Tracking Post-stroke Walking Improvements From the Clinic Into the Home: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Tracking Post-stroke Walking Improvements From the Clinic Into the Home (NCT05454007), a non-randomized na study. The primary outcome is Step Length Asymmetry.

2. Clinical Question

3. Trial Design

The registry identifies NCT05454007 as a completed, non-randomized study with a single-group model and no masking. Although the record describes participants with stroke and neurologically intact adults, it explicitly cautions that the arms are not a like-for-like control comparison. Group assignment was not randomized, and the healthy-adult group did not receive the intervention.

The study had two aims. In Aim 1, people with stroke and neurologically intact adults underwent screening and walking trials in clinic and at home. In Aim 2, people with stroke underwent a single session of step length symmetry training. The registry describes the overall purpose as using video-based pose estimation to assess gait in natural home settings and to examine how clinically derived walking improvements change in the home.

The listed enrollment is 53. The study dates in the record run from 2022-08-25 to primary completion on 2025-03-09, and the status is completed. These administrative details do not establish how many participants contributed usable data to each outcome. The registry notes that data from three individuals with stroke and three individuals without stroke were excluded from analysis of the primary outcome because of insufficient video or pose-estimation quality.

The record lists six outcome measures and no analysis entries. It does not report the statistical analysis plan, allocation procedures beyond non-randomized status, or detailed intervention procedures. It also does not provide numerical results for the primary outcome. Thus the design and aims can be described, but the posted record does not support an estimate of the observed difference between settings or a conclusion about change after training.

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

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: Data from three individuals with stroke and three individuals without stroke were excluded from analysis due to insufficient data quality. Issues included pixelated or blurry videos, failure to keep the recording device stationary, or inaccurate identification of key landmarks during pose estimation. Ultimately, the data from 43 participants were included in data analysis.

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: Data from three individuals with stroke and three individuals without stroke were excluded from analysis due to insufficient data quality. Issues included pixelated or blurry videos, failure to keep the recording device stationary, or inaccurate identification of key landmarks during pose estimation. Ultimately, the data from 43 participants were included in data analysis.

6. Primary Endpoint

The registry lists the primary outcome as Step Length Asymmetry, measured over In-clinic baseline (day 1), in-home baseline (day 1).

The primary endpoint is step length asymmetry, described as the normalized absolute difference between short and long step length, based on left and right step lengths. It is unitless. The registry specifies a mean and standard deviation as the planned summary and identifies in-clinic baseline and in-home baseline as the timeframe. The endpoint description says the measure will be compared across timepoints, but numerical endpoint results are not reported in the registry record.

Step length asymmetry is used to represent a spatial feature of walking: whether steps on one side differ in length from those on the other. A normalized measure expresses that contrast relative to step length rather than in a physical distance unit. In general, such a measure can help compare gait patterns recorded under different conditions, though its interpretation depends on the precise calculation and the quality of the underlying measurements.

The study’s motivation is to assess whether clinic-based walking reflects walking in the home. The endpoint is relevant to that question because the clinic and home are distinct measurement settings. However, the primary endpoint alone does not establish why any difference occurs, whether a change is clinically important, or whether training caused a change. The registry record does not report a threshold for meaningful change, an estimated setting contrast, or uncertainty around an estimate.

Video quality is particularly relevant to this outcome because step lengths are derived from recorded movement and pose estimation. The registry reports exclusions for pixelated or blurry videos, movement of the recording device, and inaccurate identification of key landmarks. Those issues can affect whether the measurement is usable and may also limit how broadly the analyzed observations represent all participants.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The statistical methods that can be identified from the registry are limited. The primary outcome is continuous and is summarized using a mean and standard deviation. The study is non-randomized, has no masking, and is described as a single-group model. The registry does not report a detailed statistical analysis plan, model specification, handling of repeated measurements, or inferential results. The explanations below describe general statistical concepts rather than methods confirmed to have been used in this study.

What does a mean describe?

For a continuous measure such as step length asymmetry, the mean is a summary of the central level of observations. It can be useful for describing the average measured asymmetry in a group or setting. A mean does not show the full distribution: individuals may have similar or very different measurements even when the average is the same. It can also be influenced by unusually high or low observations. The registry record does not provide the actual primary-outcome means.

What does the standard deviation describe?

In general, the standard deviation describes the spread of observations around their mean. A larger standard deviation indicates more variability among measured values, while a smaller one indicates that observations tend to be closer to the mean. It is not the same as the uncertainty in the estimated mean or in a difference between means. That uncertainty also depends on the amount of usable information and the analysis method. The registry identifies standard deviation as the dispersion measure but does not report its outcome values.

What would a confidence interval add?

In general, a confidence interval communicates uncertainty around an estimate. For a clinic-to-home comparison, an interval around the estimated contrast would help readers judge which effect sizes are compatible with the observed data under the analysis assumptions. An interval that is wide would indicate limited precision; a narrower interval would indicate greater precision. A confidence interval is not a guarantee that the true value lies within the interval. No primary-outcome confidence interval is reported in the registry record.

Why does non-randomization matter?

Randomization can help balance measured and unmeasured characteristics between comparison groups, on average. This study is explicitly non-randomized, and the registry warns that its arms are not a like-for-like control comparison. The clinic and home are measurement settings, and the neurologically intact adults participated in Aim 1 but not Aim 2. Accordingly, observed differences between settings or participant groups could reflect factors other than the contrast of interest. Statistical adjustment can address measured covariates when appropriate data and a suitable plan exist, but it cannot automatically remove all confounding. The registry record does not report an adjustment strategy.

Why can repeated measurements need special handling?

In general, measurements from the same participant are related rather than independent. When a study compares observations from the same person across settings or timepoints, an analysis may need to account for that within-person relationship. Ignoring it can lead to inappropriate estimates of uncertainty. The endpoint description mentions comparisons across timepoints, but the registry record does not specify whether the primary analysis used paired comparisons, a repeated-measures model, or another approach. No model details are reported, so a particular method should not be attributed to the study.

Why is measurement quality part of statistical interpretation?

Video-based pose estimation turns recorded movement into landmarks and gait measures. If image quality is poor, the recording device moves, or landmarks are identified inaccurately, the resulting measure may be unreliable or unavailable. The registry reports exclusions for those problems. In general, measurement error can add variability and can sometimes shift estimates, depending on how it occurs. Excluding unusable recordings avoids treating invalid measurements as reliable, but it can also affect which observations are represented. The record does not report a validation analysis or sensitivity analysis for these exclusions.

What does over-dispersion mean in general?

Over-dispersion is a general concern for count outcomes when their variability exceeds what a statistical model assumes. The primary endpoint in this record is continuous, not a count, so that issue is not the central method for interpreting step length asymmetry. More generally, the choice of model should match the outcome scale, its distribution, and the study design. The registry does not describe the modeling methods used for the other listed outcomes.

The registry includes planning assumptions for sample-size illustrations: a two-sided significance level of 0.05, equal allocation between two arms, a normal approximation, and effects expressed as fractions of the observed effect. These are described as planning illustrations rather than the study’s own design. Such illustrations are not observed power and do not substitute for outcome estimates or uncertainty intervals. No numerical planning results or protocol sample-size justification are reported in the registry record.

10. Multiplicity, Missing Data and Analysis Populations

The registry lists six outcome measures, but it does not report a multiplicity strategy. When a study evaluates several outcomes, timepoints, or comparisons, the chance of at least one apparently notable result can increase if each is tested separately. General approaches can include identifying a primary endpoint in advance, adjusting for multiple comparisons, or interpreting secondary outcomes as exploratory. The record identifies step length asymmetry as the primary outcome, but does not describe any adjustment or hierarchy for the other listed measures.

The registry reports that data from three individuals with stroke and three individuals without stroke were excluded from primary-outcome analysis because of insufficient data quality. The stated problems included blurry or pixelated video, a recording device that was not kept stationary, and inaccurate landmark identification during pose estimation. The record does not report whether missing or unusable measurements were imputed, whether analyses were limited to complete cases, or whether the results were tested under alternative assumptions about missing data.

The analysis populations are not fully described. The record gives an enrollment of 53 and identifies the outcome-specific exclusions, but does not report the number of participants contributing usable measurements to each analysis. It does not specify whether all eligible participants were included in an intended analysis population or how protocol deviations, incomplete observations, or repeated measurements were handled.

The registry lists no analysis entries and does not report numerical primary-outcome results. It also does not report interim analyses or the statistical plan for handling missing data and multiplicity. These omissions limit assessment of whether the reported analyses, if available elsewhere, followed prespecified methods. They also prevent judging how robust an estimate would be to missing recordings or to the choice of analysis population.

11. 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 serious adverse event counts of zero for Individuals With Stroke and zero for Neurologically Intact Adults. It also reports other adverse event counts of zero for both groups, and deaths of zero for both groups. The listed numbers at risk are 21 for Individuals With Stroke and 29 for Neurologically Intact Adults.

These figures describe what the registry record reports; they do not establish that no adverse events occurred outside the record’s collection or reporting process. A zero count means that no events were entered in the corresponding category in the posted results, not that the risk is absent. The record does not describe the adverse-event collection method, surveillance period, event definitions, or completeness of reporting.

The groups also had different roles in the study: neurologically intact adults did not receive the intervention, whereas participants with stroke underwent a single session of step length symmetry training in Aim 2. The counts should therefore not be read as a randomized safety comparison or as evidence comparing intervention exposure. The registry does not provide event narratives, severity details beyond the serious-event category, or a statistical comparison of safety outcomes.

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

The registry record establishes that this completed study was designed to examine video-based gait measurement in clinic and home settings and to investigate post-stroke walking. It identifies step length asymmetry as the primary outcome, describes the measurement as unitless, and specifies in-clinic and in-home baseline as the timeframe. It also documents data-quality exclusions and reports adverse-event counts.

It does not report numerical results for the primary outcome. Readers therefore cannot determine from the record the average asymmetry in either setting, the size or direction of a setting contrast, or the precision of any estimate. The record also does not provide a p-value or confidence interval for that outcome. Without those statistics, it is not possible to assess whether the data show a clear pattern or whether the estimated differences are uncertain.

The study’s non-randomized structure also matters. Clinic and home observations can reveal whether measurement differs across settings, but a difference would not automatically show that the setting itself caused the difference. Likewise, the healthy-adult participants in Aim 1 are not a randomized control group for participants with stroke, and they did not receive training. The study design does not support attributing differences between these groups to the intervention.

Step length asymmetry is a measure of walking pattern, not a complete account of mobility or rehabilitation outcomes. The registry does not report a threshold for clinically important change, and a numerical difference alone would not establish that a change matters to participants. The record also does not report whether changes after training were sustained or whether they translated into broader everyday function.

The most defensible conclusion is narrow: the registry describes a study intended to assess gait in natural home environments and compare observations across settings, but its posted record does not include the primary outcome estimates needed to determine what the data show. The record’s completion status and enrollment figure do not fill that results gap.

13. Why This Trial Matters Statistically

This study addresses a statistical and measurement problem in rehabilitation research: a measure collected under controlled clinic conditions may not represent the same behavior in a person’s home. The distinction between capacity observed in a clinic and performance in everyday surroundings is central to the study’s stated rationale. Video-based pose estimation offers a way to observe walking outside a laboratory, but its value depends on whether recordings can be made and measured consistently.

The registry’s focus on both clinic and home trials makes measurement context part of the scientific question rather than merely a logistical detail. In general, an estimate of a clinical effect is only as informative as the relationship between the measurement conditions and the real-world behavior it is intended to represent. A setting contrast can therefore matter even when it is not a treatment comparison.

The study also illustrates why data quality should be considered alongside statistical results. The record identifies exclusions tied to video clarity, camera stability, and landmark identification. These practical issues can determine whether home-based measurement is feasible and which observations can be analyzed. They may also affect whether results apply to participants whose recordings are less suitable for analysis.

The design provides useful context for interpreting comparisons: the study is non-randomized, and the healthy-adult group did not participate in the training aim. That distinction discourages treating every group contrast as a causal effect. Good statistical interpretation depends not only on estimates and tests, but also on matching conclusions to the design and measurement process.

The registry does not report the outcome estimates, so the study’s stated statistical contribution cannot be judged from the posted results alone. Still, its aims highlight an important methodological question: whether clinic-derived measures correspond to walking in the home, and whether video workflows can support such assessment. Answering that question requires transparent reporting of measurement quality, analysis populations, estimates, and uncertainty.

14. 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 consequential limitation of the posted record is that it does not report numerical results for the primary outcome. The mean and standard deviation are identified as summary measures, but their observed values are absent, as are a comparison estimate and its uncertainty. The study’s central clinic-to-home question therefore cannot be evaluated quantitatively from this record.

There are also limitations in the available design and methods information. The study is non-randomized and unmasked, and the registry explicitly cautions that the arms are not a like-for-like control comparison. The record does not report how observations across settings or timepoints were modeled, whether estimates were adjusted, or how repeated measurements were handled. Those details would be important for assessing the validity and precision of comparisons.

Measurement quality affected inclusion in the primary-outcome analysis. The registry reports exclusions for unclear video, movement of the recording device, or inaccurate pose-estimation landmarks. It does not report a comparison of included and excluded participants or sensitivity analyses addressing these exclusions. If the ability to obtain usable video is related to participant or setting characteristics, analyzed observations may not fully represent all enrolled participants.

The record gives limited detail about participant characteristics, eligibility, intervention procedures, outcome timing beyond the primary timeframe, and the analysis population. It lists six outcomes but does not describe a multiplicity plan or missing-data strategy. No analysis entries are listed. These reporting gaps limit reproducibility and make it difficult to distinguish prespecified analyses from exploratory ones.

Finally, the study’s intended real-world focus is specific to the described population and video workflow. The registry does not report evidence about performance across other settings, equipment, or populations. These limitations do not establish that the approach is unsuitable; they define what the posted record can and cannot support.

15. Related Tutorials

16. Related Statistical Calculators

17. Sources

• ClinicalTrials.gov: Tracking Post-stroke Walking Improvements From the Clinic Into the Home — NCT05454007.
• Registry record last updated 2026-07-24; status last verified 2026-06.
• Results first posted 2026-07-24.
• 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.

18. Record Summary

NCT05454007, “Tracking Post-stroke Walking Improvements From the Clinic Into the Home,” is listed as completed, with enrollment of 53 and a single-group, non-randomized design without masking. The study included people with stroke in Aim 1 and Aim 2, while neurologically intact adults participated in Aim 1 only. The intervention was step length asymmetry training, delivered as a single session in Aim 2 for participants with stroke.

The primary outcome was unitless step length asymmetry, summarized as a mean with standard deviation and assessed at in-clinic and in-home baseline. The registry reports primary-outcome exclusions for video or pose-estimation quality problems. It lists zero serious adverse events, zero other adverse events, and zero deaths in both reported groups.

Numerical primary-outcome results, a statistical analysis plan, and analysis entries are not reported in the registry record. The design and aims are documented, but the posted information does not quantify the clinic-to-home contrast or its uncertainty.