Sensory Augmentation Methods in Stroke: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Sensory Augmentation Methods in Stroke (NCT03988400), a randomized early/phase 1 study. The primary outcome is Percentage of Participants With Related or Serious Adverse Events.

2. Clinical Question

3. Trial Design

This was a randomized, triple-masked, parallel-group early-phase study of sensory augmentation methods in stroke. The registry lists an enrollment of 44 and 22 participants in each arm. The study status is completed. The recorded start date is 2022-04-04, and the primary completion date is 2023-04-26. The sponsor is the VA Office of Research and Development.

Both groups completed treadmill walking training at self-selected speed for 8 sessions over 4 weeks. The intervention contrast was how vibration was controlled: the sensory augmentation arm used vibration magnitude linked to the mechanical state of the pelvis, whereas the random vibration arm varied vibration randomly from step to step. Both interventions are classified as behavioral.

The stated rationale is that enhancing sensory information during walking may affect foot placement accuracy and balance in chronic stroke survivors. The listed secondary outcomes include intervention feasibility measured by walking time, change in fear of falling, change in Functional Gait Assessment Score, change in Activity-specific Balance Confidence Score, and change in overground self-selected gait speed.

The registry lists 10 outcome measures and 10 analyses, while 5 secondary outcomes are posted. The posted primary outcome is the count of participants with related or serious adverse events, accumulated over the 4-week intervention. The analysis population is described as all participants who began the intervention. Further details about sequence generation, concealment, the masking procedures, and protocol deviations 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.

Before assignment: Participants completed an initial assessment session prior to their assignment to a study arm/group. No qualified participants were enrolled into the study, but then excluded from the study before group assignment.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: Analysis was performed on all participants who began the intervention.

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: Information is presented for all participants who enrolled in this intervention study.

6. Primary Endpoint

The registry lists the primary outcome as Percentage of Participants With Related or Serious Adverse Events, measured over Cumulative over the course of the 4-week intervention.

The primary endpoint counts participants who experienced an adverse event potentially related to the intervention or any serious adverse event during the 4-week intervention. It is expressed as a count of participants rather than a measurement of event severity or a change in walking performance. The registry says safety information was gathered through participant self-report, or communication with a caregiver when necessary.

This type of endpoint is used to describe safety experience during an intervention. Including potentially related events and serious events gives the monitoring outcome a broad scope: an event could be considered potentially related even if it is not serious, while a serious event is included regardless of the stated relationship to the intervention. The registry gives skin irritation and falls as examples of potentially related events, and hospitalization and death as examples of serious events.

The posted counts are 1 participant in Sensory Augmentation and 0 in Random Vibration. The outcome does not establish whether an event was caused by an intervention. Nor does it directly answer whether sensory augmentation changes foot placement, balance, fear of falling, or gait speed. Those questions require their own outcome results and interpretation. The registry’s primary outcome description and its analysis population should therefore guide interpretation of this count; additional event-classification details are not reported in the registry record.

7. Results Posted in the Registry

• Random Vibration: 0.0% (0/22) versus 4.5% (1/22) on control; risk difference -4.5 percentage points (95% CI -13.2 to 4.2); unadjusted two-proportion z-test p = 0.312.

8. Statistical Interpretation

9. Statistical Methods Explained

The primary result is a comparison of participant counts between two randomized arms. For a binary outcome, each participant is classified according to whether the event occurred or did not occur during the defined observation period. Here the outcome is the count of participants with related or serious adverse events over the 4-week intervention. The arm listings show 22 participants at risk in each group, with the posted outcome counts of 1 and 0.

What the risk difference expresses

The risk difference is the event proportion in one group minus the event proportion in the reference group, using the direction specified for the comparison. It is an absolute contrast, rather than a relative measure. The reported risk difference for Random Vibration compared with Sensory Augmentation is -0.045. A negative estimate means the observed event proportion was lower in the comparison arm. Its magnitude is an estimate from the observed participants, not a guaranteed difference in a future sample.

Risk differences are useful because they express the contrast on the same scale as the event probability. They can be interpreted without converting the result into a relative risk or odds ratio. Such other measures are not needed to describe the posted result, and no additional effect measure is calculated. The risk difference may be unstable when the event is uncommon and the group counts are small; a small number of participants changing event status can materially alter the estimate.

What the confidence interval adds

A confidence interval represents uncertainty around an estimate under the statistical method used. The interval for this risk difference extends from -0.133 to 0.042. Its lower and upper limits span both negative and positive values, so the data do not pin down the direction of the population contrast. The interval also communicates that the estimate is not precise: a range of materially different contrasts remains compatible with the data and method.

General statistical explanation: a confidence interval is not a statement that the true value has a particular probability of falling inside the already calculated interval. Rather, the method is designed to produce intervals with a specified long-run coverage under its assumptions. The exact construction method for the recalculated interval is not reported in the registry record. Its limits should therefore be interpreted alongside the stated recalculation and not as a substitute for the sponsor’s own test.

Fisher’s exact test and the recalculation

The registry lists Fisher Exact as the sponsor’s method and reports p = 1 for the superiority analysis. Fisher’s exact test is a method for testing association between group assignment and a binary outcome, using the observed counts in the groups. General statistical explanation: it evaluates how unusual the observed allocation of events would be under a null model of no association, conditional on the margins of the count table. It does not measure the size of an effect, and a p-value does not show that the treatments are equivalent.

The separate recalculated result is an unadjusted comparison derived from the posted summary statistics. Its reported p-value is 0.312, and its risk difference and confidence interval are reported as -0.045, -0.133, and 0.042. The recalculated value is not the sponsor’s analysis. Different methods or test conventions can give different inferential results, particularly with sparse binary counts. The registry record does not report enough detail to determine the exact reason for the difference between the sponsor’s p-value and the recalculated p-value.

Why sparse counts need caution

General statistical explanation: when event counts are sparse, the data provide limited information about the event probability in either group. Normal approximations can be less dependable in that setting, which is one reason exact methods are often considered for binary outcomes. An exact test still cannot overcome the limited information in the sample; it only uses a different inferential procedure. Sparse data can also yield wide confidence intervals and make estimates sensitive to individual observations.

The stated planning illustrations use the observed effect as their basis, with a two-sided significance level of 0.05, equal allocation, and a normal approximation; the effect to detect is a stated fraction of the observed effect. These are planning illustrations, not observed power and not evidence that the trial had adequate power for its primary outcome. The original protocol’s sample-size justification is not reported in the registry record. No sample-size assumptions beyond those stated for the planning illustrations should be inferred.

Limits of the inferential result

The sponsor’s result and the recalculated comparison address a group-level difference in recorded event counts. Neither determines whether an event was caused by the intervention, and neither provides a detailed account of event severity or timing. General statistical explanation: statistical significance, confidence interval width, and clinical or practical importance are distinct considerations. A p-value alone does not measure the importance of an effect, while an interval helps show the range of estimates compatible with the data. In this study, the small counts and interval spanning both directions call for restrained interpretation.

10. Confidence Intervals and Precision

The 95% confidence interval for the recalculated risk difference is -0.133 to 0.042. It includes zero and extends in both directions, so the estimate does not determine whether the event proportion is higher in Sensory Augmentation or in Random Vibration. The interval is broad relative to the point estimate, reflecting limited precision in this comparison.

The posted event counts are 1 participant in Sensory Augmentation and 0 in Random Vibration, with 22 participants at risk in each arm. With so few recorded events, the observed contrast is particularly sensitive to the experience of individual participants. The point estimate should not be treated as a stable measure of the underlying difference, and the confidence interval is important context rather than a decorative range around a definitive effect.

General statistical explanation: precision describes how tightly an estimate is determined by the data and statistical method. A narrow interval generally indicates greater precision than a wide one, but precision does not guarantee that an estimate is unbiased or that it is clinically important. A confidence interval also does not establish equivalence merely because it includes zero. To support an equivalence conclusion, a study would need an appropriate design and prespecified margins; those details are not reported in the registry record.

The registry-reported Fisher Exact p-value is 1, while the recalculated unadjusted p-value is 0.312. These values arise from different analyses and should not be combined as if they were one result. Neither p-value resolves the uncertainty indicated by the interval. Further details about the recalculated interval’s construction and the sponsor’s full analysis plan are not reported in the registry record.

11. Analyses Reported by the Sponsor

The registry reports a Fisher Exact superiority analysis with p = 1. This is the sponsor’s posted inferential result. The separate recalculation reports a risk difference of -0.045 for Random Vibration relative to Sensory Augmentation, a 95% confidence interval from -0.133 to 0.042, and an unadjusted p-value of 0.312. The recalculation is a simple unadjusted analysis of the posted summary statistics and is not a replacement for the sponsor’s analysis.

The values are not identical because they come from different analysis procedures. Fisher’s exact test evaluates the binary counts using an exact testing approach; the recalculated result uses a different comparison. The registry record does not report enough methodological detail to identify every reason for the numerical difference. The reported discrepancy indicator is false, but the sponsor’s test and the recalculation should still be identified separately rather than presented as interchangeable.

Neither result establishes that the arms have the same safety experience. The sponsor’s p-value is not a probability that there is no difference, and the recalculated interval includes both negative and positive values. The clearest reading is that the posted counts are sparse and the between-group contrast is uncertain. The sponsor’s own analysis takes precedence when interpreting the sponsor-reported inferential finding; the recalculation provides an additional unadjusted description of the posted summary data.

12. Multiplicity, Missing Data and Analysis Populations

The registry lists 10 outcome measures and 10 analyses, and 5 secondary outcomes are posted. The listed secondary outcomes address feasibility and changes in fear of falling, functional gait assessment, balance confidence, and overground self-selected gait speed. Multiple outcomes can raise a multiplicity concern when many hypotheses are tested, because the chance of at least one apparently unusual result can increase across a collection of tests. Whether any multiplicity adjustment was planned or applied is not reported in the registry record.

The primary safety outcome is distinct from those secondary outcomes, but the registry record does not report a full statistical analysis plan describing how the outcomes were ordered, which tests were confirmatory, or how any multiplicity was handled. No conclusion about adjustment should be inferred from the number of analyses alone. The posted p-value for the primary outcome is identified as a Fisher Exact result; whether it was adjusted for any other analyses is not reported in the registry record.

The registry says the primary analysis was performed on all participants who began the intervention. Each arm lists 22 participants at risk. Further information about the relation between this analysis population and the randomized population, any exclusions after randomization, or the handling of participants who did not complete intervention sessions is not reported in the registry record. The treatment of missing outcome information is also not reported in the registry record. It is therefore not possible to determine from the posted description whether missingness occurred or how it would have been handled.

Details about interim analyses, stopping rules, protocol deviations, and sensitivity analyses are not reported in the registry record. These omissions limit assessment of the full inferential context, but do not alter the basic description of the posted counts. The sample-size illustrations use the observed effect and stated planning assumptions; they should not be interpreted as observed power or as confirmation of the original design rationale.

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.

Most frequent serious adverse event terms

Serious events with the highest arm-level rate: stroke (4.5%).

The primary outcome records 1 participant in Sensory Augmentation and 0 in Random Vibration. In the adverse-event listing, 22 participants were at risk for serious events in each arm; 1 participant in Sensory Augmentation and 0 in Random Vibration are listed as affected by serious events. Other adverse events are listed for 0 participants in each arm, and deaths are listed as 0 in each arm. The registry also lists Stroke as a top serious term at 4.5%.

These records describe counts, not causal conclusions. The primary endpoint combines potentially intervention-related adverse events with serious adverse events, while the adverse-event listing separates serious and other events. The registry does not report enough detail to establish whether the serious event and the primary endpoint count refer to the same classification or to describe the relationship of the event to the intervention. The Stroke term should not be assumed to identify the cause of an event or a treatment effect.

Small counts provide limited information about uncommon harms. A count of 0 in an arm does not show that the risk is absent, and a recorded event does not establish that the intervention caused it. The result is a safety snapshot over the specified intervention period, based on the registry’s stated monitoring and analysis population.

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

The result describes the recorded safety outcome in a randomized study of sensory augmentation compared with random vibration during treadmill training after stroke. The registry reports 1 participant with a related or serious adverse event in Sensory Augmentation and 0 in Random Vibration. The recalculated risk difference is -0.045 for Random Vibration relative to Sensory Augmentation, with a 95% confidence interval from -0.133 to 0.042. The sponsor’s Fisher Exact analysis reports p = 1.

Those findings do not establish that either approach is safer. The event counts are sparse, and the interval includes both directions of effect. The absence of a recorded event in one group is not proof that no event could occur. Conversely, the event recorded in the other group cannot be attributed to the intervention based on these counts alone.

The primary outcome does not answer whether sensory augmentation improves foot placement accuracy, balance, walking time, fear of falling, functional gait assessment, balance confidence, or gait speed. Those are separate questions, and their results are not included in the primary outcome details. The registry describes the study’s rationale and lists secondary outcomes, but those details should not be mistaken for evidence that a particular outcome changed.

The trial’s early-phase designation, the listed enrollment of 44, and the limited number of recorded events place the safety contrast in a preliminary setting. The result can inform description of what was recorded during this intervention period, but it cannot settle longer-term safety, generalizability to all chronic stroke survivors, or the comparative value of the approaches. Details needed to make those broader judgments are not reported in the registry record.

General statistical explanation: a non-significant test does not demonstrate no effect, and a confidence interval that includes zero does not prove equivalence. Interpretation requires attention to the estimate, its uncertainty, the design, the endpoint, and the completeness of reporting. Here, the most defensible reading is descriptive and cautious: few events were recorded, the estimated contrast is imprecise, and the posted analysis does not establish a clear between-group difference.

15. Why This Trial Matters Statistically

This trial matters statistically because it illustrates the limits of inference from a binary safety outcome with sparse events. The observed counts are 1 participant in Sensory Augmentation and 0 in Random Vibration. A numerical contrast can be calculated from those counts, but a point estimate alone can make the data appear more decisive than they are. The broad confidence interval shows that the underlying difference remains uncertain.

The sponsor’s Fisher Exact result and the unadjusted recalculation also illustrate why methods need to be named. Both concern the same posted outcome, but they do not produce the same p-value. Reporting the analysis method alongside the estimate prevents readers from treating every test result as interchangeable. The sponsor’s test is the registry-reported analysis, while the recalculation is an additional summary based on the posted counts.

The safety endpoint is also distinct from the trial’s motivating question about sensory information, foot placement, and balance. This distinction matters because a safety count cannot establish a change in gait or balance. The registry lists secondary outcomes that address several of those questions, but results for those outcomes are not included in the registry record for the primary result.

For clinical researchers and biostatisticians, the example emphasizes the value of complete reporting: clear analysis populations, endpoint definitions, missing-data handling, and a prespecified approach to multiple outcomes help readers interpret even a straightforward comparison. Several of those details are not reported in the registry record. The available data still support a careful account of the observed events and their uncertainty, without extending that account into claims of efficacy, equivalence, or causation.

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 outcome is based on few recorded events: 1 participant in Sensory Augmentation and 0 in Random Vibration. The resulting estimate is imprecise, and its confidence interval includes negative and positive risk differences. This limits the strength of any conclusion about a difference between groups. The registry record does not report event timing, detailed severity, adjudication procedures, or the classification of the event’s relationship to the intervention.

The primary analysis population is described as all participants who began the intervention, but the registry does not report whether this population matches all randomized participants or how participants who did not begin or complete training were handled. Missing-data procedures are not reported in the registry record. These gaps limit assessment of whether the analyzed counts represent the full randomized experience.

The registry reports Fisher Exact as the sponsor’s method and also supports an unadjusted recalculation with a different p-value. Additional details of the sponsor’s statistical analysis plan and the recalculated interval method are not reported in the registry record. The full approach to multiplicity, interim analyses, and protocol deviations is also not reported.

The record describes 8 training sessions over 4 weeks, but the duration of follow-up for safety beyond the intervention is not reported. The trial is early phase and lists 44 participants, which limits how broadly the observed safety experience can be generalized. The primary endpoint does not measure gait stability or the listed secondary outcomes; those findings cannot be inferred from the safety counts. Finally, the original sample-size justification is not reported in the registry record, so the planning illustrations based on the observed effect should not be mistaken for the trial’s original design rationale or observed power.

17. Related Tutorials

18. Related Statistical Calculators

19. Sources

• ClinicalTrials.gov: Sensory Augmentation Methods in Stroke — NCT03988400.
• Registry record last updated 2024-09-19; status last verified 2024-08.
• Results first posted 2024-09-19.
• 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

NCT03988400 is a completed, randomized, triple-masked, parallel-group early-phase study sponsored by the VA Office of Research and Development. It lists 44 participants and compares sensory augmentation with random vibration during treadmill training. Each group’s schedule was 8 sessions over 4 weeks at self-selected walking speed.

The primary outcome is the count of participants with potentially related or serious adverse events during the intervention. The posted counts are 1 in Sensory Augmentation and 0 in Random Vibration, with 22 participants at risk in each group. The sponsor’s Fisher Exact result is p = 1. The recalculated risk difference for Random Vibration relative to Sensory Augmentation is -0.045, with a 95% confidence interval from -0.133 to 0.042 and an unadjusted p-value of 0.312.

The low event counts and interval spanning both directions leave the comparative safety difference uncertain. The registry record does not report several details needed for fuller assessment, including missing-data procedures and the full statistical analysis plan.