Caminamos: A Smartphone App to Connect With Walking Partners: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Caminamos: A Smartphone App to Connect With Walking Partners (NCT03059901), a randomized na study. The primary outcome is System Usability Questionnaire.

2. Clinical Question

The posted comparison estimates the Normal App Notifications group’s mean score minus the More App Notifications group’s mean score. The estimate is 2.78 units on the questionnaire scale, with a confidence interval extending from -6.057 to 11.62 and an unadjusted p-value of 0.538. The estimate is small relative to the reported pooled standard deviation, and the interval includes values on either side of zero. These results describe the measured usability outcome among participants with analyzable app data; they do not answer whether the app changed walking habits or whether either notification schedule improved health.

3. Trial Design

The registry describes a randomized, single-masked, parallel-group study of the Caminamos app. Participants were assigned to More App Notifications or Normal App Notifications. The former group received more app prompts to walk; the latter received fewer. The registry does not state who was masked, how assignment was generated or concealed, or how the notification levels were operationalized.

The registry lists enrollment of 60 and a completed status. Its summary describes the work as Phase I research involving surveys of potential end-users, focus groups to shape a prototype, and development and testing to assess feasibility and functionality for a randomized Phase II intervention. The phase field itself is listed as NA. Start and primary completion dates are recorded as 2015-08-04 and 2016-12-31.

The primary outcome was the System Usability Questionnaire at 4 weeks. Although 60 women were enrolled, 6 did not use the app and were not included in that outcome analysis because they had no data to analyze. The reported arm summaries include 28 participants in More App Notifications and 26 in Normal App Notifications. The record does not explain in detail how the analyzed participants were selected beyond the stated app-use and data-availability reason.

This is an active comparison of notification frequency, so the result addresses the contrast between those two schedules among analyzed participants. It does not isolate the effect of receiving any app versus receiving no app. The registry record does not report the original design assumptions, the intended sample-size justification, or additional detail on masking and allocation procedures.

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

Analysis population for the primary outcome, as defined in the registry: Although 60 women were enrolled, 6 women did not utilize the app during the trial. These 6 women (2 from More App Notifications and 4 from Normal App Notifications) were not included in the data analysis because they did not have data to analyze.

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 System Usability Questionnaire, measured over 4 weeks.

The primary endpoint is the System Usability Questionnaire score at 4 weeks. The registry describes it as a participant-perspective measure of program feasibility, based on ten Likert-type questions about technology user-friendliness. Each question offers five response options ranging from Strongly Agree to Strongly Disagree. The overall score is reported in units on a scale from 0-100, with all scores averaged. The registry notes that 68 or above is average.

Usability is relevant in a study testing a smartphone app because participants need to be able to engage with the technology for the proposed approach to be feasible. A questionnaire about user-friendliness can capture participants’ reported experience of the app. It is distinct from an objective measure of app use, walking, or physical activity, and the registry’s endpoint description does not establish that the questionnaire measures those outcomes.

The reported mean was 60.54 in More App Notifications and 63.32 in Normal App Notifications. The posted comparison is between those group averages. The registry does not report the questionnaire’s individual item results, a measure of app functionality independent of participant ratings, or an additional primary outcome. The outcome therefore supports interpretation of reported usability scores at the specified follow-up, not conclusions about walking behavior or health outcomes.

7. Results Posted in the Registry

• Normal App Notifications: mean 63.32 versus 60.54 on control; difference 2.78 (95% CI -6.06 to 11.62), standardized effect d = 0.17; z-test p = 0.538.

8. Statistical Interpretation

9. Statistical Methods Explained

The primary outcome is continuous: it is a questionnaire score summarized as a mean and standard deviation in each group. The posted arm summaries report the average score and the spread of scores around that average. A mean difference expresses the distance between the two group averages in the questionnaire’s own units. Here, the comparison is defined as Normal App Notifications minus More App Notifications, so a positive estimate means the reported average was higher in the normal-notification group.

What does a mean difference describe?

A mean difference keeps the outcome in its original units, which makes it possible to relate the estimate to the 0-100 questionnaire scale. It describes the difference between group averages, not a change experienced by every participant and not the difference between the highest or lowest scores. The estimated difference is 2.78 units. Its interpretation should be considered alongside the uncertainty interval and the variability of individual scores, rather than on its own.

What does the standard deviation add?

The standard deviation describes how dispersed participant scores were around the mean within an arm. The reported standard deviations are 16.43 for More App Notifications and 16.67 for Normal App Notifications. These measures describe variation among observed questionnaire scores; they are not uncertainty ranges for the group means. The pooled standard deviation of 16.55 is used in the reported standardized effect estimate, which expresses the mean difference relative to a common measure of score spread. The reported standardized effect is 0.168.

How should the confidence interval and p-value be read?

A confidence interval gives a range of estimates compatible with the data under the method used to calculate it. It conveys precision as well as direction. The reported interval, -6.057 to 11.62, contains zero, so the results do not clearly distinguish a positive mean difference from no difference or a negative mean difference. A p-value describes how unusual results at least as inconsistent with a specified null model would be if that model and the analysis assumptions held. The unadjusted p-value is 0.538. It is not the probability that the null hypothesis is true, and it does not quantify the probability that a particular notification schedule is better.

The registry record does not report the detailed inferential procedure, the exact confidence-interval method, or the assumptions used for the posted comparison. The standard error estimates listed for the two arms are 3.105 and 3.269, respectively, but the record does not describe the calculation procedure behind them. The p-value is explicitly unadjusted. The record does not report a multiplicity adjustment or a broader prespecified analysis plan.

Why can variability matter?

When individual scores vary substantially, a given difference between group means is harder to estimate precisely than it would be if scores were more tightly clustered, all else being equal. This is a general statistical principle. It helps explain why the observed mean difference should be interpreted with the interval and standard deviations rather than as a stand-alone finding. A measure of variability does not itself indicate whether the groups differ, and a small p-value or a large p-value does not substitute for interpreting the effect estimate.

Over-dispersion is a general concern for count outcomes when observed variation exceeds what a chosen count model assumes. It is not the central issue for this endpoint, which is a continuous questionnaire score summarized by means and standard deviations. The registry does not report an analysis of over-dispersion or a count model for the primary outcome.

What does randomization contribute?

Random assignment is intended to create groups that can be compared without choosing the notification level based on participants’ characteristics. In general, randomization supports causal interpretation when allocation is implemented as intended and when analysis, follow-up, and missing outcomes are handled appropriately. The record identifies the study as randomized, but does not report the allocation sequence, concealment procedures, balance of participant characteristics, or an intention-to-treat analysis. The stated exclusion of participants who did not use the app means the endpoint analysis does not include all enrolled women.

The registry lists no analysis entries, while the posted outcome comparison includes a mean difference, interval, standardized effect, and p-value. The detailed statistical analysis plan and the method used to generate those estimates are not reported in the registry record. Accordingly, the available figures can be described and interpreted, but assumptions not documented in the record should not be attributed to the investigators.

10. Confidence Intervals and Precision

The estimated mean difference is 2.78 questionnaire units, with a confidence interval from -6.057 to 11.62. The interval is wide enough to include both a negative difference and a positive difference. In the stated direction of comparison, negative values would correspond to a lower average in Normal App Notifications, while positive values correspond to a higher average in that group. Zero represents no difference between the group means.

This range signals uncertainty about the size and direction of the group contrast. The estimate itself is positive, but the interval does not isolate a positive difference: it also includes zero and negative values. The unadjusted p-value of 0.538 is consistent with the lack of a precise distinction between the group averages in this sample. It should not be read as proof that the averages are identical, nor does the interval assign probabilities to individual values within its range.

Precision is affected by the amount of information in the analysis and the variability of scores. The reported standard deviations, 16.43 and 16.67, describe substantial spread relative to the estimated mean difference, while the analyzed arm counts are 28 and 26. The registry record does not report an original precision target or a planned confidence-interval width. It also does not report whether missing questionnaire data were handled in any way beyond excluding the 6 nonusers who had no outcome data.

The interval pertains to the mean usability-score difference under the analysis used for the posted comparison. It does not describe uncertainty about app use, walking, physical activity, or health outcomes. Precision should therefore be judged for this specific questionnaire contrast, not generalized to outcomes that were not reported.

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 calculated from the observed effect, rather than a statement of the trial’s original design. The listed assumptions are a two-sided significance level of 0.05, equal allocation between the two arms, a normal approximation, and an effect to detect defined as the stated fraction of the observed effect.

For 100% of the observed effect, the planning illustration gives 557 participants per arm for 80% power and 745 participants per arm for 90% power. For 75% of the observed effect, it gives 989 participants per arm for 80% power and 1,324 participants per arm for 90% power. For 50% of the observed effect, it gives 2,225 participants per arm for 80% power and 2,978 participants per arm for 90% power.

The 80% power column represents the planned chance of detecting an effect of the specified size under the stated assumptions if that effect is present. The 90% column represents a higher target chance of detection under those same assumptions. These are forward-looking planning quantities, not probabilities that the trial’s finding is correct. Smaller effects require more participants to detect with the same power under the same assumptions. This is not observed power.

The registry record does not report the original sample-size justification or design assumptions for the trial. These illustrations use the observed effect as their starting point and should not be mistaken for the investigators’ original recruitment target or for evidence that the completed study had adequate power.

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 = 2.780; pooled standard deviation σ = 16.546.
2. n = 2 × (1.96 + 0.842)² × 16.546² / 2.780² = 556.1, rounded up to 557 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 <- 2.78                      # observed difference in means
sigma <- 16.5458191326157                     # 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(3059901)
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 = 557 (80% power), 745 (90% power)
 75% of observed effect: n per arm = 989 (80% power), 1,324 (90% power)
 50% of observed effect: n per arm = 2,225 (80% power), 2,978 (90% power)
power.t.test: 557 
Simulated power at n = 557 per arm: 0.798 (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. Multiplicity, Missing Data and Analysis Populations

The primary outcome named in the registry is the System Usability Questionnaire at 4 weeks. The record lists one outcome measure and no secondary outcomes posted. It does not report a prespecified statistical analysis plan that would show whether additional outcomes, time points, subgroup comparisons, or alternative analyses were considered. The posted p-value is unadjusted. Without a complete analysis plan, it is not possible to determine whether multiplicity was an issue in the wider set of analyses conducted.

Multiplicity arises when many statistical comparisons are made: even when no true effects exist, the chance of obtaining at least one apparently unusual result can increase as the number of comparisons grows. This general concern does not establish that such comparisons occurred in this study. The registry record does not report interim analyses, stopping rules, or a correction for multiple testing, so none should be assumed.

The outcome description states that 6 enrolled women did not use the app and were not included in the data analysis because they had no data to analyze. This is a specific missing-outcome issue. The record does not report whether any additional questionnaire responses were missing, how incomplete questionnaires were handled, or whether missingness differed for reasons other than app nonuse. It also does not report imputation or sensitivity analyses.

The record does not describe an intention-to-treat analysis or another analysis population definition beyond the stated exclusion of nonusers. As a result, the comparison reflects participants with usable data under the described exclusion, and its relationship to the randomized groups as a whole is not fully documented. The registry record also does not report a detailed account of protocol deviations or loss to follow-up beyond the nonuse information.

These omissions limit assessment of whether the analysis was selected in advance, how missing outcomes may affect the estimate, and how broadly the posted comparison represents randomized participants. They do not by themselves demonstrate bias or invalidate the reported figures; they mark information that is not reported in the registry record.

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.

The registry reports 0 serious adverse events and 0 other adverse events in each notification group, with 0 deaths in each group. The safety record lists 30 participants at risk in More App Notifications and 30 in Normal App Notifications. These are the figures recorded for safety; they are distinct from the participant counts reported for the usability analysis.

A zero count means no events of the listed type were recorded in the registry’s safety figures. It does not establish that no events occurred outside the reported collection or reporting process, and it does not show that the risk is zero. The registry record does not give further detail on how adverse events were defined, collected, assessed, or attributed. It also does not describe the duration or completeness of safety surveillance.

The safety counts should be read separately from the usability outcome. They do not measure app usability, app use, walking, or physical activity, and they do not resolve the uncertainty around the mean questionnaire-score difference. The record lists no serious adverse-event terms. Interpretation is limited to the event counts and at-risk numbers recorded.

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

The posted results describe average four-week usability scores in two groups assigned to different Caminamos notification schedules. The Normal App Notifications mean was 63.32 and the More App Notifications mean was 60.54. The estimated difference was 2.78 units in favor of the normal-notification group in the direction used for the comparison, but its confidence interval, -6.057 to 11.62, includes zero and values in both directions. The unadjusted p-value was 0.538.

These results do not provide a precise estimate of a difference between the notification schedules. They are compatible with a range of possible mean differences, including no difference. The observed averages alone do not establish that one schedule produces a meaningful improvement in usability, and the registry does not report a threshold for a meaningful between-group difference.

The outcome is a questionnaire-based assessment of technology user-friendliness and program feasibility. It is not a direct measure of walking or physical activity. Although the study summary describes a broader goal of improving walking habits through social support and reduced perceived barriers, the posted primary outcome does not test whether those goals were achieved. No walking or physical activity result is reported in the registry outcome information.

The comparison also cannot establish the effect of using Caminamos versus not using an app. Both groups received app notifications; the contrast is between more and fewer notifications. The record does not report a no-app comparison group or an outcome analysis that addresses app use versus no app.

Finally, 6 enrolled women who did not use the app were excluded from the usability analysis because they had no data to analyze. The posted comparison therefore describes participants with data under that exclusion, and the registry record does not report an analysis addressing the consequences of missing outcomes. The results are informative about the recorded usability ratings, but they do not settle questions about broader effectiveness, generalizability, or outcomes beyond the questionnaire.

16. Why This Trial Matters Statistically

Statistically, this trial is useful for distinguishing a within-app comparison from a test of the app itself. Random assignment was used to compare two notification levels, but there was no reported no-app group. The primary estimate therefore concerns the difference between schedules, not whether an app-based approach changes behavior compared with usual circumstances.

The outcome also illustrates why a numerical estimate should be read with its uncertainty. The mean difference is 2.78, yet the interval extends from -6.057 to 11.62. The point estimate alone could suggest one direction, while the interval makes clear that the data do not tightly determine either the direction or size of the contrast. The p-value adds information about compatibility with a null model, but does not replace the effect estimate or interval.

The study’s enrollment and the endpoint analysis population are not identical: the registry says 60 women enrolled, while 6 nonusers without outcome data were excluded. This makes the definition of the analyzed population important. It also highlights the value of reporting how missing outcomes were handled and whether conclusions are similar under other reasonable approaches.

Finally, the planning illustrations are based on the observed effect and specified assumptions, rather than the trial’s original design justification. They show how detectable effect size, target power, and participant requirements relate in a planning exercise, but they cannot retrospectively establish the study’s original rationale or its realized ability to detect an effect. These distinctions help readers avoid treating a non-significant result as proof of equivalence or treating planning calculations as a finding about what the study observed.

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 contains limited methodological detail. It identifies the study as randomized, single-masked, and parallel, but does not report the allocation sequence, concealment, who was masked, or the exact notification schedules. The original sample-size rationale and detailed statistical analysis plan are not reported in the registry record. Those omissions make it difficult to assess the intended analysis and the implementation of key design features.

The primary outcome is a self-reported usability questionnaire. The registry describes its content and scoring range but does not report item-level results, questionnaire completion details, or a measure of change from a baseline score. The posted endpoint therefore gives a between-group comparison at 4 weeks, not evidence about within-person change or objective app performance.

The analysis excluded 6 women who did not use the app because they had no data to analyze. The reasons for nonuse, any other missing data, and the consequences of excluding these participants are not described in detail. The record does not report an intention-to-treat analysis, imputation, or sensitivity analyses. This limits assessment of how well the analyzed group represents all randomized participants.

The trial compared more with fewer app notifications, so the results cannot isolate the effect of the app versus no app. The registry’s stated broader motivation concerns walking habits, social support, and perceived barriers, but no corresponding primary result for walking or physical activity is posted in the outcome information. Generalizability beyond the enrolled participants is also difficult to assess because participant characteristics are not described in the registry record summarized for this outcome.

Lastly, the confidence interval is broad and includes zero. The record does not provide a prespecified threshold for a meaningful difference in usability, limiting interpretation of practical importance. The safety counts are recorded, but definitions and collection procedures are not described.

18. Related Tutorials

19. Related Statistical Calculators

20. Sources

• ClinicalTrials.gov: Caminamos: A Smartphone App to Connect With Walking Partners — NCT03059901.
• Registry record last updated 2018-09-20; status last verified 2018-08.
• Results first posted 2018-09-20.
• 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

NCT03059901, Caminamos: A Smartphone App to Connect With Walking Partners, is listed as a completed randomized, single-masked, parallel-group study sponsored by Klein Buendel, Inc. The registry records enrollment of 60 and describes two groups receiving more or normal app notifications. The primary outcome was the System Usability Questionnaire at 4 weeks, a participant-rated measure of technology user-friendliness and program feasibility.

The reported mean scores were 60.54 for More App Notifications and 63.32 for Normal App Notifications. The mean difference was 2.78, with a confidence interval from -6.057 to 11.62 and an unadjusted p-value of 0.538. Six women who did not use the app had no data for analysis and were excluded from the usability analysis.

The registry also records 0 serious adverse events, 0 other adverse events, and 0 deaths in each group. The available outcome information addresses usability under two notification schedules; it does not report a primary result for walking or physical activity. Detailed original design assumptions and a statistical analysis plan are not reported in the registry record.