37 versus 2.53 on control; difference -1.16 (95% CI -2.82 to 0.50), standardized effect d = -0.43; z-test p = 0.172.

8. Statistical Interpretation

9. Statistical Methods Explained

What does a difference in means represent?

The primary outcome is summarized using an arithmetic mean in each arm. A mean difference expresses the intervention-arm mean relative to the control-arm mean. The reported estimate is -1.16, with the comparison identified as intervention versus control. Its negative sign indicates that the observed intervention-arm mean is lower. The estimate describes the contrast in these posted group summaries; it does not, on its own, establish why the groups differed or whether the difference reflects an intervention effect.

General statistical context: a mean difference keeps the outcome in its original units, which are new partners during the past 30 days. This makes the contrast directly interpretable in relation to the measured outcome. It also means that the estimate depends on how the count outcome is distributed. Means and standard deviations alone cannot show the full pattern of participants' values, such as whether most values were similar or whether a smaller number of larger counts contributed substantially to the average.

What do standard deviations tell us?

The reported standard deviation is 3.6 for standard post-test counseling and 1.2 for the brief web-based intervention. A standard deviation describes the spread of observed values around an arm's mean. It is not the uncertainty interval for the mean, and it is not the standard error of the difference. The registry also reports a pooled standard deviation of 2.683 for the comparison, but does not explain further details of the calculation.

General statistical context: when an outcome has a skewed distribution or a large number of low values alongside some high values, the mean and standard deviation may not fully describe its shape. Count outcomes can have such features, but the actual distribution in this trial is not reported. It would therefore be inappropriate to assume a particular distribution or to substitute an unreported median or count model for the posted analysis.

How should the interval and p-value be read?

The reported interval for the mean difference has a lower endpoint of -2.823 and extends to a positive upper endpoint. Because zero lies between its endpoints, the interval does not rule out no difference. It also allows differences in either direction. The unadjusted p-value is 0.172. The registry does not identify the statistical test, adjustment variables, or confidence level used for this comparison, so those details should not be inferred from the reported values.

General statistical context: a confidence interval combines the estimated contrast with its uncertainty under a statistical method. A narrower interval indicates greater precision under that method; a wider interval indicates less precision. An interval that includes zero for a difference can still include potentially meaningful effects in either direction. The interval does not say that every value inside it is equally plausible, and it does not account automatically for design limitations, measurement error, or missing outcome data.

General statistical context: a p-value is conditional on a null model and the analysis used. It is not a measure of effect size, clinical importance, or the probability that a particular explanation is correct. Its interpretation depends on the analysis plan and assumptions, and those details are not fully reported in the registry record. The p-value should be read alongside the estimate and interval rather than as a standalone verdict.

What is the role of standardized effects?

The comparison record also includes a standardized mean difference and a pooled standard deviation. A standardized effect expresses a difference relative to a measure of variability, rather than in the original outcome units. General statistical context: standardization can help compare outcomes measured on different scales, but it can be less directly interpretable for a clinical audience than a difference in the original units. It also depends on the variability measure used. The registry does not report a separate inferential conclusion based on the standardized effect.

What can be inferred from the available analysis?

The record posts one primary comparison and no analyses. It does not report an adjusted model, analysis population, handling of missing outcome data, or statistical test. The posted means and comparison therefore support a descriptive reading of the reported result and the uncertainty attached to it, but not reconstruction of unreported analysis choices.

General statistical context: over-dispersion occurs when variability in count data exceeds what a simple count model assumes. In such settings, a model that allows extra variability may give more appropriate standard errors than a more restrictive model. This is a general consideration for count outcomes, not a claim about the analysis used or the distribution observed in this trial. No count model or assessment of over-dispersion is reported in the registry record.

The planning calculations are separate from the analysis of the trial's observed outcome. They show participant counts under stated assumptions and different fractions of the observed effect. They are prospective illustrations based on the posted point estimate, not evidence that the pilot had a particular probability of detecting an effect.

10. Confidence Intervals and Precision

The reported mean difference is -1.16 new partners, intervention compared with standard post-test counseling. Its interval extends from -2.823 to a value above zero, so it includes zero and values on both sides of zero. This span indicates that the point estimate should not be treated as a precise determination of the direction or magnitude of the difference.

The interval allows for a decrease in the intervention arm relative to control, but it also allows for no difference and for an increase. The point estimate alone is not enough to distinguish among these possibilities. The unadjusted p-value of 0.172 is consistent with the same caution: the reported comparison does not sharply separate the observed result from a no-difference explanation under the analysis used.

General statistical context: precision concerns the range of estimates compatible with the data and method, not whether an estimate is important. An interval may be informative even when it includes zero, because it displays both the direction and scale of uncertainty. Its interpretation depends on the method and confidence level. The registry record does not state the confidence level or the exact method used to construct this interval.

There are 20 observations listed in each arm, and the observed standard deviations are 3.6 in control and 1.2 in intervention. These summaries provide context for the comparison, but they do not replace the interval or reveal the individual outcome distribution. The interval is the clearest posted summary of uncertainty around the mean difference. The record does not report precision for other outcomes or adjusted estimates.

11. 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 total event counts. They are calculated as illustrations from the observed effect, not as the trial's original sample-size design. The stated assumptions are a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. The effect to detect is set to the stated fraction of the observed effect. The trial's own original design assumptions are not reported in the registry record.

For planning based on 100% of the observed effect, the figures are 84 participants per arm for 80% power and 113 participants per arm for 90% power. For 75% of the observed effect, they are 150 participants per arm for 80% power and 200 participants per arm for 90% power. For 50% of the observed effect, they are 336 participants per arm for 80% power and 450 participants per arm for 90% power.

In plain terms, the 80% column gives the illustrative number per arm under which the stated assumptions would yield an 80% chance of detecting the specified effect if that effect were the true one. The 90% column applies the same idea at a higher target chance of detection. These are planning probabilities under assumptions, not guarantees about a future study. Smaller effects need more participants to detect, as reflected by the larger figures for the smaller fractions of the observed effect.

This is not observed power. The values do not describe the probability that the completed pilot detected its effect, and they should not be read as the trial's original recruitment target.

12. 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 = -1.160; pooled standard deviation σ = 2.683.
2. n = 2 × (1.96 + 0.842)² × 2.683² / -1.160² = 84.0, rounded up to 84 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 <- -1.16                      # observed difference in means
sigma <- 2.68328157299975                     # 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(3435783)
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 = 84 (80% power), 113 (90% power)
 75% of observed effect: n per arm = 150 (80% power), 200 (90% power)
 50% of observed effect: n per arm = 336 (80% power), 450 (90% power)
power.t.test: 85 
Simulated power at n = 84 per arm: 0.795 (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.

13. Multiplicity, Missing Data and Analysis Populations

The registry lists five outcome measures but reports no secondary results. It also lists no analyses. The primary comparison is posted for the number of new sex partners in the past 30 days. Because results for the other listed outcome measures are not posted, the record does not show how findings across outcomes might compare or whether additional analyses were conducted.

No multiplicity procedure is reported in the registry record. Multiplicity refers, in general, to the increased chance of at least one apparently unusual finding when several hypotheses or outcomes are examined. Whether and how this issue was addressed cannot be determined from the posted information. The unadjusted p-value of 0.172 should therefore be understood as the reported unadjusted comparison, not as a value known to account for a broader set of analyses.

The registry record does not report an analysis population, such as whether all randomized participants were included in the primary analysis, or how participants were handled if they had missing outcome data. It does not describe imputation, complete-case analysis, exclusions, or sensitivity analyses. These omissions limit assessment of how missing observations or analysis choices might affect the estimate. They do not show that missing data occurred; the details are simply not reported in the registry record.

General statistical context: a prespecified analysis plan helps distinguish the principal outcome and analysis from exploratory comparisons, and a missing-data approach clarifies how incomplete records contribute to estimates. Without those details, readers should avoid assuming that a particular method was used. The record also does not state whether any interim analyses were planned or conducted.

The trial is listed as randomized and single-masked, but who was masked is not reported. No information is given about allocation concealment or the procedures used to preserve masking. These omissions do not establish that the procedures were inadequate; they limit what can be assessed from the registry record.

14. 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 20 participants at risk in each arm. For standard post-test counseling, serious adverse events are recorded as 0, other adverse events as 0, and deaths as 0. The same figures are recorded for the brief web-based intervention arm. No serious adverse-event terms are listed.

These figures describe the event counts recorded in the registry. They are not a comparison demonstrating that the arms have identical safety profiles, and they do not establish that no events occurred outside the registry's recording or observation procedures. The record does not report the adverse-event definitions, collection method, or the period over which events were assessed.

General statistical context: when event counts are zero, that does not mean the underlying event risk is known to be zero. The amount of information depends on the size of the group and the duration and completeness of observation, among other factors. The registry provides the at-risk counts but does not report enough detail to characterize those other features. These safety figures should therefore be read as the posted counts, without extending them to unreported outcomes or follow-up periods.

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

The posted result is a comparison of self-reported new sex partners in the past 30 days. The observed mean was 1.37 in the brief web-based intervention arm and 2.53 in the standard post-test counseling arm. The reported mean difference was -1.16. The interval extends from -2.823 to above zero, and the unadjusted p-value is 0.172. The result thus shows a lower observed mean in the intervention arm, with uncertainty that includes no difference and a difference in the opposite direction.

This result does not establish that the intervention caused the observed difference. Randomized allocation is recorded, but the registry does not report enough detail about implementation, analysis populations, missing data, or the statistical method to resolve all sources of uncertainty. The result also does not show whether the intervention changed heavy drinking, HIV infection, or other sexual behaviors. The primary endpoint is specifically a self-reported count of new partners over a 30-day period.

The trial is described as a pilot intended to explore potential impacts and inform further intervention development. Its posted outcome can be read as an estimate from this pilot, not as a definitive estimate for other settings or populations. The interval communicates substantial uncertainty about the contrast; it is not a guarantee that the underlying effect lies in a particular range irrespective of study design or measurement.

General statistical context: statistical evidence, clinical relevance, and causal interpretation are related but distinct questions. The size and precision of an estimate address statistical uncertainty; whether the outcome matters to participants or clinical decisions requires context not reported in this record. A randomized design can support causal comparisons when implemented and analyzed appropriately, but the registry entry alone does not document every relevant design and analysis detail.

16. Why This Trial Matters Statistically

This pilot is statistically informative because it illustrates why a point estimate should be read together with its uncertainty. The observed mean difference is negative, but the interval extends across zero. Focusing only on the direction of the estimate would omit an important part of the result: the posted data do not precisely determine whether the difference is negative, absent, or positive.

The outcome is a self-reported behavioral count, and its observed spread differs between the arms. General statistical context: counts can have distributions that are not well described by a mean and standard deviation alone. The posted summaries are useful for comparing group averages, while the absence of individual-level or distributional information limits a fuller assessment of variation and model assumptions. That distinction is especially important when interpreting a pilot result that may inform further research.

The planning illustrations also show how assumptions about the target effect affect recruitment requirements. Under the stated planning framework, smaller fractions of the observed effect correspond to larger participant requirements per arm. This is useful for understanding the sensitivity of a planning exercise, but it does not reveal the trial's original design rationale or establish a definitive target for later work.

Finally, the registry includes limited analysis and missing-data details. That makes transparent reporting of analysis populations, outcome availability, and multiplicity procedures important when interpreting the result. These points do not negate the posted estimate; they define what can be concluded from the available record and what remains uncertain.

17. 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 self-report, and the record does not report additional information about how recall or reporting error was assessed. The result concerns the past 30 days only; other time periods are not represented by this endpoint. The registry does not provide the distribution of individual counts, so the mean and standard deviation cannot show how many participants reported particular values or whether a small number of larger counts influenced the mean.

The arm descriptions are blank, and the record uses different labels for the comparator and intervention across its intervention and outcome listings. It does not describe the intervention content or the control counseling content. The record lists randomized allocation and single masking, but does not identify who was masked or give details of the randomization process.

The primary result includes an interval that crosses zero, and the unadjusted p-value is 0.172. The record does not report the statistical test, confidence level, adjustment variables, or analysis population. Missing-data procedures and the number of participants with outcome data are not reported beyond the arm counts shown with the primary result. No secondary results are posted, despite five outcome measures being listed, and no analyses are listed.

The model is classified as single-group even though the record also describes randomized allocation and lists two arms; the record does not explain this classification. The study is described as a pilot, and the sample-size illustrations are based on the observed effect rather than the trial's original design assumptions. These limitations constrain interpretation of the precision, generalizability, and analysis of the posted comparison.

18. Related Tutorials

19. Related Statistical Calculators

20. Sources

• ClinicalTrials.gov: Brief Electronic Intervention for Heavy Drinking and Sex Risk Among MSM Seeking HIV Testing — NCT03435783.
• Registry record last updated 2019-04-23; status last verified 2019-04.
• Results first posted 2019-04-23.
• 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.

21. Record Summary

NCT03435783 is a completed randomized pilot sponsored by Brown University. The registry lists single masking and a single-group model, though the record does not explain the combination of that model label with randomized allocation and two arms. The study concerns a brief electronic intervention for heavy drinking and sexual risk among men who have sex with men seeking HIV testing.

The primary outcome is self-reported number of new sex partners in the past 30 days, assessed using Timeline Follow Back. The posted means are 2.53 for standard post-test counseling and 1.37 for the brief web-based intervention. The reported mean difference is -1.16; its interval crosses zero, and the unadjusted p-value is 0.172.

The record lists five outcome measures, no secondary results, and no analyses. Adverse-event counts are recorded as zero in both arms. Details on missing data, the analysis population, multiplicity, and the original sample-size assumptions are not reported in the registry record.