GeoScan and Remote Geo Smoking Study: Neural and Behavioral Correlates of Smokers' Exposure to Retail Environments: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for GeoScan and Remote Geo Smoking Study: Neural and Behavioral Correlates of Smokers' Exposure to Retail Environments (NCT04279483), a randomized na study. The primary outcome is Self-reported Cigarette Craving During the Experimental Manipulation Period.

2. Clinical Question

3. Trial Design

NCT04279483 is listed as a randomized, single-masked, parallel-group study. The registry describes behavioral interventions involving visits to tobacco retailers or non-tobacco retailers, alongside a control condition. Enrollment is listed as 343. The phase is recorded as not applicable. The sponsor is Abramson Cancer Center at Penn Medicine.

The Tobacco Retailer Group was asked to visit a tobacco retailer 5 times per week for 4 weeks, for 20 store visits. The Non-tobacco Retailer Group was asked to visit a non-tobacco retailer on the same schedule. Control participants were not asked to change their behavior during the 4-week intervention period and were assigned 0 store visits. The registry lists these three groups, as well as an unassigned group for safety reporting. Unassigned participants did not begin the intervention period and had no primary outcome data reported.

The study began on 2019-12-17, reached primary completion on 2024-06-10, and is listed as completed. Its stated purpose was to learn about relationships between the brain, behavior, and natural daily exposure to particular environments, including places where smokers regularly spend time and specific retail outlets. The primary outcome concerns self-reported cigarette craving during the experimental manipulation period.

The registry identifies the design as single-masked but does not report who was masked. Further details about randomization procedures, allocation concealment, the original sample-size justification, or the planned statistical analysis are not reported in the registry record. The posted enrollment is not the same as the number contributing to the primary outcome averages: those averages use participants who responded to any craving-rating questions, and the outcome counts differ by group.

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: Participants completed several screening steps to confirm eligibility and interest in the study. If eligible, they completed the first online survey session, and then began the baseline period, during which they answered questions about their cigarette craving and smoking multiple times daily and recorded their geolocation through the Google Maps smartphone application. Participants who completed the baseline period and began the intervention period were assigned to an experimental condition.

Baseline Period

Intervention Period

fMRI Scans

Analysis population for the primary outcome, as defined in the registry: All participants who responded to any craving rating questions during the intervention period are included in this average. Participants in the "unassigned" group did not begin the intervention period and therefore are have no outcome data to report.

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.

Population: Baseline population includes each individual who signed the informed consent form.

6. Primary Endpoint

The registry lists the primary outcome as Self-reported Cigarette Craving During the Experimental Manipulation Period, measured over 4-week experimental manipulation period.

The primary endpoint is self-reported cigarette craving during the 4-week experimental manipulation period. Participants were asked 4 times per day for 28 days, “Right now, how much do you want to smoke a cigarette?” They answered on a 100-point scale, from 0 (not at all) to 100 (extremely).

The registry describes a sequence of averaging: craving ratings were averaged within each day for each participant, then across intervention-period days for each participant, and finally across participants in each group. Thus, the posted group value is a mean of participant-level averages, rather than a count of craving episodes or a single rating at one time point. The posted spread is a standard deviation, which describes variability among the included participant-level values.

Repeated ratings can capture reported craving across everyday experience during the intervention period, rather than relying on one assessment. The measure is also directly interpretable on its stated scale: higher scores indicate stronger reported craving at the time of the rating. This is a self-report measure, not a direct measure of smoking behavior or brain activity. The outcome population includes participants who answered any craving questions during the period; the registry does not describe how incomplete sets of daily ratings were handled beyond the averaging procedure.

7. Results Posted in the Registry

• Tobacco Retailer Group: mean 48.33 versus 47.17 on control; difference 1.16 (95% CI -6.22 to 8.54), standardized effect d = 0.05; z-test p = 0.758.
• Non-tobacco Retailer Group: mean 49.14 versus 47.17 on control; difference 1.97 (95% CI -5.07 to 9.01), standardized effect d = 0.08; z-test p = 0.583.

8. Statistical Interpretation

9. Statistical Methods Explained

The primary outcome is continuous: each participant contributes an average craving score on the 100-point scale. A group mean summarizes the center of those participant-level averages. The standard deviation describes how dispersed the values are among participants. A large spread means that individual values vary considerably around the group mean; it does not, by itself, measure uncertainty in the mean.

What does a mean difference represent?

The unadjusted recalculations compare the mean for each retailer group with the mean for Control. The Tobacco Retailer Group comparison has an estimated mean difference of 1.16; the Non-tobacco Retailer Group comparison has an estimated mean difference of 1.97. Positive estimates indicate that the corresponding retailer-group mean is higher than the Control mean under the direction used for these comparisons. These are simple comparisons of the posted summary statistics, not adjusted estimates from the sponsor’s mixed models.

The standardized effect estimates, d = 0.047 for the Tobacco Retailer Group comparison and d = 0.084 for the Non-tobacco Retailer Group comparison, express the mean differences relative to the pooled standard deviations used in the recalculations. Standardization can help describe the size of a difference in relation to the outcome’s variability, but it does not remove uncertainty or establish practical importance. The registry does not report a threshold for a meaningful change in craving.

What does the confidence interval add?

A confidence interval gives a range of estimates compatible with the data under the calculation’s assumptions. It helps show the uncertainty around a point estimate. The 95% intervals for both recalculated mean differences cross zero: -6.218 to 8.538 for the Tobacco Retailer Group comparison, and -5.07 to 9.01 for the Non-tobacco Retailer Group comparison. The intervals therefore include no difference as well as differences in either direction. A wide interval signals limited precision; it should not be interpreted as proof that all values in the interval are equally plausible, nor as a probability statement about the true value.

How do the posted analyses differ?

The registry reports mixed-model analyses for two group comparisons. Their outcome parameter is a slope: -0.64 for Tobacco Retailer Group versus Non-tobacco Retailer Group and -0.84 for Tobacco Retailer Group versus Control. The associated p-values are 0.74 and 0.69, respectively. The registry does not report confidence intervals for these slopes or the model specification and time coding needed to interpret the slope in detail. A slope analysis can use the repeated measurements over time, whereas the recalculation uses the posted period-level means and standard deviations. The two approaches consequently answer related but distinct questions.

A p-value describes how unusual data at least as incompatible with a specified null hypothesis would be under that hypothesis and the analysis assumptions. It is not the probability that the null hypothesis is true, and it does not quantify the size or importance of a difference. The recalculated unadjusted p-values are 0.758 and 0.583. The sponsor-reported p-values are 0.74 and 0.69. The registry does not report whether the sponsor analyses adjusted for multiple comparisons or other covariates.

Why can repeated measurements matter?

Repeated craving ratings from one participant are related observations, rather than independent participants. In general, mixed models can account for this clustering and can describe patterns over time. The registry identifies the sponsor’s method as mixed models but does not report enough detail to determine the exact covariance structure, fixed effects, missing-data assumptions, or slope contrast. Therefore, the method label alone does not support a more specific account of the sponsor’s model.

The posted primary outcome average follows a participant-level averaging procedure. That approach gives each included participant a period average, but it can conceal variation in daily patterns or the amount of available rating data. The registry states that participants responding to any craving questions are included; handling of partially observed days and the impact of missing ratings are not reported in the registry record.

All statistical methods involve assumptions. General examples include how observations are distributed, how missing measurements relate to observed measurements, and whether model structure reflects the data. The recalculations are explicitly unadjusted comparisons based on posted summary statistics. They should be interpreted as descriptive supplementary analyses, while the sponsor’s own reported analysis remains the registry’s analysis of record.

10. Confidence Intervals and Precision

Precision concerns how much uncertainty surrounds an estimate. The group means are 48.33 for the Tobacco Retailer Group, 49.14 for the Non-tobacco Retailer Group, and 47.17 for Control, with standard deviations of 25.3, 22.96, and 24.24. The standard deviations describe variation among participant-level averages; they are not confidence intervals and do not show how precisely each mean is estimated on their own.

The recalculated mean difference for the Tobacco Retailer Group versus Control is 1.16, with a 95% confidence interval from -6.218 to 8.538. For the Non-tobacco Retailer Group versus Control, the mean difference is 1.97, with a 95% confidence interval from -5.07 to 9.01. Both intervals include zero and extend to negative and positive differences. This means the posted summary data do not pin down the direction of either mean difference with high precision.

Confidence intervals are useful because they show a range of estimates consistent with the data under the calculation’s assumptions, rather than presenting only one point estimate. They do not establish that the true effect lies within the interval with a probability of 95%, and they do not resolve questions about bias, missing outcome data, or the appropriateness of the analysis. The registry does not report confidence intervals for the sponsor’s mixed-model slope estimates, so their precision cannot be assessed from those posted analyses.

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 illustrations give participants per arm, not total events. They show sample sizes for two-arm comparisons under the stated planning assumptions: a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. The effect to detect is set at the stated fraction of the observed effect. These are planning assumptions for the illustrations; the original trial’s design assumptions are not reported in the registry record.

For the Tobacco Retailer Group comparison, planning for 100% of the observed effect gives 7,207 participants per arm for 80% power and 9,648 participants per arm for 90% power. Planning for 75% of the observed effect gives 12,813 participants per arm for 80% power and 17,152 participants per arm for 90% power. Planning for 50% of the observed effect gives 28,828 participants per arm for 80% power and 38,592 participants per arm for 90% power.

For the Non-tobacco Retailer Group comparison, planning for 100% of the observed effect gives 2,235 participants per arm for 80% power and 2,992 participants per arm for 90% power. Planning for 75% of the observed effect gives 3,973 participants per arm for 80% power and 5,318 participants per arm for 90% power. Planning for 50% of the observed effect gives 8,938 participants per arm for 80% power and 11,966 participants per arm for 90% power.

In plain terms, the 80% power column gives the planned number per arm for an 80% chance of detecting the specified effect under the assumptions; the 90% column uses a higher detection probability and therefore lists more participants. Smaller effects need more participants to detect under the same assumptions. These figures are planning illustrations based on the observed effects, not observed power and not the trial’s original sample-size justification.

12. How the Sample Size Is Derived

Power is the probability that a study of a given size rejects the null hypothesis when a specific true effect exists. It depends on four things: the effect to detect, the variability of the outcome, the significance level (here a two-sided α of 0.05, so z_1−α/2 = 1.960) and the desired power (z_1−β = 0.842 for 80% and 1.282 for 90%).

Formula

n is the number of participants in each arm, Δ is the difference in means to detect and σ is the common standard deviation.

Worked calculation for this trial

1. Observed difference Δ = 1.160; pooled standard deviation σ = 24.855.
2. n = 2 × (1.96 + 0.842)² × 24.855² / 1.160² = 7,206.8, rounded up to 7,207 participants per arm for 80% power.

The smaller-effect rows in the table above repeat the same calculation with the effect multiplied by 0.75 and 0.5 (for ratios, applied to the logarithm of the ratio).

R code

alpha <- 0.05
z_a   <- qnorm(1 - alpha / 2)          # two-sided
delta <- 1.16                      # observed difference in means
sigma <- 24.8547206829987                     # pooled standard deviation
n_per_arm <- function(power, frac = 1) {
  ceiling(round(2 * (z_a + qnorm(power))^2 * sigma^2 / (delta * frac)^2, 6))
}
for (f in c(1, 0.75, 0.5))
  cat(sprintf("%3.0f%% of observed effect: n per arm = %s (80%% power), %s (90%% power)\n", 100 * f,
              format(n_per_arm(0.80, f), big.mark = ","), format(n_per_arm(0.90, f), big.mark = ",")))

# Check against R's built-in function (t-based, so very slightly larger)
cat("power.t.test:", round(power.t.test(delta = delta, sd = sigma, power = 0.80)$n, 1), "\n")

# Simulation: exact two-sample t-test using simulated means and variances
set.seed(4279483)
n    <- n_per_arm(0.80)
reps <- 20000
m_diff <- rnorm(reps, delta, sigma * sqrt(2 / n))             # difference in sample means
s2     <- sigma^2 * rchisq(reps, 2 * (n - 1)) / (2 * (n - 1))   # pooled sample variance
tstat  <- m_diff / sqrt(s2 * 2 / n)
cat(sprintf("Simulated power at n = %s per arm: %.3f (%d replicates)\n", format(n, big.mark = ","), mean(abs(tstat) > qt(0.975, 2 * (n - 1))), reps))

Output from running the code

100% of observed effect: n per arm = 7,207 (80% power), 9,648 (90% power)
 75% of observed effect: n per arm = 12,813 (80% power), 17,152 (90% power)
 50% of observed effect: n per arm = 28,828 (80% power), 38,592 (90% power)
power.t.test: 7207.7 
Simulated power at n = 7,207 per arm: 0.803 (20000 replicates)

The simulation draws many sets of group means and variances at the planned sample size and counts how often the two-sample t-test rejects the null. It should land close to the target power; at very small required sample sizes it can fall below it, because the normal approximation is rough there.

13. Analyses Reported by the Sponsor

The registry’s sponsor-reported analyses are mixed-model slope comparisons. They report a slope of -0.64 with p = 0.74 for the Tobacco Retailer Group versus the Non-tobacco Retailer Group, and a slope of -0.84 with p = 0.69 for the Tobacco Retailer Group versus Control. Confidence intervals for these slopes are not reported. No sponsor-reported comparison of the Non-tobacco Retailer Group with Control appears in the registry record.

The unadjusted recalculations instead compare period-level group means using the posted averages and standard deviations. They estimate a difference of 1.16 for Tobacco Retailer Group versus Control and 1.97 for Non-tobacco Retailer Group versus Control, with their respective 95% confidence intervals and unadjusted p-values. Since slope estimates and period-average differences concern different summaries of the outcome, they are not direct substitutes for one another. The registry does not provide enough detail about the mixed-model specification to translate the reported slopes into mean differences.

No discrepancy is flagged between the sponsor’s analysis and the recalculations. The sponsor’s analyses are the registry-reported inferential results; the recalculations are supplementary, simple unadjusted comparisons and should be read with that distinction in mind.

14. Multiplicity, Missing Data and Analysis Populations

The registry lists four outcome measures and four analyses, with 0 secondary outcomes posted. It reports two mixed-model comparisons for the primary outcome: Tobacco Retailer Group versus Non-tobacco Retailer Group, and Tobacco Retailer Group versus Control. The recalculations include comparisons of each retailer group with Control. The registry does not report a sponsor analysis comparing Non-tobacco Retailer Group with Control.

When several outcomes or group comparisons are examined, the chance of at least one apparently notable result can increase. Statistical methods can address this multiplicity, but whether an adjustment was planned or applied is not reported in the registry record. The sponsor’s reported p-values are presented as posted; the registry does not describe them as adjusted. The recalculated p-values are unadjusted. They should not be treated as if they account for the full set of outcomes or comparisons.

The primary outcome population includes all participants who responded to any craving-rating questions during the intervention period. Unassigned participants did not begin the intervention and had no primary outcome data. The registry does not report a full accounting of missing craving responses within the intervention groups, reasons for missingness, or the method used to handle partial rating records in the mixed models. The described averaging procedure indicates how ratings were summarized, but does not specify how missing daily ratings affected each participant’s average.

The outcome counts differ across groups, and the safety populations are reported separately from the primary outcome groups. The registry does not report whether the sponsor used an intention-to-treat population, a per-protocol population, or another named analysis population for the mixed models. It also does not report sensitivity analyses for missing data. Consequently, the posted results describe the analyzed outcome data, but the extent to which missing responses could affect the estimates cannot be determined from the registry record.

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 5% reporting threshold.

Most frequent serious adverse event terms

Serious events with the highest arm-level rate: hospitalization (4.6%).

The registry reports serious adverse events for 109 participants at risk in the Tobacco Retailer Group, 109 in the Non-tobacco Retailer Group, 76 in Control, and 49 in the unassigned group. The corresponding serious-event counts are 2, 5, 0, and 0. Other adverse events are recorded as 0 for each group, and deaths are recorded as 0 for each group.

Hospitalization is listed among the serious adverse event terms, at 4.6%. The registry does not provide further detail in the posted summary about the event term’s group distribution or its relationship to the reported serious-event counts. These safety figures are descriptive counts within the listed at-risk groups; they do not establish whether an event was caused by a study assignment. The registry record does not provide additional information needed to interpret causality, timing, severity beyond the serious classification, or event-by-event clinical context.

The safety group sizes should not be substituted for the primary outcome sample sizes. The outcome averages include participants who responded to craving questions, whereas the serious-event reporting lists separate at-risk counts, including an unassigned group.

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

The posted outcome data describe average self-reported craving during the intervention period among participants who answered at least one craving-rating question. Group averages were 48.33 for the Tobacco Retailer Group, 49.14 for the Non-tobacco Retailer Group, and 47.17 for Control. The standard deviations indicate substantial variation in participant-level averages within each group.

The unadjusted comparisons produce mean differences of 1.16 for Tobacco Retailer Group versus Control and 1.97 for Non-tobacco Retailer Group versus Control. The corresponding 95% confidence intervals, -6.218 to 8.538 and -5.07 to 9.01, include zero and span differences in either direction. The sponsor’s posted mixed-model analyses report slopes with p-values of 0.74 and 0.69 for the two listed comparisons, but do not include confidence intervals. Taken together, these summaries do not precisely establish the direction or size of group differences in craving.

These results do not show that the groups are identical. A result with a confidence interval that includes zero can still be compatible with nonzero differences; the interval’s range communicates uncertainty. Likewise, a p-value does not measure the probability that a treatment effect exists or its practical importance. The registry does not report a threshold for a meaningful craving change, detailed model assumptions, or an adjusted analysis for multiplicity.

The endpoint is self-reported craving, not smoking behavior, cessation, or a direct measurement of neural activity. The registry’s broad study purpose concerns relationships among brain, behavior, and retail exposure, but the primary outcome is craving. The posted outcome summaries cannot, on their own, establish how the assigned visits affected other behaviors or mechanisms. Nor do the results provide medical advice about smoking or retail exposure.

17. Why This Trial Matters Statistically

This study is statistically informative because the outcome comes from repeated real-world craving reports gathered during a structured behavioral manipulation. Its endpoint description makes clear that the analysis is based on participant-level averages assembled from daily ratings, rather than on a single assessment. That distinction matters when interpreting both group means and variability.

The record also illustrates why estimates and precision belong together. The simple mean differences are small relative to the standard deviations, while the confidence intervals extend across zero and allow differences in either direction. Reporting only group means or p-values would leave out important information about how uncertain the comparisons are. The sponsor’s mixed-model slopes add a different view of the repeated data, but the lack of model detail and confidence intervals limits how fully those estimates can be interpreted.

Another useful statistical distinction is between trial design and retrospective planning illustrations. The illustrations translate observed effects into hypothetical participant requirements under stated assumptions. They do not explain the original enrollment target and do not represent observed power. The original design assumptions are not reported in the registry record.

Finally, the outcome and safety counts are drawn from different reporting populations. The primary averages include people who answered any craving questions, while the safety information lists separate at-risk counts, including unassigned participants. Recognizing those populations helps prevent direct comparisons between counts that do not have the same basis. The record therefore highlights the importance of clear estimands, analysis populations, missing-data reporting, and uncertainty when interpreting behavioral trial results.

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 reports the primary outcome averages and standard deviations but does not provide confidence intervals for the group means. The recalculated comparisons offer intervals for the mean differences, but they are unadjusted summaries and do not replace the sponsor’s mixed-model analyses. The registry does not report confidence intervals for the sponsor-reported slopes or enough model detail to determine precisely how to interpret them.

The primary outcome includes participants who answered any craving questions during the intervention period. Details about missing ratings within participants, reasons for missing responses, and the handling of incomplete days are not reported in the registry record. The posted outcome sample sizes therefore describe contributors to the reported averages, but the possible influence of missing ratings cannot be assessed from the record.

The registry describes randomized, single-masked, parallel-group assignment, but does not report who was masked, the randomization procedure, allocation concealment, or adherence to the assigned store-visit schedules. It also does not report the original sample-size justification or detailed analysis plan. Whether any multiplicity adjustment was used is not reported.

The main endpoint is self-reported craving. The registry’s primary outcome summary does not report smoking behavior, cessation, or a neural measure, despite the study’s broader interest in neural and geospatial correlates. The threshold for a practically meaningful change in craving is not reported. Safety results are summary counts, with limited detail about event timing, group-specific event terms, or causality. These omissions restrict interpretation of both the outcome comparisons and safety counts.

19. Related Tutorials

20. Related Statistical Calculators

21. Sources

• ClinicalTrials.gov: GeoScan and Remote Geo Smoking Study: Neural and Behavioral Correlates of Smokers' Exposure to Retail Environments — NCT04279483.
• Registry record last updated 2025-08-14; status last verified 2025-07.
• Results first posted 2025-08-14.
• 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

NCT04279483, titled “GeoScan and Remote Geo Smoking Study: Neural and Behavioral Correlates of Smokers' Exposure to Retail Environments,” is listed as a completed randomized, single-masked, parallel-group study sponsored by Abramson Cancer Center at Penn Medicine. Enrollment is listed as 343. The study began on 2019-12-17 and reached primary completion on 2024-06-10.

The intervention groups were assigned visits to tobacco retailers or non-tobacco retailers during a 4-week period; Control participants were not asked to alter their behavior. The primary outcome was average self-reported cigarette craving during that period. The registry reports group means, sponsor-reported mixed-model slope analyses for two comparisons, and serious adverse event counts. Unadjusted recalculations compare each retailer group with Control and include confidence intervals. The registry does not report the original sample-size justification, detailed mixed-model specification, or whether multiplicity adjustments were used.