Walk Assist Device to Improve Community Ambulation: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Walk Assist Device to Improve Community Ambulation (NCT03222505), a randomized na study. The primary outcome is Change in Six Minute Walk Test Between Device Turned ON and OFF - Distance Traveled.

2. Clinical Question

3. Trial Design

The registry record describes NCT03222505, “Walk Assist Device to Improve Community Ambulation,” as a completed, randomized, single-masked crossover study sponsored by Shirley Ryan AbilityLab. The listed condition is stroke, and enrollment was 12. The listed start date is 2010-01 and the primary completion date is 2018-05. The record describes the research as examining immediate effects of a prototype wearable robotic stride management assist device designed by Honda R&D®.

A crossover design means that participants are assigned to experience more than one study condition, rather than being assigned to only one condition throughout the study. That general feature is consistent with the listed Device ON and Device OFF conditions. The registry record does not report the sequence, the interval between conditions, the number of visits, or whether any period or carryover effects were considered. It also does not identify who was masked. The arm descriptions labeled Treadmill and Overground Walking are blank, so their relationship to the conditions cannot be established from the registry record.

The primary outcome is the change in six-minute walk distance between the device-on and device-off conditions, with a Day 1 timeframe. The posted group summaries include 12 participants in each condition. The record lists four outcome measures and two secondary outcomes relating to transcranial magnetic stimulation excitability in rectus femoris and tibialis anterior. It does not report a statistical analysis plan or analysis details for those secondary outcomes. No phase is listed. Details about the original design rationale and sample-size justification are not reported in the registry 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 Change in Six Minute Walk Test Between Device Turned ON and OFF - Distance Traveled, measured over Day 1.

The primary endpoint is distance traveled during the Six Minute Walk Test (6MWT), recorded in feet. The test measures how far a participant can walk indoors on a flat, hard surface over six minutes. The registry description allows assistive devices as needed and instructs participants to cover as much ground as possible, with the option to stop to sit or stand if needed. Distance is measured with a measuring wheel.

As a continuous measure, distance retains information about the scale of walking performance rather than reducing the outcome to a category such as success or failure. In general, a fixed-duration walking test can provide a standardized way to describe walking endurance under the test conditions. It does not, by itself, describe every aspect of community ambulation or explain how a participant moves in all settings.

The endpoint compares distance when the device was turned on with distance when it was turned off. The registry record gives the timeframe as Day 1 and reports a mean and standard deviation for each condition. It does not report further details about the testing environment, the exact procedure beyond the endpoint description, or whether any additional measurement safeguards were used. The result should therefore be read as a comparison of the recorded test distances under the two listed conditions, not as a complete account of mobility in everyday life.

7. Results Posted in the Registry

• Device OFF: mean 804.80 versus 837.10 on control; difference -32.30 (95% CI -312.71 to 248.11), standardized effect d = -0.09; z-test p = 0.821.

8. Statistical Interpretation

9. Statistical Methods Explained

The primary outcome is continuous: it is a measured distance in feet. The posted summaries report the mean and standard deviation in each condition, followed by a mean difference, confidence interval, and unadjusted p-value. These statistics answer related but distinct questions. The means describe the average distances in each condition, the mean difference describes the contrast, the confidence interval describes uncertainty about that contrast, and the p-value assesses compatibility with a specified null model under the analysis assumptions.

What do the means and standard deviations describe?

The mean is the arithmetic average of the recorded distances in a condition. The Device ON mean is 837.1 feet, and the Device OFF mean is 804.8 feet. A mean is a useful summary of the group’s central level, but it does not show whether participant values cluster closely around that level or vary widely. The standard deviation describes the spread of the observations around the mean. The reported standard deviations are 342.2 feet for Device ON and 358.5 feet for Device OFF.

In general, a standard deviation is a description of variation among measurements, not a confidence interval for the mean and not a measure of uncertainty in the treatment contrast by itself. When outcomes are measured repeatedly on the same participants, as in a crossover design, the relationship between each participant’s measurements can also matter. The registry record does not report participant-level paired distances or the method used to account for that relationship.

How should the mean difference be read?

The reported contrast is the Device OFF mean relative to the Device ON mean. Its estimate is -32.3 feet. The negative sign indicates that the reported mean in the Device OFF condition is lower than the mean in Device ON. The estimate summarizes the direction and size of the difference in the posted summaries; it does not establish that the same direction or size applies to every participant.

In general, a mean difference is expressed in the outcome’s original units, which makes it possible to interpret the contrast as a distance rather than as a standardized score. The registry record does not report a prespecified threshold for interpreting the size of this difference, nor does it provide a clinical-importance criterion. Statistical interpretation should therefore focus on the estimate and its uncertainty rather than attaching an unreported threshold to the value.

What does the p-value tell a reader?

The unadjusted p-value is 0.821. In general, a p-value is calculated under a null hypothesis and a statistical model; it describes how unusual data at least as discrepant as those observed would be if that null model and its assumptions held. It is not the probability that the null is true, nor does it quantify the size or importance of an effect. The word “unadjusted” indicates that the listed value is not described as adjusted for multiple comparisons. The registry record does not report the test used, its assumptions, or whether any adjustment was planned.

The p-value should not be read in isolation. A large p-value does not establish equivalence, and it does not show that the conditions have identical mean distances. Equivalence or noninferiority conclusions require a prespecified margin and an analysis suited to that question; such a margin is not reported in the registry record. The confidence interval gives a more direct view of the range of contrasts compatible with the data under the analysis assumptions.

Why does a crossover design matter statistically?

In general, a crossover design can use within-participant comparisons because participants experience more than one condition. The statistical analysis may therefore depend on the order in which conditions occur, the association between repeated measurements, and whether a condition’s effect carries into a later period. These are general design considerations, not evidence that any particular issue occurred in this study. The registry record identifies a randomized crossover model but does not report sequence allocation, washout timing, period-specific data, or a model accounting for participant and period effects.

The condition summaries each list 12 participants, but those counts do not reveal how the paired observations were analyzed. The reported comparison includes a confidence interval and p-value, yet the detailed method behind them is not reported in the registry record. The record also does not report whether the analysis followed an intention-to-treat, per-protocol, or other analysis population. Without these details, readers can describe the posted estimates but cannot fully assess the analysis choices or reproduce the comparison from participant-level data.

In general, over-dispersion means that observed variability is greater than a model expects. When that occurs, standard errors based on an overly restrictive variance assumption can be too small, making confidence intervals too narrow and tests too optimistic. The posted standard deviations indicate the spread of distances, but there is no reported model or diagnostic information from which to assess over-dispersion in this study. This general issue is not a claim that over-dispersion affected the reported result.

10. Confidence Intervals and Precision

The reported mean difference is -32.3 feet, and its confidence interval extends from -312.7 to 248.1 feet. This interval is wide relative to the point estimate and includes zero. It also includes negative and positive contrasts, so the data are compatible with the Device OFF condition having a lower mean, a higher mean, or little difference under the analysis that generated the interval. The interval communicates substantial uncertainty about the size and direction of the contrast.

In general, a confidence interval is a range constructed by a statistical procedure that would capture the target parameter at its stated rate across repeated uses of that procedure, if the model assumptions hold. The registry record does not state the confidence level, so it is not appropriate to assign one to this interval. The interval should not be interpreted as containing every plausible individual participant response; it concerns the estimated group contrast.

The wide interval is consistent with limited precision in a study enrolling 12 participants, but the registry record does not report a formal explanation for the width. The listed standard deviations, 342.2 feet for Device ON and 358.5 feet for Device OFF, describe variation in the condition measurements. In a crossover study, the uncertainty of a paired contrast can also depend on the within-participant relationship, which is not reported. The interval is therefore more informative about precision of the estimated mean difference than either group standard deviation alone.

The interval’s inclusion of zero does not establish that the conditions are equivalent. To support an equivalence interpretation, a meaningful margin would need to be defined and the uncertainty assessed against it; no such margin is reported in the registry record. Conversely, the point estimate alone should not be treated as a precise estimate of an effect, because the interval spans a broad range of contrasts. The most defensible reading is that the posted comparison is imprecise.

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 report participants per arm, not total events. They use a normal approximation, two-sided significance level 0.05, and equal allocation between two arms. Each row considers an effect to detect set at a stated fraction of the observed effect. These are planning assumptions for an illustration; the original trial’s design assumptions and sample-size justification are not reported in the registry record.

The 80% power column describes the number of participants per arm that the illustration estimates would be needed to have an 80% chance of detecting the chosen effect, if the planning assumptions and model were appropriate. The 90% power column gives the corresponding planning target for a 90% chance. Power is a property of a design under specified assumptions, not a guarantee that a study will detect an effect.

The table considers planning for 100%, 75%, and 50% of the observed effect. The listed participant targets increase as the assumed effect becomes smaller. This is a general feature of study planning: smaller effects are harder to distinguish from random variation, and typically require more participants for the same planned power. The observed effect in a small study may itself be uncertain, so basing a future design directly on it can be an unstable choice.

These figures are not observed power. They are planning illustrations based on the observed effect and the stated assumptions. The registry record does not report that these assumptions were used when the trial was designed, and the illustrations should not be read as the study’s original target enrollment or as a retrospective measure of whether the trial was adequately powered.

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 = -32.300; pooled standard deviation σ = 350.445.
2. n = 2 × (1.96 + 0.842)² × 350.445² / -32.300² = 1,847.9, rounded up to 1,848 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 <- -32.3000000000001                      # observed difference in means
sigma <- 350.44478167038                     # 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(3222505)
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 = 1,848 (80% power), 2,474 (90% power)
 75% of observed effect: n per arm = 3,286 (80% power), 4,398 (90% power)
 50% of observed effect: n per arm = 7,392 (80% power), 9,896 (90% power)
power.t.test: 1848.8 
Simulated power at n = 1,848 per arm: 0.799 (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 registry record lists a primary endpoint and two secondary outcome names. It also reports four outcome measures in total. The secondary outcomes concern percent change in transcranial magnetic stimulation excitability for rectus femoris and change in percent excitability for tibialis anterior between Device ON and Device OFF. Secondary numerical results are not reported in the registry record.

When several outcomes or comparisons are tested, the chance of obtaining at least one apparently unusual result can increase. In general, multiplicity refers to the need to consider a family of related tests and how conclusions should account for them. Some analysis plans adjust for multiple comparisons; others define a primary outcome and treat additional outcomes as exploratory. The registry record does not report a multiplicity strategy, a hierarchy among outcomes, or planned adjustments. The primary comparison’s p-value is described as unadjusted, so it should not be assumed to account for other outcomes or tests.

Missing-data handling is also not reported in the registry record. The posted primary summaries list 12 participants in each condition, matching the enrollment count, but those counts do not explain whether every participant had usable measurements, how incomplete observations would have been handled, or whether any data were imputed. The record does not report reasons for missing measurements or a sensitivity analysis for missingness.

The analysis population is not reported in the registry record. The record does not say whether the comparison included all randomized participants, only those with complete measurements, or another defined group. Nor does it provide participant-level data or a detailed statistical analysis plan. These omissions matter because choices about outcome selection, multiplicity, missing data, and analysis population can affect estimates and their interpretation. They do not demonstrate that a particular bias occurred; they limit what can be assessed from the posted 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 adverse-event record lists 12 participants at risk in each condition. For Device ON, it reports zero serious adverse events, zero other adverse events, and zero deaths. For Device OFF, it also reports zero serious adverse events, zero other adverse events, and zero deaths. No serious adverse-event terms are listed.

These figures describe what was recorded for the study conditions and the participants listed as at risk. A zero count means that no events in the listed category were reported in the registry record; it does not establish that an event is impossible or that the conditions have identical safety profiles. With a small number of participants, event counts provide limited information about uncommon events. The record does not report additional detail about how events were defined, collected, monitored, or attributed to a condition, nor does it give follow-up duration for safety reporting. The counts should therefore be read as a narrow description of the posted adverse-event data, not as a complete safety evaluation.

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

The posted primary result is a comparison of six-minute walk distance under Device ON and Device OFF conditions in a randomized crossover study enrolling 12 people with stroke. The means were 837.1 feet and 804.8 feet, respectively. The reported contrast, expressed as Device OFF relative to Device ON, was -32.3 feet. Its confidence interval, from -312.7 to 248.1 feet, is broad, and the unadjusted p-value is 0.821. Together, these statistics indicate that the estimated mean difference is imprecise.

The result does not establish that the two conditions are equivalent. No equivalence margin is reported, and the interval covers both negative and positive contrasts. It also does not establish that a difference of any particular size is present. The p-value is not a measure of the probability that one condition is better, and the point estimate should not be treated as a precise description of the underlying difference.

The outcome concerns a six-minute indoor walking test on a flat, hard surface, with assistive devices permitted as needed. It is a specific measurement under the test conditions. The result alone does not describe walking in every community setting, sustained use over time, or other outcomes named in the study aim. The registry record identifies additional interests in walking energetics, gait kinematics, and corticospinal excitability, but the primary comparison cannot stand in for results on those domains.

The record’s zero adverse-event counts are likewise limited to the listed categories and participants at risk. They do not establish absence of all possible harms. Important design and analysis details—including the crossover sequence, masking identity, missing-data approach, and statistical plan—are not reported in the registry record. The most cautious interpretation is that this small study reports an uncertain primary comparison, with limited posted information for assessing the analysis and broader implications.

16. Why This Trial Matters Statistically

This trial is statistically informative as an example of the distinction between an estimated effect and the precision of that estimate. The mean difference is -32.3 feet, but its confidence interval runs from -312.7 to 248.1 feet. A point estimate can appear straightforward while the interval shows that the data leave substantial uncertainty about both direction and magnitude. Reading the estimate together with its interval avoids treating a single number as a complete account of the evidence.

The crossover model also makes the structure of the data important. In general, participants serving in more than one condition can provide within-person comparisons, but analysis may depend on paired measurements and design features such as treatment order and period. The registry record identifies a randomized crossover design but does not report the information needed to assess those details. This illustrates why the design label alone is not enough to reconstruct the statistical analysis.

The study also demonstrates why a nonsmall p-value is not equivalent to evidence of no difference. The unadjusted p-value is 0.821, while the confidence interval remains broad. Neither result establishes equivalence. Interpretation depends on the study question, a clearly defined threshold if equivalence is intended, and the precision of the estimate.

Finally, planning based on an observed effect can produce very large participant targets when the assumed effect is reduced. The planning illustrations are not the trial’s original justification and are not observed power. Their value is to show how assumptions about the effect size and desired power shape a future planning exercise. The registry record’s limited analysis details make transparent reporting especially important for readers who want to assess how the posted comparison was obtained.

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 enrollment was 12, and the confidence interval for the primary mean difference is broad. This limits precision and makes the estimate compatible with substantially different contrasts. The registry record does not report participant characteristics beyond the listed condition, so readers cannot assess how the enrolled group compares with other populations or which characteristics might affect the outcome.

The study is identified as randomized, single-masked, and crossover, but several design details are not reported in the registry record. These include the randomization sequence, the timing and order of conditions, any washout interval, the identity of the masked party, and whether period or carryover effects were assessed. The arm descriptions for Treadmill and Overground Walking are blank. As a result, the relationship between those labels and the primary comparison is unclear.

The record does not report a detailed statistical analysis plan, the analysis population, missing-data methods, or the procedure used to calculate the primary comparison. Participant-level paired measurements are not posted in the record summarized by the reported statistics. The record also lists secondary outcomes without numerical secondary results, and does not describe a multiplicity strategy. These omissions restrict assessment of robustness and of the broader study aims.

The endpoint is a six-minute indoor walk test on a flat, hard surface, with assistive devices allowed as needed. It is a defined measure of walking endurance, but the result is not a direct measurement of every aspect of community ambulation. The registry record’s safety counts are limited to the reported categories and do not provide detailed event definitions, monitoring procedures, or follow-up duration. These are limitations of what can be assessed from the record, not evidence that an unreported procedure or event did or did not occur.

18. Related Tutorials

19. Related Statistical Calculators

20. Sources

• ClinicalTrials.gov: Walk Assist Device to Improve Community Ambulation — NCT03222505.
• Registry record last updated 2019-08-28; status last verified 2019-08.
• Results first posted 2019-08-28.
• 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

NCT03222505 is a completed randomized, single-masked crossover study sponsored by Shirley Ryan AbilityLab. It enrolled 12 people with stroke and examined a prototype Honda stride-assist device. The primary endpoint was Day 1 six-minute walk distance in feet, comparing Device ON with Device OFF.

The posted means were 837.1 feet with Device ON and 804.8 feet with Device OFF. The reported mean difference for Device OFF relative to Device ON was -32.3 feet, with a confidence interval from -312.7 to 248.1 feet and an unadjusted p-value of 0.821. The interval indicates substantial uncertainty around the estimated contrast. The record lists two secondary outcome names but does not report their numerical results. It reports zero serious adverse events, zero other adverse events, and zero deaths in each condition. Details about the analysis plan, missing data, and crossover sequence are not reported in the registry record.