CARES: A Mobile Health Program for Alcohol Risk Reduction: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for CARES: A Mobile Health Program for Alcohol Risk Reduction (NCT03927482), a randomized na study. The primary outcome is Drinks Per Week.

2. Clinical Question

The central statistical comparison is between the mean outcome in the CARES group and the mean outcome in the Alcohol Education group. In the posted results, the reported mean is lower in the CARES group. The comparison estimate is expressed as Alcohol Education minus CARES, so a positive difference corresponds to a higher average in the Alcohol Education group. The direction of that estimate describes the observed group contrast; by itself, it does not establish why the groups differed or how broadly the result applies.

The registry brief describes a study of an app intended to reduce risky drinking and improve user safety. It also says six-month outcomes would assess changes in drinking-related behaviors. The primary outcome listing specifies a 30-day measurement window, but the timing of the posted primary result within the study follow-up is not reported in the registry record. The record also does not give a detailed account of how participants were recruited or how the under-served population was defined. These details matter when considering which populations the comparison may represent.

3. Trial Design

The registry record classifies the study as randomized, single masked, and parallel. Participants were assigned to one of two intervention groups: CARES Mobile App or Alcohol Education. The trial status is completed, and the registry lists enrollment of 239. The recorded start date is 2018-08-02 and the primary completion date is 2021-09-15. The listed phase is NA.

The brief summary describes two parts of the project. It says Phase 1 would develop and test the app among 40 adult community college students, and Phase 2 would test CARES with 200 adult community college students who drink alcohol. In the second part, the summary describes random assignment to CARES or an alcohol education control condition and 12 weeks of app use. These are statements in the study description. The registry does not report how the listed total enrollment of 239 was divided between project phases, or whether all enrolled participants contributed to the posted outcome groups.

The intervention descriptions distinguish the groups. CARES included use of the smartphone app and text messages intended to support alcohol risk reduction. Alcohol Education involved access to Check Your Drinking, described as giving users normative feedback about their drinking relative to peers. Both interventions are behavioral. The registry says the study was single masked, but does not identify which people involved in the trial were masked or how masking was maintained. The allocation ratio and details of sequence generation are not reported in the registry record.

The results list 118 participants in the CARES group and 121 in the Alcohol Education group for the primary outcome, with the same group counts appearing for the posted secondary outcomes. These outcome-group counts need not explain the complete flow of enrollment: recruitment, withdrawals, outcome availability, and analysis-population definitions are not reported in the registry record. Details of the trial's original sample-size justification, planned statistical analyses, interim analyses, and protocol deviations are also not reported in the registry record. The posted group summaries permit descriptive comparison, but cannot replace those missing design and analysis details.

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 Drinks Per Week, measured over 30 days.

The primary endpoint is titled Drinks Per Week and is measured in drinks. It is described as the average number of drinks based on a 30-day timeline follow-back of all alcohol consumption. A timeline follow-back is a structured way to report alcohol use over a defined period by recalling consumption across that period. In general, summarizing reported consumption as a mean gives a measure of average use across the participants contributing to the group result.

This measure is used because the study concerns alcohol-related behavior and the intervention is intended to support alcohol risk reduction. The group means describe the average reported drinks per week for each arm; the standard deviations describe how much individual observations vary around those means. Neither quantity alone describes every participant's experience. The primary results list a mean of 4.05 drinks in the CARES group and 5.58 drinks in the Alcohol Education group, with standard deviations of 4.83 and 6.29, respectively.

The outcome definition specifies a 30-day reporting window, while the study brief refers to six-month outcomes. The relationship between that follow-up description and the timing of this primary result is not reported in the registry record. The record does not specify the scoring details used to convert the timeline follow-back reports into the weekly average, how incomplete reports were handled, or whether the posted endpoint represents a particular scheduled visit. Those details limit how precisely the measurement process can be reconstructed.

7. Results Posted in the Registry

• Alcohol Education: mean 5.58 versus 4.05 on control; difference 1.53 (95% CI 0.11 to 2.95), standardized effect d = 0.27; z-test p = 0.035.

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.

Situational Confidence

• 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:
  • Alcohol Education: mean 74.60 versus 72.60 on control; difference 2.00 (95% CI -3.21 to 7.21), standardized effect d = 0.10; z-test p = 0.451.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Alcohol Education and the control. The interval gives the range of effects that remain compatible with the data.

Protective Strategies

• 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:
  • Alcohol Education: mean 3.48 versus 3.65 on control; difference -0.17 (95% CI -0.35 to 0.01), standardized effect d = -0.24; z-test p = 0.064.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Alcohol Education and the control. The interval gives the range of effects that remain compatible with the data.

The posted secondary outcomes cover situational confidence and protective strategies, both measured at 12 weeks. Together, they add context about behavioral and confidence-related measures, but they do not form a single consistent pattern that independently reinforces the primary drinks-per-week contrast. The two comparisons point in different directions, and their uncertainty intervals include no difference. This broad pattern is more informative than treating either secondary measure as a standalone verdict.

These measures are conceptually related to alcohol risk reduction, but they are not substitutes for the primary outcome. A difference in a confidence score or protective-strategy score does not establish a corresponding change in drinking behavior. Conversely, the primary outcome contrast does not establish a change in every behavioral process that might influence alcohol use. Each outcome captures a distinct aspect of the study's stated aims.

The registry lists no reported analyses for the secondary outcomes. The displayed comparisons are unadjusted re-calculations from the group summaries, and the record does not say whether these endpoints were prespecified, how they were scored beyond their description as scores on a scale, or whether missing values were handled consistently across measures. It also does not report an adjustment for testing multiple outcomes. These gaps mean that the secondary findings should be treated as exploratory context rather than as confirmatory evidence.

In general, when several outcomes are examined, some estimates may appear more or less unusual because of chance alone. A declared analysis hierarchy or multiplicity adjustment can help readers distinguish a primary confirmatory test from additional exploratory comparisons. Neither is reported in the registry record. The secondary outcomes can therefore help characterize what was measured beyond drinking behavior, but their unadjusted results carry less weight than a prespecified and fully described analysis would. They do not resolve uncertainty about the primary result or explain why the group averages differed.

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

• Situational Confidence
• Protective Strategies

9. Statistical Interpretation

10. Statistical Methods Explained

The posted primary and secondary outcomes are continuous measures summarized as group means and standard deviations. The registry also includes calculated contrasts between the CARES Mobile App and Alcohol Education groups. The explanations below describe the general statistical ideas relevant to those summaries; the registry record does not report the trial's full statistical analysis plan.

What does a difference in means describe?

A mean is the arithmetic average of a set of observations. A difference in means compares the two group averages on the outcome's original scale. For the primary outcome, the contrast is expressed as Alcohol Education minus CARES. This sign convention means a positive value corresponds to a larger mean in Alcohol Education; a negative value would correspond to a larger mean in CARES. The reported primary contrast is 1.53 drinks, so its positive sign matches the higher posted mean in Alcohol Education.

A difference in means is directly interpretable in the outcome's units. Here, the units are drinks per week as defined by the outcome description. The contrast does not indicate that each person in one group consumed that amount more than each person in the other group. It compares group averages, and substantial individual variation may exist within both arms. The group standard deviations, 4.83 and 6.29 for the primary outcome, provide information about that variation but do not show the full distributions.

How should a confidence interval be understood?

In general, a confidence interval expresses the uncertainty around an estimate using a range of values generated by a statistical procedure. A narrow interval indicates greater precision under the procedure's assumptions than a wider interval. For the primary mean difference, the interval extends from 0.110 to 2.95. It communicates that the exact size of the contrast is not known precisely from the estimate alone. The interval's endpoints are not a range of participant-level responses, and they do not imply that every value inside the interval is equally plausible.

The registry record does not specify the confidence level or the procedure used to construct the posted intervals. The comparisons are simple unadjusted re-calculations from group summaries, rather than a registry-listed sponsor analysis. Accordingly, the interval can be discussed as a measure of uncertainty attached to that calculation, but it should not be treated as evidence that the trial's original planned analysis used the same method.

What does the p-value contribute?

In general, a p-value measures how unusual data at least as incompatible with a specified null hypothesis would be under a statistical model. It is not the probability that the null hypothesis is true, the probability that the result arose by chance, or a measure of the size or importance of a difference. The unadjusted primary p-value is 0.035. It should be interpreted alongside the mean difference, the interval, the group summaries, and the uncertainty about analysis choices. The registry does not report a multiplicity adjustment or a prespecified testing hierarchy.

Random assignment is intended to make treatment groups comparable on average with respect to factors that might influence outcomes. It can support causal interpretation when the design and conduct are sound, outcomes are handled appropriately, and the analysis corresponds to the randomized assignment. The registry identifies randomized allocation, but does not provide the sequence-generation details, allocation concealment methods, or full account of outcome availability. Randomization alone does not answer those questions or eliminate uncertainty in a finite sample.

Why is spread relevant?

The standard deviation describes typical dispersion around a mean, in the same units as the measure. Greater spread can make a group mean more uncertain when the sample size is fixed. In general, a difference that is small relative to the variability can be estimated less precisely than a larger difference under otherwise similar conditions. The primary outcome standard deviations are 4.83 in CARES and 6.29 in Alcohol Education, compared with a mean difference of 1.53. The standardized effect-size estimate listed for this contrast is 0.272. Such a standardized summary expresses a mean contrast relative to pooled variability, but it does not replace the original-unit interpretation.

Many simple mean-comparison procedures rely on assumptions about the distribution of observations and the behavior of the estimated mean. The registry does not report diagnostic checks, transformations, robust methods, or distributional analyses for this outcome. Alcohol consumption reports can have uneven distributions in general, and when a distribution is strongly skewed or includes influential values, the mean and its standard error may be sensitive to those observations. The posted summaries do not permit assessment of whether that occurred in this trial.

Over-dispersion is a general term for variability that is greater than a statistical model expects. It can lead model-based standard errors and intervals to be too small if not accommodated. The trial record does not describe a count model or report an over-dispersion assessment; the outcome is posted as a mean with standard deviation, and the comparison is a difference in means. It would therefore be inappropriate to infer that a particular count-model correction was used.

What does unadjusted mean in this setting?

An unadjusted comparison uses the group summaries without accounting for other measured characteristics or a broader model. This can be useful for a direct description of the observed group contrast. In a randomized study, an unadjusted contrast may be a reasonable summary, but the choice between unadjusted and adjusted analysis is ordinarily part of the planned statistical approach. The registry lists no reported analyses and does not state whether the posted group summaries are the intended primary analysis or a secondary summary. No adjustment for baseline values, covariates, or missingness should be assumed.

The registry lists four outcome measures in total and identifies two posted secondary outcomes. It does not report the full outcome set or the testing plan for those measures. When multiple outcomes or comparisons are examined, the chance of at least one apparently unusual result can increase. In general, multiplicity procedures or a clear hierarchy of hypotheses can help control interpretation, but no such procedure is reported. This is why unadjusted p-values, especially for secondary outcomes, should not be treated as definitive evidence in isolation.

How should the calculated planning estimates be distinguished from trial results?

The sample-size illustrations are calculations based on the observed effect, not the trial's own original design assumptions. The listed assumptions are a two-sided significance level of 0.05, equal allocation between groups, and a normal approximation. The calculations consider detecting the full observed effect and smaller fractions of that effect, and state the required number of participants per arm for the specified power levels. These are planning scenarios, not trial outcomes and not evidence that the completed study had a particular probability of detecting an effect. The registry's original sample-size justification is not reported in the registry record.

In general, a power calculation asks how often a planned statistical test would meet its decision criterion if a particular effect and a set of assumptions were true. It depends on the effect size being targeted, variability, allocation, significance threshold, and analysis method. A post hoc calculation using an observed effect does not reveal the actual probability that the trial detected a true effect. The sample-size figures should therefore be read as hypothetical planning quantities only, with the assumptions attached to them, rather than as a retrospective rating of the trial.

11. Confidence Intervals and Precision

Precision concerns how tightly the data estimate a group contrast. The primary mean difference is 1.53 drinks, and its interval runs from 0.110 to 2.95. This interval gives a materially wider range than the single point estimate and shows that the exact magnitude of the contrast remains uncertain. Under the calculation used, the range is on the positive side of zero, but its lower endpoint is close to zero relative to the scale of the outcome. The result therefore supports a directional description of the posted means more strongly than a precise statement about the size of the difference.

The interval's endpoints should not be interpreted as limits on individual responses. They refer to uncertainty around the estimated difference in group means. The standard deviations of 4.83 and 6.29 describe spread among observations within the groups; the interval describes uncertainty in the comparison estimate. These are related but distinct quantities. A reader should not use the group standard deviations as if they were confidence intervals or assume that the mean difference captures the range of outcomes among participants.

The registry does not report the confidence level or the method used for the intervals, and no sponsor analysis is listed. The displayed comparison is an unadjusted re-calculation from posted summaries. Precision therefore depends on the assumptions of that calculation, and may differ from an estimate under a prespecified adjusted model or a procedure accounting for missing data. The record does not report enough analysis detail to compare such alternatives.

The posted sample sizes for the primary outcome are 118 in CARES and 121 in Alcohol Education. Larger samples can generally yield more precise estimates when other features remain similar, but precision also depends on outcome variability and the analysis approach. The posted primary outcome standard deviations indicate appreciable variation in reported drinks per week. No further precision can be inferred from the group sizes alone, and the registry does not report the number of participants with missing primary outcome data or the completeness of follow-up.

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 sample-size illustrations report participants per arm for the continuous outcomes. They are not total-event counts: the record's planning assumptions specify per-arm quantities for outcomes other than time-to-event outcomes, and these outcomes are continuous. The calculations use 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 original trial design assumptions and sample-size justification are not reported in the registry record.

The 80% power column gives the calculated number of participants per arm for a planning scenario intended to have an 80% chance of meeting the specified test criterion if the assumed effect and model conditions hold. The 90% power column gives the corresponding per-arm planning quantity for a higher target chance under those same assumptions. These columns describe hypothetical designs; they do not report how many participants were actually analyzed or the probability that this completed trial detected a real effect.

For each outcome, the planning scenarios range from 100% of the observed effect to 75% and 50% of that effect. A smaller effect is harder to distinguish from sampling variation, so the calculations require more participants per arm as the assumed effect becomes smaller. This relationship is general and is reflected in the posted illustrations. The values should be read as planning estimates conditional on the assumptions, not as a recommended design or a claim about the adequacy of this trial. This is not observed power.

The plotted curve relates required sample size to the effect size under the same planning framework. It helps convey the direction of the relationship: targeting a smaller effect generally calls for a larger sample. It does not add information about the trial's original planning, because that justification is not reported in the registry record. The calculations use the observed effect as their starting point, so uncertainty in that estimated effect also limits how literally the illustrations can be used for future planning.

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 = 1.530; pooled standard deviation σ = 5.617.
2. n = 2 × (1.96 + 0.842)² × 5.617² / 1.530² = 211.6, rounded up to 212 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 <- 1.53                      # observed difference in means
sigma <- 5.61687229963258                     # 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(3927482)
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 = 212 (80% power), 284 (90% power)
 75% of observed effect: n per arm = 377 (80% power), 504 (90% power)
 50% of observed effect: n per arm = 847 (80% power), 1,133 (90% power)
power.t.test: 212.5 
Simulated power at n = 212 per arm: 0.797 (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 lists four outcome measures in total, with two secondary outcomes posted alongside the primary outcome. It lists no reported analyses. The record does not explain which outcomes were designated as confirmatory, whether there were additional planned comparisons, or whether any hierarchy governed the testing. A multiplicity adjustment is not reported in the registry record. Because the primary p-value is unadjusted and the secondary comparisons are also unadjusted, the p-values do not account for the possibility of examining multiple outcomes or analyses.

In general, testing several hypotheses can increase the chance of obtaining at least one result that appears unusual even when there is no underlying effect on the outcomes being tested. Multiplicity methods address that issue in different ways, but which method is suitable depends on the design and analysis plan. No particular correction should be presumed for this study. The primary comparison can still be described as posted, but its p-value should not be read as though the registry had documented a broader testing strategy.

Missing-data handling is also not reported. The posted outcome groups list 118 participants for CARES and 121 for Alcohol Education, but the registry does not say how many were randomized, how many completed the intervention or follow-up, whether every participant contributed an outcome, or what approach was used for unavailable measurements. The enrollment total of 239 does not by itself answer these questions. A complete-case summary, an imputed analysis, or another method cannot be inferred from the group counts.

The analysis population is not defined in the registry record. It is therefore unclear whether the posted comparison follows assignment regardless of intervention use, includes only participants with observed outcomes, or uses another population. The record does not report intention-to-treat procedures, per-protocol criteria, baseline adjustment, or sensitivity analyses for missingness. These details can affect both the interpretation of a randomized comparison and the precision of its estimate. Without them, the group summaries should be treated as descriptive posted results rather than as a fully documented account of the trial's intended analysis.

Interim analyses and stopping rules are also not reported in the registry record. No conclusion about their presence or absence should be drawn from the completed status alone. The record likewise does not state whether the statistical analysis plan was finalized before outcomes were examined. These omissions prevent a complete assessment of multiplicity, analysis flexibility, and how closely the posted results correspond to prespecified hypotheses.

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 registry records zero serious adverse events and zero deaths in both groups. It also records zero other adverse events in each group. The counts at risk are 118 for CARES Mobile App and 121 for Alcohol Education. No serious adverse-event terms are listed. These figures describe what is recorded in the registry, not a complete account of every possible symptom, event, or burden during the study.

Zero recorded events does not establish that no adverse experiences occurred. The registry does not report the safety surveillance period, event definitions, collection procedures, severity grading, relatedness assessment, or whether the figures include all enrolled participants or only the listed group counts. It also does not describe how adverse-event information was solicited or whether events could have gone unreported. Those details are necessary to interpret the scope and completeness of safety reporting.

The counts should therefore be read narrowly: the posted record lists no serious events, other events, or deaths for the numbers at risk shown. They cannot establish that the app or education condition was free of risk, and they do not allow comparison of event rates beyond the reported counts. No medical conclusion follows from these registry figures alone.

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

The posted primary outcome shows lower mean drinks per week in the CARES group than in the Alcohol Education group. The estimated difference, expressed as Alcohol Education minus CARES, is 1.53 drinks, with an interval from 0.110 to 2.95 and an unadjusted p-value of 0.035. This is a description of the result in the registry and of its statistical uncertainty. It is not, by itself, a judgment that either intervention is effective or ineffective.

The point estimate is not the only relevant feature. The interval indicates uncertainty about the size of the group contrast, while the relatively large standard deviations show that individual outcome values varied within both groups. The mean comparison does not say that all participants had outcomes near their group average, nor does it describe how many individuals changed their drinking behavior. The registry does not provide participant-level data or the distribution of changes needed to answer those questions.

Random assignment supports comparison between groups, but interpretation also depends on implementation, outcome availability, analysis choices, and adherence to the planned approach. The registry does not report the detailed analysis plan, missing-data methods, participant flow, or the timing of the posted primary outcome relative to the six-month follow-up described in the brief summary. It is not possible to resolve these uncertainties from the posted means alone.

The study concerns adult community college students who drink alcohol, as described for its Phase 2 component, and is titled for an under-served college population. Further characteristics of participants and recruitment are not reported in the registry record. The result should not automatically be generalized to other age groups, settings, or populations, and the record does not establish that the enrolled participants represent all students in the target population.

The secondary outcomes offer additional behavioral context, but their directions are mixed and their analyses are unadjusted. They do not provide a clear basis for broad conclusions about mechanisms or outcomes beyond those measures. The safety record lists zero adverse events and deaths in the group counts shown, but incomplete reporting details prevent interpreting this as proof that no adverse experiences occurred. The registry results are best treated as a limited statistical record of group summaries and contrasts, not as individualized guidance or a complete evaluation of every possible benefit and risk.

17. Why This Trial Matters Statistically

This trial is statistically interesting because it compares a mobile health intervention with an active online alcohol education condition rather than with no intervention. The comparator is described as providing normative feedback about drinking relative to peers. That context matters: the primary contrast is between two behavioral approaches, and the posted estimate describes their relative group means. It should not be interpreted as a comparison with no support or no information.

The primary endpoint is a continuous measure of reported drinking, and the registry provides both group means and standard deviations. This makes it possible to discuss the difference in the outcome's original units as well as its uncertainty. The interval and unadjusted p-value show why statistical interpretation should consider effect magnitude and precision together rather than rely on a single threshold. The variation in group outcomes also underlines that average contrasts do not describe every participant.

The record also illustrates the distinction between outcome results and study planning. Its sample-size illustrations use the observed effect and explicit hypothetical assumptions, while the original sample-size justification is not reported. Those calculations can inform how sample requirements change with a targeted effect, but they cannot be taken as evidence about the design decisions made before the trial began or the probability that the trial detected a true effect.

Finally, the registry makes clear how much matters for interpretation beyond a numerical contrast. The available record identifies randomization and masking status, yet leaves the masked party, missing-data procedures, analysis population, multiplicity plan, and detailed statistical plan unspecified. Those omissions do not erase the posted result; they define the limits of what can be concluded from it. For clinical researchers, the trial is a useful example of why transparent methods and outcome reporting are needed alongside group estimates.

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.

Several limitations specific to this trial constrain interpretation. The registry describes a project with a Phase 1 development component and a Phase 2 test, but does not explain how the total enrollment of 239 relates to the brief summary's descriptions of 40 participants in Phase 1 and 200 in Phase 2. The posted outcome-group counts are 118 and 121, but participant flow from enrollment to outcome measurement is not described. The record does not state whether these group counts represent all randomized participants or only those with an available outcome.

The primary outcome is described using a 30-day timeline follow-back, whereas the brief summary refers to six-month outcomes. The registry does not clarify the timing of the posted primary results relative to that follow-up description. It also does not report the exact assessment schedule, the timing of baseline measurements, or the scoring procedures used to derive the weekly average. The secondary outcomes are listed at 12 weeks, but their scoring details and completeness are not described.

The trial is classified as single masked, but the registry does not say who was masked. It does not report sequence generation or allocation concealment procedures, adherence to either intervention, or how much participants used the app or online education material. These are relevant features for understanding how the assigned interventions were delivered and how closely the observed comparison reflects the intended contrast.

Statistical reporting is limited: the registry lists no reported analyses and does not give the original sample-size justification, an analysis population, missing-data methods, covariate adjustment plan, multiplicity procedure, or interim-analysis approach. The posted comparisons are unadjusted re-calculations from group summaries. Their interpretation may differ from a prespecified analysis that accounts for design features or missing outcomes. The registry record also does not provide participant characteristics in enough detail to assess representativeness or variation in response across subgroups.

Finally, the safety counts have no reported surveillance window or detailed event-collection methods. Zero recorded events should not be treated as proof that no adverse experiences occurred. These reporting gaps limit assessment of safety, precision, and generalizability, and they make the registry record an incomplete account of the trial's conduct and analysis.

19. Related Tutorials

20. Related Statistical Calculators

21. Sources

• ClinicalTrials.gov: CARES: A Mobile Health Program for Alcohol Risk Reduction — NCT03927482.
• Registry record last updated 2025-03-26; status last verified 2025-03.
• Results first posted 2022-08-31.
• 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 completed randomized parallel study compared the CARES mobile app with online Alcohol Education using Check Your Drinking. The registry lists 239 participants enrolled and identifies an under-served college population in the study title. Its brief summary describes app development and testing among adult community college students, including a Phase 2 comparison for students who drink alcohol.

The primary outcome is mean drinks per week based on a 30-day timeline follow-back. Posted results show means of 4.05 drinks for CARES and 5.58 for Alcohol Education; the unadjusted mean difference, expressed as Alcohol Education minus CARES, is 1.53 drinks, with an interval from 0.110 to 2.95 and p = 0.035. Situational confidence and protective strategies are posted as secondary outcomes at 12 weeks. The record lists zero serious adverse events, other adverse events, and deaths in both groups.

The registry does not report a full statistical analysis plan, missing-data methods, analysis populations, or multiplicity procedures. The posted contrasts should be interpreted with those limits in mind.