Smoking Cessation Interventions for People Living With HIV in Nairobi, Kenya: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Smoking Cessation Interventions for People Living With HIV in Nairobi, Kenya (NCT03342027), a randomized phase 3 study. The primary outcome is 7-day Abstinence.

2. Clinical Question

The central statistical question concerns a binary outcome: whether each participant met the abstinence definition at the relevant assessment. The posted results therefore describe participants meeting the endpoint, rather than the number of cigarettes smoked or the duration of abstinence. The randomized comparisons can describe outcome patterns among the assigned combinations. They do not, on their own, establish why those patterns occurred or whether they reflect a medication effect, a behavioral-program effect, or a combination of both.

The four-arm arrangement matters to interpretation. A direct comparison of one combination with the reference arm changes more than one treatment component in some cases, so it should not be read as an isolated test of a single component. The record does not post an analysis that separates these components. Accordingly, the comparison-specific estimates are best read as results for the named combinations, against the stated reference, rather than as a complete account of the factorial research question.

3. Trial Design

The registry describes a Phase 3, randomized, double-masked, factorial trial. The listed enrollment is 300, and the status is COMPLETED. The record lists a start date of 2020-08-01 and a primary completion date of 2023-07-25. The sponsor is the University of Maryland, Baltimore. The study concerns smoking cessation among people living with HIV in Nairobi, Kenya.

Participants were assigned to one of four combinations: Bupropion + Positively Smoke Free, Bupropion + Standard of Care, Placebo + Positively Smoke Free, or Placebo + Standard of Care. Positively Smoke Free is described as an 8-session tailored behavioral treatment. Standard of Care is described as brief advice to quit, delivered in a standardized format. The placebo is described as matched to bupropion. The registry calls the trial factorial, but does not report the planned factorial model, the specific main-effect contrasts, or an interaction analysis.

For the primary outcome, the recorded participant counts are 72.0 for Bupropion + Positively Smoke Free, 72.0 for Bupropion + Standard of Care, 74.0 for Placebo + Positively Smoke Free, and 76.0 for Placebo + Standard of Care. These are the arm denominators attached to the posted outcome counts. The registry does not explain how they relate to the listed enrollment of 300, so the difference should not be interpreted as evidence of a particular type of missingness, exclusion, or follow-up pattern.

The registry record does not report the randomization procedure, allocation concealment, masking procedures, adherence measures, or the planned handling of protocol deviations. It also does not report the original sample-size justification or detailed design assumptions. Those omissions limit assessment of how the trial was planned and conducted beyond the design features stated in the record.

4. Participant Flow and Accounting

The registry reports how participants moved through each stage of the study. All counts below are exactly as posted; missing counts are not inferred from other stages.

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 7-day Abstinence, measured over Up to 36 Weeks.

The primary endpoint is 7-day Abstinence, measured as a count of participants. It is defined as self-reported no smoking in the past 7 days plus CO <7 ppm, with a timeframe of up to 36 Weeks. The endpoint therefore combines a participant’s report about recent smoking with a carbon monoxide measurement below the stated threshold. The posted outcome is the number of participants meeting that definition in each arm.

As a general statistical concept, a binary endpoint makes each participant’s status classifiable into one of two categories for the defined assessment: meeting or not meeting the endpoint. Reporting the count makes the observed number of participants meeting the definition visible, while the arm denominator provides the context for that count. The endpoint’s combined self-report and measurement criteria are part of the registry definition; the record does not provide further details about measurement timing, verification procedures, or how unavailable assessments were classified.

A short-window abstinence definition captures recent smoking status at assessment, rather than establishing continuous abstinence over the whole study period. The phrase “up to 36 Weeks” sets the reported timeframe, but the registry record does not specify the schedule of assessments or which assessment determined the posted result. The endpoint should therefore be interpreted according to the definition and timeframe recorded, without assuming a particular follow-up schedule or a longer-term cessation measure.

7. Results Posted in the Registry

• Bupropion + Positively Smoke Free: 5.6% (4/72) versus 23.6% (17/72) on control; risk difference -18.1 percentage points (95% CI -29.2 to -6.9), risk ratio 0.24 (0.08–0.67); unadjusted two-proportion z-test p = 0.002.
• Placebo + Positively Smoke Free: 20.3% (15/74) versus 23.6% (17/72) on control; risk difference -3.3 percentage points (95% CI -16.8 to 10.1), risk ratio 0.86 (0.46–1.59); unadjusted two-proportion z-test p = 0.626.
• Placebo + Standard of Care.: 38.2% (29/76) versus 23.6% (17/72) on control; risk difference 14.5 percentage points (95% CI -0.1 to 29.2), risk ratio 1.62 (0.98–2.68); unadjusted two-proportion z-test p = 0.056.

8. Statistical Interpretation

9. Statistical Methods Explained

What does a risk difference express?

A risk difference compares the proportion meeting a binary endpoint in one arm with the proportion meeting it in a reference arm. In these posted comparisons, a negative estimate means the named arm had a lower observed endpoint proportion than Bupropion + Standard of Care; a positive estimate means it had a higher observed endpoint proportion. The estimate is an absolute contrast on the outcome scale, not a relative measure. It refers to the specific pair of arms being compared and does not automatically isolate one component of a factorial intervention.

The comparison summaries also contain relative measures, but the registry record does not describe a planned analysis for them. The central interpretation in the posted results can be made using the counts, risk differences, confidence intervals, and unadjusted p-values. No additional adjusted effect estimate is reported in the registry analysis records.

How should the confidence intervals be read?

A confidence interval describes statistical uncertainty around an estimate under the method used to calculate it. As a general explanation, a wider interval indicates less precision, while a narrower interval indicates greater precision. For a risk difference, an interval that spans zero includes values corresponding to no difference between the compared arms, as well as values in either direction. An interval wholly on one side of zero indicates that the values in that interval point in one direction, although it does not establish clinical importance or remove concerns about the analysis plan.

For the bupropion-plus-behavioral-program comparison, the posted interval is below zero. The intervals for the placebo-plus-behavioral-program comparison and the placebo-plus-standard-advice comparison span zero. These statements describe the reported intervals; they do not substitute for the trial’s original statistical analysis plan, which is not reported in the registry record.

What does an unadjusted p-value tell the reader?

A p-value is a measure of how compatible the observed data are with a specified null model, under the assumptions of the calculation. It is not the probability that the null hypothesis is true, nor is it a measure of the size or clinical importance of a difference. The posted p-values are identified as unadjusted. The registry does not state whether they came from the sponsor’s planned analysis, identify the test procedure, or report adjustment for the multiple comparisons.

The p-values should be considered alongside the arm counts and confidence intervals. A small p-value does not explain the source of a difference, and a larger p-value does not demonstrate that two approaches are equivalent. The record does not report an equivalence or non-inferiority framework.

Why does the factorial design matter?

In general, a factorial design assigns combinations of intervention components so that investigators can study component effects and potentially whether components work differently in combination. Here, the four combinations pair bupropion or placebo with Positively Smoke Free or Standard of Care. To interpret the design as a test of separate component effects, the analysis would need to specify how results are combined across arms and whether an interaction is assessed. The registry record does not report those model details or an interaction result.

The posted comparisons instead contrast each of three named arms with Bupropion + Standard of Care. Each contrast has a clear arm-level meaning, but they are not interchangeable with an estimated overall medication effect or an overall behavioral-program effect. A comparison between two arms differing in both components cannot separate their contributions. The outcome counts alone do not resolve that issue.

What can be said about the analysis population?

The registry reports arm denominators for the primary endpoint, but does not report whether the results use an intention-to-treat population, a per-protocol population, or another analysis population. It does not describe exclusions after randomization, treatment adherence, or how participants with no endpoint assessment were handled. The denominators should be read as the numbers attached to the reported arm outcomes, not as proof of a particular analysis rule.

Likewise, the record does not report a missing-data method, imputation procedure, sensitivity analysis, or interim analysis. These are important features for evaluating an outcome analysis, especially when an endpoint depends on follow-up assessment. Their absence from the record means that the posted counts cannot be used to infer how unobserved outcomes were classified.

What is the scope of these calculations?

The posted comparison estimates are unadjusted calculations based on the reported summary statistics. The registry analysis-record count is 0, so a sponsor analysis with model specifications and adjustment details is not reported in the registry record. No discrepancy is flagged between the available summary information and these calculations. The values should consequently be interpreted as simple comparisons of the posted arm-level results, not as evidence that a more detailed planned analysis was conducted.

10. Confidence Intervals and Precision

Precision is reflected in the confidence interval around each risk-difference estimate. The interval for Bupropion + Positively Smoke Free compared with Bupropion + Standard of Care lies below zero. The intervals for Placebo + Positively Smoke Free and Placebo + Standard of Care compared with the reference arm span zero. The latter comparisons therefore leave uncertainty about whether the underlying difference is in one direction or the other.

The observed counts provide important context for that uncertainty. The first comparison has 4.0 participants meeting the definition in one arm and 17.0 in the reference arm, with 72.0 participants in each. The other comparisons also have outcome counts based on arm denominators in the low-to-mid 70s. Those denominators are the numbers attached to the posted results, not a guarantee that every randomized participant had an observed endpoint.

As a general statistical point, an interval’s width reflects both the estimated contrast and the information in the data, as well as the method used to construct the interval. A confidence interval is not a guarantee that the true value lies within its endpoints. It summarizes uncertainty under the calculation’s assumptions. The record does not identify the confidence-interval method or describe adjustments for multiple comparisons.

The precision of a comparison should not be confused with the importance of the outcome or the quality of trial conduct. The record does not report an analysis population, missing-data approach, or detailed model specification. Those omissions limit how fully the reported intervals can be assessed. The intervals and unadjusted p-values are best read as uncertainty summaries for the particular arm comparisons, rather than as a complete statistical analysis of the factorial trial.

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 for these binary-outcome comparisons, not total events. They are illustrations based on the observed effect, with a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. The scenarios consider an effect equal to 100%, 75%, or 50% of the observed effect. Smaller effects require more participants to detect under the same planning framework.

The 80% power column describes the planned number of participants per arm for a calculation with an 80% chance of detecting the specified effect if that effect were the true one, under the stated assumptions. The 90% power column uses a higher target chance of detection and therefore generally calls for a larger per-arm sample. For example, for the Bupropion + Positively Smoke Free comparison, the 100% observed-effect scenario lists 59 participants per arm for 80% power and 79 for 90% power. At 75% of the observed effect, the corresponding figures are 119 and 159 participants per arm; at 50%, they are 297 and 397.

These are planning illustrations, not observed power. The trial’s original design assumptions and sample-size justification are not reported in the registry record. The planning figures do not establish that the trial was designed around the observed effect, nor do they show the probability that the reported findings are true. For comparisons with a smaller observed effect, the planning counts are much larger; that pattern follows from the calculation’s stated assumption that a smaller difference is harder to distinguish from no difference.

The equal-allocation assumption applies to the planning illustration, while the posted outcome denominators differ across some arms. The planning counts should not be interpreted as the trial’s actual enrollment or as evidence of the sponsor’s intended allocation scheme. No time-to-event endpoint is reported among the posted primary results; the figures relevant to these comparisons are participant counts per arm.

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

Worked calculation for this trial

1. Control proportion p₀ = 0.236; treatment proportion p₁ = 0.056; average p̄ = 0.146.
2. First term: z_1−α/2 × √[2p̄(1−p̄)] = 1.96 × 0.4991 = 0.9783.
3. Second term: z_1−β × √[p₀(1−p₀) + p₁(1−p₁)] = 0.842 × 0.4825 = 0.4061.
4. n = (0.9783 + 0.4061)² / (-0.1806)² = 58.8, rounded up to 59 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 <- 0.236111111111111                          # control proportion
p1 <- 0.0555555555555556                          # 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(3342027)
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 = 59 (80% power), 79 (90% power)
 75% of observed effect: n per arm = 119 (80% power), 159 (90% power)
 50% of observed effect: n per arm = 297 (80% power), 397 (90% power)
power.prop.test: 58.8 
Simulated power at n = 59 per arm: 0.836 (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.

13. Multiplicity, Missing Data and Analysis Populations

The registry record does not report a multiplicity plan. The posted results contain three comparisons with a shared reference arm, and each is labeled unadjusted. When several comparisons are examined, the chance of at least one apparently unusual result can differ from the chance associated with a single pre-specified test. As a general statistical principle, a multiplicity procedure may be used to control that issue, but no such procedure is reported. The displayed p-values should therefore be understood as unadjusted values, not as values corrected for the set of comparisons.

The record also does not report whether the comparisons were pre-specified, whether the primary statistical test was one of these arm-to-reference contrasts, or whether any other outcomes were examined but not posted. The registry lists one outcome measure and zero analysis records. It lists no posted secondary outcomes. These entries do not provide a full account of the analysis plan or of any analyses performed outside the registry record.

Missing-data handling is not reported in the registry record. The endpoint depends on both self-report and a carbon monoxide criterion, but the record does not explain how missing self-reports, unavailable measurements, missed visits, or incomplete follow-up were handled. It does not state whether participants without an endpoint assessment were counted as not abstinent, excluded, or addressed through another method. The outcome denominators alone cannot answer those questions.

The analysis population is also not reported. The record does not say whether the estimates follow intention-to-treat, per-protocol, or another rule, and it does not describe post-randomization exclusions or protocol deviations. Interim analyses, stopping rules, sensitivity analyses, and model covariates are not reported in the registry record. Without those details, the posted results can be described as arm-level summaries and unadjusted comparisons, but the robustness of a sponsor-planned analysis cannot be evaluated from the registry record alone.

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.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: hospitalization (2.7%).

The adverse-event results list serious events among participants at risk in each randomized arm. The serious-event counts are 1 among 72 in Bupropion + Positively Smoke Free, 1 among 72 in Bupropion + Standard of Care, 4 among 74 in Placebo + Positively Smoke Free, and 2 among 76 in Placebo + Standard of Care. The record also lists deaths as 3, 0, 3, and 0 in those respective arms.

Hospitalization is the leading serious-event term listed, at 2.7%. The registry does not provide a corresponding term-specific count in the record summarized by these figures, so that percentage should not be converted into an inferred number of participants. The adverse-event section also reports other, non-serious events, but those counts are distinct from the serious-event counts.

These figures are descriptive. The registry record does not give event-level timing, relatedness assessments, severity detail beyond the serious-event category, or an analysis comparing safety between groups. The counts do not establish that an event was caused by a particular treatment, and differences in raw counts should not be interpreted as comparative risk estimates without an analysis that accounts for the relevant denominators and follow-up.

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

The posted results establish that the four arms had different observed counts of participants meeting the registry’s 7-day abstinence definition. The largest count is in Placebo + Standard of Care, with 29.0 participants among 76.0; the reference Bupropion + Standard of Care arm has 17.0 among 72.0. Bupropion + Positively Smoke Free has 4.0 among 72.0, and Placebo + Positively Smoke Free has 15.0 among 74.0. These are descriptions of the recorded outcome, not explanations of the differences.

The comparison estimates give more information than counts alone, but they remain limited to the stated contrasts. The unadjusted interval for Bupropion + Positively Smoke Free lies below zero relative to the reference. The other two unadjusted intervals include zero. These patterns describe the direction and uncertainty of the posted comparisons; they do not determine whether an observed difference is important to participants or whether it would be replicated.

The results do not isolate the effects of bupropion and Positively Smoke Free, despite the factorial design. Doing so would require a reported analysis that appropriately combines the arms and addresses the factorial structure. No such analysis or interaction estimate is reported in the registry record. Nor do the counts establish a sustained cessation outcome beyond the specified 7-day definition at an assessment within the timeframe up to 36 Weeks.

The registry record does not report enough detail to evaluate missing outcomes, the analysis population, treatment adherence, multiplicity, or the trial’s original statistical plan. Those gaps constrain interpretation of the precision and robustness of the posted comparisons. The results should be read as a set of trial-specific arm counts and unadjusted estimates, not as medical advice, a definitive causal explanation, or a general conclusion about smoking cessation treatments.

16. Why This Trial Matters Statistically

This trial is statistically notable for pairing a factorial design with a binary clinical endpoint and posting four arm-level combinations. That structure raises an important distinction between comparing named treatment combinations and estimating separate component effects. The registry reports the former through comparisons with a single reference arm, but does not report the factorial model needed to describe medication and behavioral effects separately or assess whether those components interact.

The outcome definition also illustrates how a clinical endpoint can combine different forms of measurement. Self-reported no smoking in the past 7 days is paired with a carbon monoxide criterion of CO <7 ppm. The resulting endpoint is a participant count, which is readily interpretable, but the record does not report the assessment schedule or how missing components of the definition were handled. Those details are important for understanding what the count represents.

The posted confidence intervals reinforce why point estimates and p-values should not be read alone. One comparison has an interval below zero, while two span zero. The latter intervals leave uncertainty about the direction of the underlying difference. The record does not report multiplicity adjustment, a formal analysis plan, or sponsor analysis records, so the comparisons are not a complete inferential account of the trial.

Finally, the record shows how registry summaries can inform statistical reading while leaving key design and analysis questions unanswered. The trial status, design label, interventions, endpoint, arm counts, and adverse-event summaries are available, but details such as missing-data handling and analysis populations are not reported. Keeping those boundaries clear helps readers distinguish observed results from unreported methods and from conclusions that the posted data cannot establish.

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 registry record gives limited information about eligibility, recruitment, and participant characteristics beyond identifying people living with HIV who smoke and the Nairobi clinic setting. It does not report demographic or clinical characteristics, so the degree to which the enrolled participants represent the intended population cannot be assessed from these results alone.

Although the record identifies the study as randomized and double-masked, it does not describe the randomization method, allocation concealment, or how masking was maintained. It does not report adherence to bupropion or placebo, attendance at the behavioral sessions, or receipt of brief advice. Those omissions limit interpretation of how the assigned interventions were implemented.

The primary endpoint is specified, but the assessment schedule within the “Up to 36 Weeks” timeframe is not reported. The record does not say how abstinence was determined when either self-report or carbon monoxide measurement was unavailable, or how participants lost to follow-up were handled. It also does not describe whether the results use all randomized participants or another analysis population.

The posted outcome denominators differ across arms, while the listed enrollment is 300. The registry does not explain this relationship. It would be inappropriate to infer from those figures alone whether the difference reflects missing observations, eligibility decisions, or another reason. The record also does not report the original sample-size justification, the protocol’s intended contrasts, an interaction analysis, multiplicity control, interim monitoring, or sensitivity analyses.

Finally, the comparisons are unadjusted and the registry lists zero analysis records. The available estimates therefore do not replace a full trial report with prespecified methods and analysis details. These limitations do not invalidate the arm counts, but they restrict what can be concluded from them.

18. Related Tutorials

19. Related Statistical Calculators

20. Sources

• ClinicalTrials.gov: Smoking Cessation Interventions for People Living With HIV in Nairobi, Kenya — NCT03342027.
• Registry record last updated 2025-02-05; status last verified 2025-02.
• Results first posted 2025-02-05.
• 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

NCT03342027 is a completed, randomized, double-masked, Phase 3 factorial trial sponsored by the University of Maryland, Baltimore. It evaluated four combinations of bupropion or matched placebo with Positively Smoke Free or standardized brief advice among people living with HIV who smoke in Nairobi, Kenya. The listed enrollment is 300.

The primary outcome is the count of participants meeting the 7-day Abstinence definition—self-reported no smoking in the past 7 days plus CO <7 ppm—up to 36 Weeks. The registry posts arm counts and three unadjusted comparisons against Bupropion + Standard of Care. One comparison’s confidence interval lies below zero; two span zero. The registry record does not report a factorial analysis, missing-data approach, analysis population, multiplicity procedure, or original sample-size justification.