Passive Tactile Stimulation for Stroke Rehabilitation: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Passive Tactile Stimulation for Stroke Rehabilitation (NCT03814889), a randomized na study. The primary outcome is Change in Modified Ashworth at 12 Weeks.

2. Clinical Question

3. Trial Design

This was a randomized, parallel study with no masking recorded in the registry. Its phase is listed as not applicable. The registry lists an enrollment of 35 and marks the study completed. The recorded start date is 2019-05-15, and the primary completion date is 2022-12-31.

The study summary describes an effort to optimize gentle vibratory stimulation delivered to the hands of people with chronic stroke and to explore effects on range of movement and spasticity. The registry lists vibration pattern and no vibration as device interventions. Its arm descriptions distinguish acute testing of vibration patterns 1–3 with sham control from longitudinal stimulation using vibration pattern 4.

In the acute description, participants try wearable prototypes in a laboratory. Sensors or electrodes are taped onto the arm and hand to sense muscle activity and record changes during vibration and during periods when vibration is turned off. The longitudinal description says that, if a patient receives Botox injections in the hand and arm, arm function is measured during three months of that standard care. It then describes a take-home wearable device for vibration pattern 4. Further details of that description are not complete in the registry record.

For the primary endpoint, the record says the measure was conducted on only one group. The longitudinal stimulation group has a reported outcome sample of 20, while the acute stimulation and sham-control row has an outcome sample of 0 and no endpoint value. This means the posted primary outcome does not give the randomized, between-group comparison that would directly estimate a difference between vibration and sham. The registry record does not report the allocation procedure, masking beyond “none,” or additional details of the analysis population.

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.

Acute Testing

Longitudinal Testing

Analysis population for the primary outcome, as defined in the registry: This measure was conducted on only one group.

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 Modified Ashworth at 12 Weeks, measured over Change from Baseline Modified Ashworth at 12 weeks.

The primary endpoint is change from baseline in Modified Ashworth at 12 weeks, reported in units on a scale. The Modified Ashworth Scale assesses muscle tone by manually moving an affected joint at set velocities. The registry describes score categories ranging from no increase in muscle tone to a rigid affected part. It also states that lower scale values indicate a better outcome.

This measure is relevant to the study’s focus on spasticity after stroke. Changes in the score are intended to capture change in muscle tone over the assessment period. The registry reports a mean change of 1.1 units for the longitudinal stimulation group, with standard error 0.22, and says the measure was conducted on only one group. It does not report a value for the acute stimulation and sham-control group.

The direction of a change score depends on how it is calculated—for example, whether the recorded change is follow-up minus baseline or baseline minus follow-up. That convention is not reported in the registry record. Consequently, the mean change cannot be labelled as an improvement or worsening using the scale’s lower-is-better direction alone. Nor does a within-group change establish what would have happened under sham stimulation or no vibration.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The endpoint is continuous: it is reported as a mean change in units on a scale. A mean describes the average observed change in the analyzed group. It does not describe every participant’s change, and it can be affected by the range and distribution of individual observations. The registry reports the mean as 1.1 and its standard error as 0.22 for the longitudinal stimulation group.

What does the standard error describe?

General statistical explanation: A standard error describes the sampling uncertainty of an estimated quantity, such as a mean. It is not the same as the spread of individual participant measurements. A smaller standard error generally indicates a more precisely estimated mean under the model or calculation used, while the estimate can still be affected by study design, missing observations, and measurement quality. The registry record does not give the individual-level spread for this endpoint or explain how the standard error was calculated.

A confidence interval is a way of displaying uncertainty around an estimate. It gives a range of values compatible with the data and the method at a stated confidence level. No confidence interval is reported for the primary mean change. Therefore, the standard error alone does not provide the same direct interval interpretation, and an interval should not be inferred by adding or subtracting a quantity from the mean.

What does the Wilcoxon signed-rank test do?

The registry lists a Wilcoxon signed-rank analysis for the longitudinal stimulation group and reports p < 0.001. General statistical explanation: This nonparametric test is commonly used for paired observations, such as measurements on the same participants at two occasions. It ranks the magnitudes of nonzero paired differences and considers their signs. Its assumptions and interpretation depend on how the paired observations and null hypothesis are defined. The registry does not report the exact test setup, the number of paired observations included, or an effect estimate from this test.

A p-value is not the probability that the null hypothesis is true, and it is not a measure of the size or practical importance of an effect. The reported value is evidence against the test’s null model under the test assumptions. Since the result is not a between-group contrast, it cannot establish that the change is attributable to vibration rather than time, standard care, or other factors.

Why does a comparator matter?

General statistical explanation: In a randomized parallel trial, a comparison of outcomes between assigned groups can estimate the effect of assignment, subject to the analysis and follow-up assumptions. A change within one group answers a different question: it describes movement from baseline in that group. Participants might change over time for reasons unrelated to the intervention, so a within-group test does not substitute for a randomized comparison.

The registry lists no vibration and sham control among the study interventions, but the primary endpoint results do not include a control-group outcome. The recorded endpoint sample for the acute stimulation and sham-control row is 0, with no mean or spread. As a result, no two-group difference, confidence interval for a difference, or recalculated comparison is available. The sample sizes and estimates needed for an unadjusted two-arm recalculation are not reported for this endpoint.

How should the scale and change score be read?

The Modified Ashworth Scale measures resistance encountered when an affected joint is moved, and the registry states that lower values indicate a better outcome. But a positive value for change does not by itself reveal whether the underlying score rose or fell: that depends on the subtraction convention. The convention is not reported in the registry record. In addition, the reported mean can conceal variation in individual responses. The posted aggregate figures do not show that variation.

The registry’s general planning note describes planning illustrations based on an observed effect, a two-sided significance level of 0.05, equal allocation, and a normal approximation. Such calculations are prospective planning aids, not observed trial findings or observed power. The record does not provide the numeric planning estimates in the registry record for this explanation, so no sample-size illustration can be interpreted numerically.

10. Confidence Intervals for Each Arm

No two-arm comparison is shown for this trial, but each arm's mean can be given with a t-based 95% confidence interval computed from the posted mean, spread and number of participants.

R code and output

arm <- c("Longitudinal Stimulation: Vibration Pattern 4")
n   <- c(20)
m   <- c(1.1)
sp  <- c(0.22)
sd  <- sp * sqrt(n)      # standard errors converted to SDs
for (i in seq_along(n)) {
  se <- sd[i] / sqrt(n[i]); h <- qt(0.975, n[i] - 1) * se
  cat(sprintf("CARM|%s|%d|%.3f|%.3f|%.3f|%.3f\n", arm[i], n[i], m[i], sd[i], m[i] - h, m[i] + h))
}

CARM  Longitudinal Stimulation: Vibration Pattern 4  20  1.100  0.984  0.640  1.560

11. Analyses Reported by the Sponsor

The registry lists the Wilcoxon signed-rank test as the reported analysis for the longitudinal stimulation group, with p < 0.001. It does not report a test-based effect estimate or confidence interval, and it does not report an endpoint estimate for the acute stimulation and sham-control group.

A recalculation of a two-arm effect from posted summary statistics is not available for this endpoint because the needed comparator outcome is absent. The registry record describes the primary measure as conducted on only one group. Consequently, there is no posted two-group contrast to compare with the reported signed-rank result. The p-value should be understood as the result of the listed within-group analysis, not as evidence from a recalculated vibration-versus-sham comparison.

The registry does not report further details of the sponsor’s analysis, such as the precise paired data included or an effect estimate from the test. The reported analysis therefore offers a statistical test result, while the size and direction of the change remain incompletely characterized. The record does not flag a discrepancy between a sponsor analysis and a recalculation.

12. Multiplicity, Missing Data and Analysis Populations

The registry lists 8 outcome measures and 5 analyses. It reports one primary endpoint and one analysis result in the registry record described for this outcome. The registry record does not report whether the analyses were adjusted for multiple outcomes or multiple tests, whether a testing hierarchy was prespecified, or how the primary endpoint relates to the other outcomes in the analysis plan. Without that information, the reported p-value should not be interpreted as though a multiplicity adjustment was or was not applied.

Missing-data handling is not reported in the registry record. The endpoint has a reported sample of 20 in the longitudinal stimulation group and 0 in the acute stimulation and sham-control row. The registry does not explain whether the endpoint sample reflects all participants assigned to that group, all participants with measurements at baseline and follow-up, or another analysis population. It also does not report reasons for missing outcome values, imputation methods, or sensitivity analyses.

The analysis population is not described in sufficient detail to determine whether the analysis followed randomized assignment, included only participants who completed follow-up, or used another rule. The registry record does not report protocol deviations, exclusions, or how participants receiving standard care were handled in the primary endpoint analysis. The longitudinal description notes a period of standard care for patients who receive Botox injections, but it does not establish how that care was incorporated into the endpoint analysis.

These omissions matter because the interpretation of a change estimate and its p-value depends on which participants contributed data and how missing observations were addressed. They also limit assessment of whether the test result is robust to alternative assumptions. The posted record supports describing the reported group estimate and test, but it does not support reconstructing the full analysis plan or the handling of all outcomes and missing measurements.

13. 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 0 serious adverse events and 0 other adverse events for each of vibration patterns 1, 2, and 3 and for sham control, with 14 at risk in each of those groups. For vibration pattern 4, it lists 0 serious adverse events and 0 other adverse events, with 21 at risk. The record lists 0 deaths in every group.

These are the event counts and at-risk group sizes recorded in the registry. Zero recorded events means that no events of those categories are listed for the stated groups in this record; it does not demonstrate that an event is impossible or that the intervention has no safety risks. The record does not provide further detail about adverse-event definitions, ascertainment procedures, follow-up windows, or whether all events were captured in the same way across groups.

The safety group counts are reported separately from the primary endpoint sample. They should not be treated as the denominator for the Modified Ashworth analysis, whose outcome sample is listed separately. Nor do these counts establish comparative safety, because the registry record does not provide a comparative analysis of adverse-event risk.

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

The posted primary result documents a mean change in Modified Ashworth at 12 weeks of 1.1 units for the longitudinal stimulation group, with standard error 0.22. The registry also reports a Wilcoxon signed-rank p-value of < 0.001 for that group. Together, these figures describe a measured change and a test result under the listed analysis. The scale is used to assess muscle tone, and the registry states that lower values indicate a better outcome.

They do not show whether the underlying score increased or decreased, because the change-score convention is not reported. They also do not estimate the difference between vibration and sham or no vibration: the primary endpoint is described as measured in only one group, and the control row has no reported endpoint value. Thus, the posted results cannot establish that vibration caused the observed change or determine how the group would have fared under a comparator condition.

The p-value does not state the magnitude or clinical importance of the change. The record supplies no confidence interval for the endpoint and no two-group effect estimate. The standard error gives information about uncertainty in the reported mean, but it does not describe the spread of individual changes or resolve the missing direction and comparator information.

The randomized design is relevant to the study’s intended comparisons, but the primary outcome data as posted do not present a randomized contrast. The registry record also does not describe enough about missing data, analysis populations, or multiplicity handling to fully assess the reliability and scope of the reported test. The result is best understood as a within-group statistical finding, with important limits on causal and comparative interpretation.

15. Why This Trial Matters Statistically

This trial is statistically instructive because its design and its posted primary outcome answer different levels of question. The registry describes a randomized parallel study with vibration and no-vibration interventions, but the primary endpoint is stated to have been conducted on only one group. The longitudinal stimulation group has a reported change estimate, while the acute stimulation and sham-control outcome row has no endpoint value. This distinction is central: a trial’s allocation design does not make every posted result a between-group comparison.

The example also shows why the outcome scale and the definition of change must be read together. Lower Modified Ashworth values are described as better, but without the subtraction convention the sign of a mean change cannot be interpreted as improvement or worsening. The reported p-value adds information about the listed test under its assumptions, but it cannot supply the missing direction, effect size, or comparison.

Finally, the record illustrates how incomplete reporting constrains statistical interpretation. No confidence interval is reported for the primary outcome, and the registry does not describe the analysis population, missing-data approach, or multiplicity handling. Those gaps do not erase the posted result, but they limit what can be concluded from it and what can be recalculated independently. Clear reporting of endpoint definitions, group-specific results, and analysis decisions is important for readers assessing both precision and causal relevance.

16. 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 most direct limitation for interpreting this endpoint is that the registry reports primary outcome data for one group only. Although the study is listed as randomized and includes vibration and no-vibration interventions, the acute stimulation and sham-control row has an endpoint sample of 0 and no outcome estimate. A direct treatment contrast cannot be derived from the posted primary outcome results.

The mean change is reported as 1.1 units with standard error 0.22, but the registry does not state the sign convention for calculating change. Since lower Modified Ashworth values are described as better, the missing convention prevents a confident interpretation of whether the reported change corresponds to a rise or fall in the scale. The record also provides no confidence interval or individual-level distribution for the endpoint.

The listed signed-rank test has p < 0.001, but the registry does not explain the exact paired observations analyzed, the number contributing to that test, or the test’s full specification. Missing-data procedures and the analysis population are not reported. The record lists 8 outcome measures and 5 analyses, but does not describe multiplicity control or a testing hierarchy.

The longitudinal arm description includes standard care for patients who receive Botox injections, followed by a take-home device, but the registry record does not explain how this care relates to the endpoint analysis. The description is incomplete, and the record does not provide further details of treatment exposure or adherence for interpreting the 12-week measurement. These limitations do not alter the reported figures, but they restrict conclusions about the direction, precision, and cause of the observed change.

17. Related Tutorials

18. Related Statistical Calculators

19. Sources

• ClinicalTrials.gov: Passive Tactile Stimulation for Stroke Rehabilitation — NCT03814889.
• Registry record last updated 2024-04-08; status last verified 2024-04.
• Results first posted 2024-04-08.
• 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.

20. Record Summary

NCT03814889, titled “Passive Tactile Stimulation for Stroke Rehabilitation,” is recorded as a completed, randomized, parallel study with no masking and an enrollment of 35. The registry lists vibration patterns and no vibration as device interventions. Its summary describes optimizing gentle vibration for people with chronic stroke and exploring range of movement and spasticity.

The primary endpoint is change from baseline in Modified Ashworth at 12 weeks. The registry reports a mean change of 1.1 units, standard error 0.22, for the longitudinal stimulation group, and a Wilcoxon signed-rank p-value of < 0.001. It says the primary measure was conducted on only one group; the acute stimulation and sham-control row has no endpoint value. The change-score direction is not specified, so the result does not establish whether the scale moved in a better direction or provide a vibration-versus-control effect estimate.

The registry lists 0 serious adverse events, 0 other adverse events, and 0 deaths across the recorded groups. Details of missing-data handling, analysis populations, and multiplicity procedures are not reported in the registry record.