Worksite Phone Counseling for Smoking Cessation: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Worksite Phone Counseling for Smoking Cessation (NCT02730260), a randomized na study. The primary outcome is Smoking Abstinence for 7 Days at Last Contact.

2. Clinical Question

3. Trial Design

This completed trial is registered as NCT02730260, with the title “Worksite Phone Counseling for Smoking Cessation.” It is described as randomized, with a parallel-group model and no masking. The record lists 518 participants enrolled. The sponsor is Washington University School of Medicine. The recorded start was 2005-02, and primary completion was 2008-09.

The randomized groups received different forms of telephone coaching through a quitline: directive coaching or nondirective coaching. Each arm description allows for up to 7 calls over 9 weeks. The registry summary frames the comparison around the possible role of social support in behavior change and notes that interactions involving race or income have been reported in some contexts. It does not report trial results for those interactions in the primary outcome.

The primary outcome is a binary self-reported abstinence measure assessed at the participant’s last post-intervention contact, occurring 6 to 12 months after enrollment. The outcome definition counts people reporting no cigarettes in the preceding 7 days. The population description says that participants successfully contacted at 6 or 12 months are included and that other participants are counted as continuing smokers. The arm counts posted for this measure are 68 of 176 for Directive and 77 of 185 for Nondirective.

The registry record does not report the original sample-size justification, the randomization procedure or concealment details, the intended analysis population, or the complete statistical analysis plan. The posted analyses include a chi-squared result and two logistic-regression results, but their model specifications are not described. The record also does not report secondary outcomes.

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.

Before assignment: 978 persons inquired about the program. 847 took steps to qualify for an employer health insurance discount Of these, 294 declined to enroll, e.g. for being in the precontemplative stage of change. 553 persons participated in coaching. Of these, 35 worked with coaches in training. 518 participants enrolled in the randomized trial.

Overall Study

Analysis population for the primary outcome, as defined in the registry: These are participants who were successfully contacted at 6 or 12 months following enrollment. Other participants are counted as continuing smokers.

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 Smoking Abstinence for 7 Days at Last Contact, measured over 6-12 months.

The primary endpoint is “Smoking Abstinence for 7 Days at Last Contact.” It counts participants who self-reported smoking no cigarettes during the 7 days before their last post-intervention assessment. That assessment took place 6 to 12 months after enrollment. The outcome is binary at the participant level: each person is classified according to whether the abstinence definition was met.

A short, defined reporting window makes the outcome interpretable as a measure of recent smoking status at follow-up. It is not a measure of continuous abstinence throughout the period since enrollment. Because the measure is self-reported, it describes participants’ reported smoking rather than an independently verified status; the registry record does not report biochemical verification.

The endpoint’s population definition is important for interpreting the counts. Participants successfully contacted at 6 or 12 months contribute according to their reported status, while other participants are counted as continuing smokers. The posted counts are 68 in the Directive group and 77 in the Nondirective group, with arm counts of 176 and 185, respectively. These counts describe the endpoint as registered; they should not be confused with the total enrollment of 518.

Using a binary outcome supports comparisons of the proportion meeting a prespecified definition, but the record does not explain why this particular follow-up window or classification rule was selected. The endpoint does not, by itself, describe how much participants smoked outside the 7-day window or whether any reported abstinence continued after assessment.

7. Results Posted in the Registry

• Nondirective: 41.6% (77/185) versus 38.6% (68/176) on control; risk difference 3.0 percentage points (95% CI -7.1 to 13.1), risk ratio 1.08 (0.84–1.39); unadjusted two-proportion z-test p = 0.563.

8. Statistical Interpretation

9. Statistical Methods Explained

The primary outcome is a count of participants who met a binary definition: self-report of no cigarettes during the 7 days before the last post-intervention assessment. For this type of outcome, a group’s count can be compared with its group size to describe the fraction meeting the definition. The posted counts are 68 of 176 for Directive and 77 of 185 for Nondirective. The record does not provide a separate adjusted estimate for the primary outcome with enough detail to identify its model specification.

What does the unadjusted risk difference describe?

The risk difference is an absolute contrast between the outcome proportions in the two groups. The recalculated contrast is defined as Nondirective minus Directive and is 0.030. A value of zero corresponds to equal proportions under this contrast definition; a positive value indicates a larger proportion in Nondirective, while a negative value indicates a larger proportion in Directive. The confidence interval extends from -0.071 to 0.131, so it includes zero and values in either direction. This is a simple comparison based on the posted summary counts, not a reproduction of the sponsor’s logistic-regression models.

How does the risk ratio differ?

A risk ratio compares the outcome proportions by division rather than subtraction. The recalculated risk ratio is 1.077, with a confidence interval from 0.837 to 1.387. A ratio of 1 represents equal proportions. Values greater than 1 indicate a greater proportion in the numerator group under the stated comparison order; values less than 1 indicate a smaller proportion. The interval includes 1, so this recalculated ratio is compatible with no ratio difference as well as contrasts in either direction. The ratio’s interpretation is not the same as that of a risk difference: it is relative rather than absolute.

What does an odds ratio measure?

Odds are formed by comparing the number meeting the outcome definition with the number not meeting it. An odds ratio compares those odds between groups. The unadjusted recalculated odds ratio is 1.132, with a confidence interval from 0.743 to 1.726. A ratio of 1 corresponds to equal odds. Odds ratios are not risk ratios, and they are not direct percentage-point differences. The registry’s sponsor-reported logistic odds ratios are 1.8 and 1.6, both reported with p < 0.03. The record does not report enough model detail to explain how those estimates relate to the unadjusted odds ratio.

How should p-values be read?

A p-value summarizes how incompatible the observed data are with a specified null model, assuming that model and its assumptions. It is not the probability that the null hypothesis is true, and it does not quantify the probability that a result is due to chance. The unadjusted comparison has p = 0.563. The registry reports p = 0.41 for a chi-squared analysis and p < 0.03 for each of two logistic-regression analyses. These values come from distinct reported analyses; the record does not describe the models sufficiently to reconcile them. Since a discrepancy is flagged, the sponsor’s own analysis takes precedence.

In general, chi-squared methods compare categorical outcome distributions across groups. Logistic regression models the probability of a binary outcome through odds and can incorporate covariates, but the registry record does not say which covariates, if any, were used in these analyses. The label “Regression, Logistic” alone does not establish whether an estimate is adjusted, which participants were included, or which category was modeled as the event. Those specifications are not reported in the registry record.

General statistical interpretation also depends on the sampling and outcome process. Random allocation can support a between-group comparison, but the record does not give details of how allocation was generated or concealed. Self-reported status and the rule counting participants not successfully contacted as continuing smokers are features of the endpoint definition. They affect what the posted comparison represents and do not establish the value of an alternative analysis.

For counts, over-dispersion is generally a concern when observed variability exceeds what a model assumes; it is especially relevant to count models. This primary endpoint is binary rather than a recurring event count, and no over-dispersion assessment is reported. No additional statistical assumptions or diagnostics should be inferred from the reported method labels.

10. Confidence Intervals and Precision

A confidence interval gives a range of values compatible with the data and the calculation’s assumptions. It communicates uncertainty about an estimated contrast; it does not guarantee that the true value lies inside the interval. The confidence level for the recalculated intervals is not stated in the registry record, so it should not be inferred.

For the unadjusted risk difference, the estimate is 0.030 and the interval runs from -0.071 to 0.131. Because the interval includes zero, the recalculation does not exclude equal outcome proportions. Its endpoints also allow contrasts in either direction. The interval’s width indicates that the estimate is not highly precise: it is compatible with a range extending from a negative to a positive risk difference.

The recalculated risk ratio is 1.077, with an interval from 0.837 to 1.387; the recalculated odds ratio is 1.132, with an interval from 0.743 to 1.726. Both intervals include 1, the value corresponding to no ratio contrast. The ratios and risk difference use distinct scales, so their interval widths should be interpreted in relation to their own measures rather than compared directly.

The sponsor-reported logistic odds ratios of 1.8 and 1.6 have no confidence intervals posted. Their precision therefore cannot be evaluated from interval estimates in the registry record. The reported p-values do not supply the missing interval endpoints or identify the range of effect sizes compatible with those models. Given the discrepancy between sponsor-reported analyses and recalculation, the recalculated intervals describe only the simple unadjusted comparison and do not resolve the sponsor results.

11. Planning a Similar Trial: Sample Size

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

The planning figures are participants per arm, not total events. They are illustrations based on the observed effect, not the trial’s original design calculation. The stated assumptions are a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation; the effect to detect is set to the stated fraction of the observed effect. The trial’s own original design assumptions are not reported in the registry record.

For planning at 100% of the observed effect, the figures are 4,231 participants per arm for 80% power and 5,664 participants per arm for 90% power. At 75% of the observed effect, they are 7,499 participants per arm for 80% power and 10,039 participants per arm for 90% power. At 50% of the observed effect, they are 16,820 participants per arm for 80% power and 22,516 participants per arm for 90% power.

In general, power is the chance that a study will detect an effect of a specified size under the planning assumptions if that effect is present. The 80% column represents planning for an 80% chance of detection under those assumptions; the 90% column represents planning for a 90% chance. Smaller effects require more participants because they are harder to distinguish from random variation. The planning curve expresses this same relationship between assumed effect size and required sample size; its values should be read as planning illustrations under the stated assumptions, not as observed trial outcomes.

These figures are not observed power. They do not describe the probability that the completed trial’s findings are correct or the probability that a treatment effect exists.

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 = 0.386; treatment proportion p₁ = 0.416; average p̄ = 0.401.
2. First term: z_1−α/2 × √[2p̄(1−p̄)] = 1.96 × 0.6932 = 1.3586.
3. Second term: z_1−β × √[p₀(1−p₀) + p₁(1−p₁)] = 0.842 × 0.6929 = 0.5831.
4. n = (1.3586 + 0.5831)² / (0.0299)² = 4,230.9, rounded up to 4,231 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 <- 0.386363636363636                          # control proportion
p1 <- 0.416216216216216                          # 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(2730260)
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 = 4,231 (80% power), 5,664 (90% power)
 75% of observed effect: n per arm = 7,499 (80% power), 10,039 (90% power)
 50% of observed effect: n per arm = 16,820 (80% power), 22,516 (90% power)
power.prop.test: 4230.9 
Simulated power at n = 4,231 per arm: 0.796 (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. Analyses Reported by the Sponsor

The registry reports three sponsor analyses for the comparison. The chi-squared analysis has p = 0.41. Two logistic-regression analyses report odds ratios of 1.8 and 1.6, each with p < 0.03. The registry does not report confidence intervals for those odds ratios or explain their model specifications, covariates, outcome coding, or the difference between the two regression analyses.

The simple recalculation from the posted arm counts gives a risk difference of 0.030, with a confidence interval from -0.071 to 0.131 and p = 0.563. It also gives a risk ratio of 1.077 and an odds ratio of 1.132. These are unadjusted calculations from summary counts, not the sponsor’s regression analysis. The results differ from the sponsor’s posted analyses, including in the reported p-values.

The record flags this discrepancy. The sponsor’s own analysis therefore takes precedence; the recalculation is useful as a transparent description of what follows from the posted counts, but it should not be substituted for the sponsor’s analysis. The available record does not establish the reason for the difference. It would be inappropriate to attribute it to adjustment, participant selection, coding, or another analytic choice without model details.

14. Multiplicity, Missing Data and Analysis Populations

The registry posts one primary outcome measure and reports three analyses comparing the two groups: a chi-squared analysis and two logistic-regression analyses. It reports no secondary outcome measures. The record does not explain whether these analyses were planned in advance, whether they address distinct questions, or whether one was designated as the primary statistical test.

Multiplicity arises when several statistical tests or outcomes are considered. In general, repeated testing can increase the chance of at least one result appearing statistically notable even when no corresponding effect exists. Methods such as prespecifying a primary analysis or adjusting for multiple comparisons can address that issue, but the registry record does not report a multiplicity strategy. It is not possible to determine from the listed analyses whether any adjustment was used or was considered necessary.

The record also does not report a full statistical analysis plan, a protocol-level definition of analysis populations, or details of how participants were handled in the sponsor’s analyses. The primary endpoint description does state that participants successfully contacted at 6 or 12 months are assessed and that other participants are counted as continuing smokers. That is the stated outcome rule, but the registry does not report sensitivity analyses using alternative assumptions for participants without successful follow-up contact.

Further details that are not reported include the handling of missing covariates, the inclusion or exclusion rules for each analysis, and whether any subgroup or interim analyses were conducted. No such analyses or methods should be assumed from the existence of multiple posted results. The reported logistic-regression model specifications and the reason for two separate odds-ratio results are also not reported in the registry record.

These omissions limit how fully the analyses can be reproduced and compared. In particular, the difference between the sponsor’s results and the recalculation cannot be resolved without details about the underlying models and analysis populations. The discrepancy should remain explicit rather than being explained by an unreported assumption.

15. Safety Results

The registry reports serious adverse events as affected participants divided by participants at risk.

Other (non-serious) adverse events

Participants with at least one non-serious event at or above the registry's 0% reporting threshold.

The registry lists 258 participants at risk for serious adverse events in the Directive group and 260 in the Nondirective group. It reports 0 serious adverse events for each group. The listed counts for other adverse events are also 0 in each group, with 258 and 260 participants at risk, respectively.

These figures describe what is recorded for the specified safety categories and denominators. They should not be read as proof that no adverse experiences occurred outside the registry’s recorded categories or reporting process. The record does not provide a list of serious adverse-event terms; none are listed. Death counts are not reported in the registry record.

The safety denominators are distinct from the posted primary-outcome arm counts of 176 and 185. The record does not explain the difference between these sets of counts, so they should not be treated as interchangeable. The registry’s zero counts are descriptive and do not establish a comparative safety estimate or quantify uncertainty about unrecorded events.

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

The result concerns a specific comparison: two telephone-coaching approaches offered through a smoking quitline to employees of two large corporations. The primary outcome is a self-reported 7-day abstinence status at the last post-intervention assessment, occurring 6 to 12 months after enrollment. It is not a measure of sustained abstinence throughout follow-up, smoking intensity, or outcomes beyond the recorded endpoint.

The posted counts are 68 of 176 in the Directive group and 77 of 185 in the Nondirective group. The unadjusted risk difference is 0.030, with an interval from -0.071 to 0.131. This interval includes zero and spans both negative and positive contrasts, so the unadjusted estimate is compatible with multiple possible directions and magnitudes. It should not be described as establishing a definitive difference or equivalence.

The sponsor’s reported analyses differ: the chi-squared result has p = 0.41, while two logistic-regression results report odds ratios of 1.8 and 1.6, each with p < 0.03. Since the record flags a discrepancy, the sponsor’s own analysis takes precedence. However, important details needed to interpret or reproduce those analyses, including model specifications and confidence intervals, are not reported. The unadjusted calculations are not a replacement for those results.

The record also does not establish why any difference appears between the analyses. It does not report the original sample-size justification or a detailed plan for missing data, multiplicity, or analysis populations. The primary outcome rule counts participants not successfully contacted as continuing smokers, which is part of the stated endpoint definition and shapes what the counts represent.

These findings are descriptions of the posted record, not medical advice and not a general conclusion about smoking-cessation coaching. They do not establish what would occur in other settings or populations, or what the outcome would be under a different follow-up or missing-data definition.

17. Why This Trial Matters Statistically

This trial is statistically instructive because the registry presents a simple comparison from arm counts alongside sponsor-reported analyses that do not align numerically with that recalculation. The unadjusted risk difference is 0.030, with a confidence interval from -0.071 to 0.131 and p = 0.563. The registry also reports a chi-squared p = 0.41 and logistic-regression odds ratios of 1.8 and 1.6, each with p < 0.03. The record does not give enough model detail to reconcile these estimates.

That contrast illustrates why method labels and p-values are not sufficient for reproducing a result. A comparison of posted counts, a chi-squared test, and logistic regression can answer related but not necessarily identical statistical questions, particularly if the participant sets, outcome coding, or model specifications differ. Those details are not reported, so the reason for the discrepancy remains uncertain. The explicit precedence of the sponsor’s analysis is important when describing the posted results, while the recalculation remains a distinct unadjusted summary.

The endpoint definition also demonstrates how follow-up and missingness rules affect interpretation. The record counts participants not successfully contacted as continuing smokers, rather than leaving their outcome unspecified in the stated endpoint description. This rule is part of the estimand readers encounter in the registry and should be considered when interpreting the count-based comparison.

Finally, the planning illustrations make clear that the sample size required depends on the effect assumed for planning and the desired power. They are not evidence about the trial’s original design choices or a measure of the completed trial’s observed power. Together, these features make transparent reporting of analysis plans, model details, uncertainty intervals, and follow-up handling central to statistical interpretation.

18. Limitations

• The posted summary statistics are the only data these results rest on; patient-level data, full protocols and statistical analysis plans are not available.
• Recalculated results are simple, unadjusted calculations from posted summaries. They do not replace the sponsor's prespecified analysis.

The registry record leaves several trial-specific limitations unresolved. The primary endpoint is self-reported abstinence during the 7 days before the last post-intervention assessment. It does not describe verification of smoking status, nor does it capture abstinence over the full period since enrollment. The record does not report how outcomes might differ under another definition or follow-up window.

Follow-up was specified at 6 or 12 months after enrollment, and participants not successfully contacted are counted as continuing smokers. The record does not give the number successfully contacted by arm, explain the reasons for unsuccessful contact, or report sensitivity analyses using alternative missing-data assumptions. The posted primary outcome arm counts are 176 and 185, while total enrollment is 518; the record does not explain the relationship between those counts beyond the stated endpoint population definition.

The sponsor analyses cannot be fully assessed from the available reporting. The registry gives a chi-squared p-value and two logistic-regression odds ratios with p-values, but not confidence intervals or model specifications. It does not explain the difference between the two regression results. The recalculation from posted counts differs, and the reason cannot be determined from the record.

The record does not report the original sample-size justification, detailed allocation procedures, a complete statistical analysis plan, multiplicity handling, or analysis-population definitions for each reported result. It also does not report secondary outcomes. As a result, the extent of prespecification and the reproducibility of the sponsor analyses cannot be established from the registry record.

The trial involved employees of two large corporations receiving quitline coaching. The record does not provide further details about participant characteristics or the settings needed to assess how well the findings apply beyond that context. No broader generalization should be inferred from the information recorded.

19. Related Tutorials

20. Related Statistical Calculators

21. Sources

• ClinicalTrials.gov: Worksite Phone Counseling for Smoking Cessation — NCT02730260.
• Registry record last updated 2016-12-23; status last verified 2016-10.
• Results first posted 2016-12-23.
• Data retrieved from ClinicalTrials.gov on 2026-11-09.
• Group sizes, means, event counts and sponsor-reported p-values are reproduced as posted. Effect estimates, confidence intervals and p-values not attributed to the sponsor are independently calculated from the posted summary statistics.

22. Record Summary

NCT02730260 is a completed randomized, parallel-group trial of directive versus nondirective smoking-cessation phone coaching for employees of two large corporations. The record lists enrollment of 518, a start of 2005-02, and primary completion in 2008-09. Participants could receive up to 7 quitline calls over 9 weeks.

The primary outcome is self-reported abstinence from cigarettes during the preceding 7 days at the last assessment, 6 to 12 months after enrollment. Posted counts are 68 of 176 for Directive and 77 of 185 for Nondirective. The unadjusted risk difference is 0.030, with a confidence interval from -0.071 to 0.131 and p = 0.563.

The registry also reports a chi-squared p = 0.41 and logistic-regression odds ratios of 1.8 and 1.6, each with p < 0.03. A discrepancy is flagged, so the sponsor’s own analysis takes precedence. The registry does not report enough model detail to explain the difference.