Twitter-enabled Mobile Messaging for Smoking Relapse Prevention: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Twitter-enabled Mobile Messaging for Smoking Relapse Prevention (NCT01602536), a randomized na study. The primary outcome is Self-reported 7-day Point Prevalence Abstinence From Smoking That Was Sustained at 60 Days After the Quit Date.

2. Clinical Question

The clinical question is about a behavioral support strategy for people attempting to quit smoking. The experimental arm combined peer interaction in a small private online group with daily prompts, while the control arm did not include Twitter-group assignment. Because smoking cessation aides were provided in both arms, the trial’s primary contrast is the incremental outcome associated with the Twitter-enabled arrangement as implemented, not the standalone effect of a social-media platform.

The endpoint refers to reported abstinence at a specified time after the quit date. It is not a measure of continuous abstinence throughout the entire interval: the stated outcome is seven-day point-prevalence abstinence sustained at the 60-day assessment. The outcome description says it was based on survey, so it represents self-report rather than a separately described biochemical measure. The registry record does not report further details of the survey assessment or outcome verification.

3. Trial Design

The registry record describes a completed, randomized, two-arm trial with parallel groups. Allocation is listed as randomized, masking as none, and the model as parallel. The study title is “Twitter-enabled Mobile Messaging for Smoking Relapse Prevention.” The condition listed is smoking, the sponsor is the University of California, Irvine, and the phase is listed as NA. Enrollment is recorded as 160. The study start is recorded as 2012-01 and primary completion as 2014-06.

The intervention was behavioral. Participants assigned to the Twitter arm were placed in 20-person quit-smoking groups and instructed to use Twitter-enabled interactive peer messaging. Daily messages encouraged interaction, and a bot posted a cessation-related discussion topic each day, timed to the group’s quit date. Participants in that arm also received the baseline smoking cessation aides. Control participants received those aides but were not assigned to a Twitter group. The design therefore combines a shared baseline intervention with an additional social-messaging component in the experimental arm.

The brief summary describes a two-arm trial with N=160 and 20 people per group. The primary outcome record lists 65 participants for Twitter and 70 for control, while the secondary engagement outcome and adverse-event records list 80 per arm. These are the counts recorded for their respective outcomes; the registry record does not explain the differences among the counts or state how missing outcome data were handled. The outcome-specific counts should therefore be read in the context of the measure being reported rather than treated as interchangeable denominators.

The primary endpoint was self-reported seven-day point-prevalence abstinence sustained at 60 days after the quit date. A secondary outcome assessed engagement through tweet volume over 100 days. The record lists two outcome measures, one posted secondary outcome, and zero reported analyses. It does not report masking procedures beyond the NONE designation, implementation fidelity, an analysis population, or a detailed statistical analysis plan. The registry record also does not report the original sample-size justification or planned interim analyses.

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 Self-reported 7-day Point Prevalence Abstinence From Smoking That Was Sustained at 60 Days After the Quit Date, measured over 60 days after quit date.

The primary endpoint counts participants who reported seven-day point-prevalence abstinence from smoking that was sustained at 60 days after the quit date. The registry description specifies a survey-based measure and gives the timeframe as 60 days after the quit date. This is a binary outcome for each participant: the report records how many participants met the stated abstinence definition in each arm.

Point-prevalence abstinence describes smoking status over a defined recent period rather than documenting uninterrupted abstinence over the entire period since quitting. The wording “sustained at 60 days” identifies the assessment timing and the seven-day abstinence criterion, but the registry record does not provide further operational details about the survey questions, verification, or handling of missing responses. Those details should not be inferred from the outcome label alone.

General: a binary endpoint can be useful when the clinical question is whether participants achieved a clearly defined state, rather than how much a continuous measurement changed. Counts make the observed outcome in each randomized arm directly interpretable, while a between-arm effect estimate expresses the contrast in outcome occurrence. The endpoint is relevant to the trial’s relapse-prevention focus because it records abstinence status at the stated post-quit assessment. It does not measure every aspect of quitting, such as duration of abstinence beyond the stated period, changes in smoking intensity, or longer-term status.

Interpretation also depends on how outcomes were collected and which participants contributed data. The registry record identifies self-report and provides outcome-specific participant counts, but does not report whether additional verification was used or why the endpoint counts differ from enrollment. These omissions limit assessment of outcome ascertainment and missing-data effects.

7. Results Posted in the Registry

• Twitter: 40.0% (26/65) versus 20.0% (14/70) on control; risk difference 20.0 percentage points (95% CI 4.8 to 35.2), risk ratio 2.00 (1.15–3.48); unadjusted two-proportion z-test p = 0.011.

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.

Level of Engagement Based on Tweet Volume

• 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:
  • Twitter: mean 58.86 versus 0.00 on control; difference 58.86 (95% CI 43.95 to 73.77), standardized effect d = 1.22; z-test p = <0.001.
• Interpretation:
  • The 95% confidence interval excludes no difference, so the mean is statistically higher for Twitter than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.

The registry lists one posted secondary outcome: level of engagement based on tweet volume over 100 days. It measures participation in social-media messaging rather than smoking status. The overall pattern is therefore a positive contrast in the intervention-linked behavior alongside the primary outcome contrast in reported abstinence. These outcomes address related but distinct questions: engagement can indicate whether the messaging channel was used, while abstinence records the trial’s primary cessation outcome.

Engagement should not be treated as a substitute for the primary endpoint or as proof of a mechanism. The posted summaries do not establish whether participants with greater tweet volume were more likely to report abstinence, whether engagement preceded a change in smoking status, or whether the messaging itself caused either outcome. The intervention bundled group assignment, peer messaging, daily encouragement, and automated topics, so the summaries do not isolate the contribution of any one component.

The secondary result is reported with an unadjusted p-value. Unadjusted secondary tests carry less weight when considered alongside a primary comparison because testing several questions can increase the chance of at least one apparently notable result. The registry record does not report a multiplicity plan, and it lists zero reported analyses. It is not possible to determine whether this secondary outcome was prespecified, whether other secondary tests were conducted but not posted, or whether the inference was adjusted for multiple comparisons.

General: secondary outcomes can provide context, assess feasibility, or suggest directions for further study, but their interpretation depends on how they were selected and analyzed. A result that aligns with an intervention’s intended behavior may be informative without establishing that behavior mediates a clinical outcome. The registry record does not report a formal mediation analysis or a direct statistical relationship between tweet volume and abstinence. Accordingly, engagement adds descriptive context to the trial but does not independently strengthen a causal explanation for the primary finding.

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

• Level of Engagement Based on Tweet Volume

9. Statistical Interpretation

10. Statistical Methods Explained

General statistical explanations follow; they describe the meaning of common methods rather than claiming a particular unreported analysis was used.

What does the primary comparison measure?

The primary outcome is binary at the participant level: a person either met the registered abstinence definition or did not, according to the outcome record. The posted arm summaries are counts of participants meeting that definition and the outcome-specific numbers of participants listed for each arm. A risk is the proportion of participants with the outcome in an arm. A risk difference subtracts one arm’s risk from the other in the stated comparison direction. It is expressed on the absolute probability scale, so zero corresponds to equal risks. The posted risk difference is 0.2.

A relative risk compares the risks by division. A value of one represents equal risks, while values above or below one indicate different relative outcome occurrence according to the comparison direction. The posted relative risk is 2.0. An odds ratio compares the odds, where odds are the probability of an outcome divided by the probability of no outcome. An odds ratio of one represents equal odds. The posted odds ratio is 2.667. These measures describe the same binary comparison using different scales; their numerical values are not interchangeable, and an odds ratio is not a risk difference.

The registry gives confidence intervals for the risk difference, relative risk, and odds ratio. A confidence interval represents uncertainty in an estimate under the assumptions of its statistical procedure. A narrow interval indicates greater precision on that measure’s scale than a wider interval. The record does not state the interval construction method, so no specific method should be attributed to these limits.

How should the p-value be read?

The primary comparison has an unadjusted p-value of 0.011. In general, a p-value describes how incompatible the observed data, or data more extreme, are with a specified null model if that model and the test assumptions hold. It is not the probability that the null is true, nor does it measure the chance that the finding will replicate. “Unadjusted” means the posted value is not presented as adjusted for a family of multiple comparisons. The registry record does not report a multiplicity procedure or provide a detailed test specification.

A p-value and an effect estimate answer different questions. The estimate describes the magnitude and direction of the observed comparison; the p-value describes evidence relative to a null under a testing framework. The interval adds information about precision and a range of values compatible with the data under its procedure. These should be interpreted together rather than treating any one as a complete summary.

What does the engagement summary represent?

The secondary endpoint is mean tweet volume over 100 days, reported as a continuous outcome with standard deviations. The posted mean difference is 58.86, and the reported interval extends from 43.95 to 73.77. The comparison includes a p-value listed as <0.001. The control-arm mean and standard deviation are recorded as 0.0, so the outcome distribution in that arm has no variation in the posted summary. A pooled standard deviation is recorded as 48.1, and a standardized effect size d is listed as 1.224.

General: for a continuous measure, a mean difference expresses an absolute contrast in the units of measurement, while a standardized effect size expresses a difference relative to a measure of variability. Standardization can help compare scales, but it does not replace knowing what the original units mean. A mean and standard deviation alone do not describe every feature of a distribution, especially when observations may be unevenly distributed or highly variable. The registry record does not report a model or any additional distributional diagnostics for tweet volume.

The outcome reports a mean number of social media posts, and its standard deviation in the Twitter arm is 68.02. The record labels the standard deviation directly and does not identify an approximation. Since the control value is zero, the engagement comparison is also a comparison with a group that was not assigned to a Twitter group. The observed difference is therefore closely tied to the intervention assignment and measurement definition; it should not be interpreted as an isolated causal effect of peer support, daily messages, or the bot separately.

Why can variability matter?

General: over-dispersion means that variation in count-like outcomes is greater than a simple model expects. When present, a model that understates variation can produce uncertainty estimates that are too small. The registry describes tweet volume using a mean and standard deviation, but does not report a count model or an over-dispersion assessment. It is therefore not possible to say that a particular over-dispersion adjustment was or was not applied.

Likewise, the registry does not report whether the group-based delivery was accounted for statistically. Participants were organized in 20-person online groups in the study description. In general, outcomes of people sharing a group may be correlated, which can affect uncertainty estimates if a design or analysis treats all observations as independent. This is a general methodological consideration, not evidence that the trial’s analysis failed to account for clustering. The analysis details are not reported in the registry record.

What is known about the analysis plan?

The record lists zero reported analyses and does not provide a statistical analysis plan, named test for the primary result, adjustment variables, subgroup analyses, interim monitoring, or analysis-population definition. Those details are not reported in the registry record. The reported comparisons and p-values can be described as posted, but it is not possible to establish from the record whether they were prespecified, adjusted, or selected from a larger set of analyses. The registry also does not state how missing values were handled, whether all randomized participants were included in the primary analysis, or how participants were classified if abstinence status could not be determined.

These limits matter because analysis choices influence interpretation. For example, a missing-data rule can affect who contributes to an outcome estimate, and a multiplicity strategy can affect how a collection of tests is evaluated. Without those details, readers should treat the registry’s results as reported summaries and avoid inferring unreported design or analysis decisions.

11. Confidence Intervals and Precision

The primary risk-difference estimate is 0.2, with a 95% confidence interval from 0.049 to 0.351. The interval is on the absolute risk-difference scale and lies above zero. Its lower and upper limits indicate that the data are compatible with a range of positive differences under the procedure used; they do not pinpoint a single true value. The registry does not report how the interval was constructed, so its exact statistical assumptions cannot be assessed from the record.

The relative-risk interval runs from 1.148 to 3.484 around the estimate of 2.0. The odds-ratio interval runs from 1.238 to 5.746 around the estimate of 2.667. Both are on relative scales, so their reference value for no between-arm contrast is one rather than zero. The distinct ranges reflect different measures of association and should not be compared as though they were on the same scale. The risk-difference interval can be useful for understanding an absolute contrast, while the relative measures express the contrast in proportional terms.

General: a 95% confidence interval is produced by a method designed to cover the underlying parameter at its stated long-run rate when its assumptions hold. It is not a guarantee that the parameter lies within the particular interval, and it is not a range containing 95% of individual participant outcomes. Precision is reflected in the interval’s width, but whether a range is narrow enough for a decision depends on the question and the consequences of uncertainty.

The secondary mean-difference interval runs from 43.95 to 73.77. It describes uncertainty on the posts-per-participant scale for the mean contrast, not uncertainty about the primary abstinence outcome. Its p-value is listed as <0.001. The registry does not report the interval method for this outcome either.

Confidence intervals do not account automatically for every limitation. They do not resolve uncertainty about missing participants, self-report, analysis choices, generalizability, or multiplicity. The primary endpoint’s listed arm counts are 65 and 70, whereas the total enrollment is 160; the registry record does not explain the difference. Thus, the intervals quantify sampling uncertainty as calculated for the posted comparison, but they cannot by themselves establish that missing-data or measurement concerns are negligible.

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 for 80% power and participants per arm for 90% power. They are not total events: the total-events convention applies to time-to-event outcomes, and the primary endpoint in this trial is a binary participant-count outcome. The planning illustration uses a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. It considers detecting the stated fraction of the observed effect. The registry record does not report the assumptions used for the trial’s original design or its original sample-size justification.

The 80% power column gives the planned number of participants in each arm for a hypothetical design that would have an 80% chance of detecting the specified effect under the illustration’s assumptions, if that effect and model assumptions held. The 90% column describes the corresponding planning target for a 90% chance. These are design-planning quantities, not probabilities that the observed trial result is correct and not estimates of the trial’s achieved power. This is not observed power.

The rows consider 100%, 75%, and 50% of the observed effect. For the primary outcome, the associated participants-per-arm figures are 82 and 109 for the full observed effect, 138 and 185 for 75% of that effect, and 294 and 392 for 50% of that effect, for 80% and 90% power respectively. Smaller effects are harder to distinguish from random variation, so detecting them generally requires more participants under the same planning assumptions.

These calculations are illustrations based on the observed effect, not a reconstruction of the trial’s protocol or a claim about what the investigators planned. They do not explain the enrollment or outcome-specific counts. Their values depend on the stated assumptions and on the observed effect used as the reference. A different target effect or different design assumptions would yield different planning requirements.

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, p₀ and p₁ are the control and treatment proportions, and p̄ is their average.

Worked calculation for this trial

1. Control proportion p₀ = 0.200; treatment proportion p₁ = 0.400; average p̄ = 0.300.
2. First term: z_1−α/2 × √[2p̄(1−p̄)] = 1.96 × 0.6481 = 1.2702.
3. Second term: z_1−β × √[p₀(1−p₀) + p₁(1−p₁)] = 0.842 × 0.6325 = 0.5323.
4. n = (1.2702 + 0.5323)² / (0.2000)² = 81.2, rounded up to 82 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
p0 <- 0.2                          # control proportion
p1 <- 0.4                          # treatment proportion
n_per_arm <- function(power, frac = 1) {
  q1 <- p0 + (p1 - p0) * frac               # smaller effect = fraction of the observed difference
  pbar <- (p0 + q1) / 2
  z_b <- qnorm(power)
  ceiling(round((z_a * sqrt(2 * pbar * (1 - pbar)) + z_b * sqrt(p0 * (1 - p0) + q1 * (1 - q1)))^2 / (q1 - p0)^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 (unrounded n per arm)
cat("power.prop.test:", round(power.prop.test(p1 = p0, p2 = p1, power = 0.80)$n, 1), "\n")

# Simulation: how often does the z-test reject at the 80%-power sample size?
set.seed(1602536)
n    <- n_per_arm(0.80)
reps <- 20000
x0 <- rbinom(reps, n, p0); x1 <- rbinom(reps, n, p1)
ph <- (x0 + x1) / (2 * n)
z  <- ((x1 - x0) / n) / sqrt(ph * (1 - ph) * 2 / n)
cat(sprintf("Simulated power at n = %s per arm: %.3f (%d replicates)\n", format(n, big.mark = ","), mean(abs(z) > z_a, na.rm = TRUE), reps))

Output from running the code

100% of observed effect: n per arm = 82 (80% power), 109 (90% power)
 75% of observed effect: n per arm = 138 (80% power), 185 (90% power)
 50% of observed effect: n per arm = 294 (80% power), 392 (90% power)
power.prop.test: 81.2 
Simulated power at n = 82 per arm: 0.806 (20000 replicates)

The simulation draws many pairs of event counts at the planned sample size and counts how often the two-proportion z-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 does not report a multiplicity strategy. It lists one primary endpoint, one posted secondary endpoint, and zero reported analyses, but those entries do not establish the full set of outcomes or tests that may have been planned or conducted. The primary p-value is explicitly unadjusted, and the secondary p-value is also presented without a reported adjustment. When several hypotheses are tested, the chance of at least one small p-value can increase. General methods for addressing this include defining a hierarchy of outcomes or adjusting inference across a specified family of tests, but the record does not state whether any such approach was planned or used.

Missing-data handling is also not reported. The enrollment is 160, while the primary outcome summaries list 65 participants in the Twitter arm and 70 in control. The record does not explain the difference, identify reasons an outcome may be unavailable, or state whether missing outcomes were excluded, imputed, or classified in another way. It does not report a sensitivity analysis for missing data. These omissions prevent assessment of whether the posted comparison might be affected by differential availability of outcome information.

The record also does not define the analysis population. It does not state whether the primary analysis followed intention-to-treat principles, whether participants were analyzed according to randomized assignment, or whether exclusions were made after randomization. No such population should be inferred solely from the word “randomized.” The record does not report protocol deviations or how they were addressed.

The primary endpoint is self-reported by survey. The registry does not report whether outcome assessors were masked, whether the survey was completed independently, whether smoking status was verified, or how ambiguous responses were handled. Masking is listed as NONE for the trial, but the record does not give additional outcome-assessment details. These gaps concern potential measurement and analysis uncertainty; they do not establish that bias occurred.

Finally, the registry does not report an interim-analysis plan, stopping rule, detailed statistical analysis plan, or the procedures used to account for the group-based setting. Such information is relevant to interpreting inferential results but should not be guessed. The posted estimates remain descriptive of the registered results, while these unreported elements limit a fuller audit of how those estimates were produced.

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 80 participants at risk in each arm. It records zero serious adverse events, zero other adverse events, and zero deaths in the Twitter group, and the same zero counts in control. No serious adverse event terms are listed. These figures describe the events recorded for the stated at-risk groups in the registry record.

A zero count means that no event in the relevant category was recorded in these summaries. It does not prove that the intervention has no risk, nor does it show that rare events could not occur. The amount of information available about safety depends on the number of participants observed, follow-up, event definitions, and completeness of reporting. The registry record does not provide further details about event collection, severity criteria beyond the serious-event category, follow-up duration for safety, or whether events were independently assessed.

The adverse-event denominator differs from the primary outcome counts, which are 65 and 70. The record does not explain the distinction, so those denominators should not be combined or substituted for one another. These summaries should be read as registry-recorded counts, not as a comprehensive safety assessment or a comparison capable of ruling out uncommon harms.

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

The posted primary result is a between-arm comparison of self-reported seven-day point-prevalence abstinence sustained at 60 days after the quit date. The recorded outcome counts and effect estimates favor a higher outcome proportion in the Twitter arm, and the risk-difference interval lies above zero. The unadjusted p-value is 0.011. These figures describe the result in this randomized trial as recorded; they do not establish that every part of the Twitter intervention contributed equally or that the outcome would be the same elsewhere.

The intervention was a package: group assignment, interactive peer messaging, daily encouragement, and automated discussion topics, alongside smoking cessation aides that were also provided to controls. The comparison cannot separate the effects of these individual components. Nor does the engagement outcome prove a pathway from messaging activity to abstinence. The registry does not report an analysis linking an individual’s tweet volume to their primary outcome.

The endpoint is based on self-report and covers the stated seven-day point-prevalence definition at the 60-day assessment. It should not be read as evidence about continuous abstinence for the whole period, long-term cessation, or other smoking outcomes not listed. The registry record does not provide further verification details, the reason for the difference between enrollment and primary outcome counts, or a missing-data analysis.

General: statistical evidence in a trial is conditional on how participants were enrolled, treated, followed, measured, and analyzed. Randomization is intended to make assigned groups comparable on average, but the registry summary alone cannot establish that all post-randomization data were complete or that every relevant analysis choice was appropriate. A confidence interval expresses sampling uncertainty, not uncertainty from every possible source of bias or limited generalizability.

The registry describes people seeking to quit smoking and a small-group online intervention. It does not report enough detail to establish how representative participants were of broader populations, how the intervention was implemented across settings, or whether the support arrangement is reproducible in the same form. The result should therefore be understood as evidence about the recorded trial contrast and endpoint, not a guarantee of outcomes in other circumstances. No treatment recommendation follows from the registry summaries alone.

17. Why This Trial Matters Statistically

This trial raises several useful statistical issues in a compact setting. First, it distinguishes an absolute effect measure from relative measures. The risk difference communicates the contrast in outcome probability on an absolute scale, while the relative risk and odds ratio describe relative contrasts. Because these measures answer different questions and use different reference values for no effect, readers need to identify the scale before interpreting an estimate or interval.

Second, the endpoint is a participant count derived from a self-reported binary outcome. That makes outcome definition and denominator transparency especially important. The registry reports enrollment and outcome-specific counts that do not match, without explaining why. This is a reminder that an effect estimate should be read with its contributing population and missing-data information, not as a free-standing number.

Third, the intervention is delivered through small online groups and combines several behavioral components. General statistical reasoning suggests that shared group experiences may make participant outcomes correlated, while bundled interventions make component effects difficult to distinguish. The registry does not say how group structure was addressed analytically, so these are questions for interpretation rather than claims about the actual analysis.

Fourth, the engagement measure illustrates how a secondary endpoint can document exposure to an intervention while remaining distinct from a clinical outcome. Tweet volume may describe participation, but it does not by itself show that participation caused abstinence. Unadjusted secondary results also require caution when the full testing plan is not reported.

Finally, the planning illustration demonstrates that sample-size needs depend on the effect targeted. Smaller effects require more participants under the stated assumptions. Those figures are not observed power and do not reveal the original design rationale, which the registry record does not report. Taken together, the record offers an opportunity to distinguish observed estimates, uncertainty, design assumptions, and unreported methodological details rather than treating them as one result.

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 are specific to interpreting this registry record. The primary endpoint is self-reported survey abstinence, and the record does not describe biochemical confirmation, survey wording, response rules, or assessor procedures. The endpoint is defined as seven-day point-prevalence abstinence sustained at 60 days after the quit date; it does not directly describe continuous abstinence over the entire period or outcomes at later follow-up.

The number enrolled is 160, but the primary outcome arm counts are 65 and 70. The registry record does not explain the discrepancy, report reasons for unavailable outcomes, or describe how missing data were handled. It also does not state the primary analysis population. These gaps limit evaluation of follow-up completeness and possible selection effects. The secondary engagement and safety records use 80 participants per arm, and the record does not explain how these populations relate to the primary endpoint population.

The trial is listed as unmasked. The record does not give details about outcome-assessment masking or whether participants knew the study’s specific hypotheses. The intervention itself includes several features, so the randomized contrast does not isolate the effect of peer messaging, daily reminders, or the automated topic bot. The 20-person group structure also raises a general question about within-group dependence; the registry does not report whether or how that structure was accounted for in the analysis.

The record lists zero reported analyses and provides no detailed statistical analysis plan, multiplicity strategy, interim-analysis plan, original sample-size justification, or named inferential method. The posted p-values are unadjusted. There is no reported formal analysis connecting tweet volume with abstinence, and the registry does not describe whether other secondary analyses were conducted. These omissions limit assessment of prespecification and the full scope of statistical testing.

Finally, the record’s brief population description is people seeking to quit smoking, but further eligibility and representativeness details are not reported. Generalizability to other groups, settings, or delivery formats cannot be determined from these summaries alone.

19. Related Tutorials

20. Related Statistical Calculators

21. Sources

• ClinicalTrials.gov: Twitter-enabled Mobile Messaging for Smoking Relapse Prevention — NCT01602536.
• Registry record last updated 2024-10-31; status last verified 2024-10.
• Results first posted 2024-10-09.
• 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, parallel, two-arm trial of Twitter-enabled mobile messaging for smoking relapse prevention. The experimental arm received smoking cessation aides plus assignment to a private Twitter quit-smoking group, interactive peer messaging, daily encouragement, and automated daily discussion topics. Control participants received the same baseline aides without a Twitter group. Enrollment is recorded as 160, and masking is listed as none.

The primary outcome was self-reported seven-day point-prevalence abstinence sustained at 60 days after the quit date. The posted comparison reports a risk difference of 0.2, a 95% confidence interval from 0.049 to 0.351, and an unadjusted p-value of 0.011. A secondary outcome reports tweet-volume engagement over 100 days. The record lists zero serious adverse events, other adverse events, and deaths in each arm among 80 participants at risk per arm.

The registry does not report the original sample-size justification, detailed statistical analysis plan, missing-data handling, analysis population, multiplicity procedure, or reasons for the differences among outcome-specific counts. Those omissions constrain interpretation of the posted results.