Cerebral Palsy: Short-burst Interval Training: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Cerebral Palsy: Short-burst Interval Training (NCT03373890), a randomized na study. The primary outcome is Change in Walking Performance High Versus Low Frequency SBLTT Groups.

2. Clinical Question

The clinical question is therefore about the timing and frequency of a fixed course of treadmill training. Both groups were assigned the same total number of sessions; the schedule differed in how those sessions were distributed over time. The primary endpoint focuses on walking activity measured with StepWatch accelerometry, rather than on a laboratory measure of walking capacity alone.

The registry describes the study population broadly as ambulatory children with cerebral palsy and identifies spastic diplegia as the condition. It does not report the age range, how ambulatory status was defined, or whether other clinical characteristics were required for enrollment. The practical setting and any restrictions on usual therapy are also not reported in the registry record.

The comparison can address whether the two schedules were associated with different recorded changes in daily stride counts over the stated measurement interval. It cannot, by itself, isolate the effect of receiving treadmill training from the effect of receiving no training, because both randomized groups received SBLTT. Nor does the registered primary outcome establish whether any measured difference was meaningful to participants or persisted after the immediate post-training assessment.

3. Trial Design

The registry describes a completed, randomized, single-masked, parallel-group study sponsored by Seattle Children's Hospital. The condition is cerebral palsy spastic diplegia, and the brief description identifies the intended participants as ambulatory children with cerebral palsy. Enrollment is reported as 12. The registry lists the study start as 2014-08-01 and primary completion as 2016-07-31.

Participants were randomized to one of two schedules of short-burst interval locomotor treadmill training. The high-frequency group was scheduled for 20 sessions at 5x/week for 4 weeks. The low-frequency group was scheduled for 20 sessions at 2x/week for 10 weeks. The randomized contrast is thus between the frequency and duration of delivery, with the total number of training sessions held constant in the arm descriptions. The registry reports 6 participants in each arm for the primary outcome.

A parallel-group design assigns participants to a group and compares outcomes across those groups. In contrast to a crossover design, the registry does not describe participants switching from one schedule to the other. Random allocation can reduce systematic differences between groups on average, although it does not guarantee that every baseline characteristic is balanced, especially in a small study. Details about the randomization sequence and allocation concealment are not reported in the registry record.

The masking classification is single, but the registry does not state which parties were masked or how masking was maintained. It does not report a phase designation beyond NA. The registry lists 8 outcome measures and 8 analyses, including the primary walking outcome and secondary measures related to muscle performance and architecture. Additional information about the protocol, recruitment, adherence, and any changes to the planned design is 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: In eligible if unable to walk at greater than their self selected walking speeds with orthotics on a treadmill.

Overall Study

Analysis population for the primary outcome, as defined in the registry: Change in average strides/day from baseline to immediately post SBLTT were compared between the high and low frequency groups/arms.

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 Walking Performance High Versus Low Frequency SBLTT Groups, measured over Change from baseline to immediately post SBLTT.

The primary endpoint is change in average strides per day from baseline to immediately after SBLTT, compared between the high- and low-frequency groups. The registry reports the unit as strides per day and the measure as a mean of a continuous outcome. It describes the endpoint as average stride counts recorded by a StepWatch accelerometer.

For the recorded daily average, participants were to wear the StepWatch for a minimum of 8 hrs/day. Stride counts were averaged across 5 days: 4 weekdays and 1 weekend day. Including both weekdays and a weekend day is intended to capture walking activity across more than one type of day. The registry does not report details such as the number of valid monitoring days achieved by each participant or how incomplete wear time was handled.

A change score compares a participant's post-training value with their baseline value. Comparing these changes between randomized groups addresses whether the recorded shift over the training period differed by schedule. Daily stride counts provide a measure of activity in everyday walking, which complements outcomes focused on muscle performance or architecture. They do not directly measure every aspect of mobility, walking quality, or participation.

The registry names the immediate post-SBLTT assessment as the endpoint timeframe. It does not report later follow-up for this primary outcome. Consequently, the posted result concerns the change over the specified baseline-to-immediate-post period, not whether a difference persisted over a longer interval.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The registry identifies the allocation as randomized, the study model as parallel, and masking as single. Randomization assigns participants to groups using a chance-based process. Its purpose is to reduce systematic selection differences between the groups; it does not ensure identical groups, particularly when group sizes are small. The registry does not report how the random sequence was generated, whether assignment was concealed, or which participants or study staff were masked. Those details are not reported in the registry record.

What does the parallel comparison estimate?

In a parallel-group study, outcomes are compared between groups assigned to different strategies. Here, the strategies are schedules for the same SBLTT program: 20 sessions scheduled at 5x/week for 4 weeks or 20 sessions scheduled at 2x/week for 10 weeks. This design targets the contrast between those schedules among participants in the study. Because there is no group assigned to no training, the comparison cannot estimate the effect of training versus no training.

The arm descriptions state the intended schedules. They do not establish how many participants completed all sessions or whether any participants changed schedules. Adherence, deviations from assigned schedules, and how participants were analyzed after such deviations are not reported in the registry record. Those distinctions matter because the result may be affected by the difference between the assigned regimen and the regimen actually received.

What does the mean and standard deviation describe?

The primary endpoint is a continuous measure expressed in strides per day. A mean summarizes the average recorded change in a group. A standard deviation describes variation among individual observations around that mean; it is not the uncertainty of the mean itself and is not a confidence interval for a between-group difference. The registry reports means and standard deviations for each group, but does not report a confidence interval for the contrast.

Stride counts came from StepWatch accelerometry. The registry describes averaging counts across 5 days, made up of 4 weekdays and 1 weekend day, with a minimum of 8 hrs/day of wearing time. This approach aims to represent daily walking activity across different days. In general, wearable measures can be affected by wear time, device processing, and missing or invalid observations. How those issues were handled in this study is not reported in the registry record.

How should the Wilcoxon (Mann-Whitney) result be read?

The posted analysis names the Wilcoxon (Mann-Whitney) test and reports p .39. This rank-based method compares the relative ordering of observations in the groups. In general, it is useful when an analysis is based on ranks rather than relying on a model that directly compares means. Its interpretation is not automatically a test of equality of arithmetic means. The exact estimand can depend on assumptions about the shapes and spread of the group distributions.

A p-value describes how unusual a test result at least as extreme as the observed result would be if the test's null hypothesis and assumptions held. It is not the probability that the null hypothesis is true, nor is it the probability that the observed result arose by chance. A p-value alone does not show the size or clinical importance of a contrast. The registry does not report the test statistic, the test's detailed assumptions, or a confidence interval.

Why does uncertainty matter?

A confidence interval gives a range of values compatible with the data and the statistical procedure, under its assumptions. It communicates both the direction and precision of an estimate more directly than a p-value alone. A narrow interval indicates greater precision than a wide interval, while an interval spanning meaningfully different possibilities leaves more uncertainty about the contrast. No confidence interval for the primary group comparison is reported in the registry record.

Variation in the individual measurements, the number of observations, and the analysis method all affect precision. The reported standard deviations describe observed within-group spread, while the small number of participants in each group limits the information available to estimate the between-group contrast. The available summaries do not permit a full assessment of distribution shape, outliers, or individual changes.

What do planning illustrations mean?

Sample-size illustrations are planning calculations, not observations from the trial. The registry's stated assumptions describe an illustration based on a fraction of the observed effect, equal allocation, a two-sided significance level of 0.05, and a normal approximation. Such calculations can show how sample-size requirements depend on the effect one aims to detect and the desired power. They do not establish the trial's actual design rationale, and they are not observed power. The sample-size justification and the study's own planning calculations are not reported in the registry record.

In general, power is the chance that a specified analysis will detect an effect of a specified size under its assumptions. It is not a property that can be inferred reliably from the observed p-value after a study is complete. Planning should instead be judged from the assumptions and methods set in advance. The registry does not report the protocol-level assumptions needed to assess whether the enrollment target was chosen to detect a particular difference.

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("High Frequency 5x/Week", "Low Frequency 2x Week")
n   <- c(6, 6)
m   <- c(1872, 1551)
sp  <- c(880, 952)
sd  <- sp      # posted standard deviations
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  High Frequency 5x/Week  6  1872.000  880.000  948.497  2795.503
CARM  Low Frequency 2x Week  6  1551.000  952.000  551.937  2550.063

11. Analyses Reported by the Sponsor

The registry's posted analysis is a Wilcoxon (Mann-Whitney) comparison of the High Frequency 5x/Week and Low Frequency 2x Week groups. It reports p .39, but does not provide the test statistic, a confidence interval, or a numerical estimate of the between-group contrast. The group summaries are reported separately as means and standard deviations.

A separate recalculated estimate or confidence interval is not reported in the registry record, so there is no independent numerical contrast to set beside the sponsor's analysis. The posted p-value should be interpreted as the result of the named rank-based procedure, rather than as a direct test of the difference between the two arithmetic means. No discrepancy between the sponsor's reported result and a separate recalculation is identified in the registry record.

When a sponsor analysis and a recalculation differ, the sponsor's own analysis takes precedence for describing the trial's reported analysis, while any recalculation should be clearly identified as an unadjusted analysis of posted summary statistics. In this record, the information available does not support a separate numerical recalculation. The main interpretive limitation remains the absence of an interval estimate and a directly reported effect estimate for the group contrast.

12. Multiplicity, Missing Data and Analysis Populations

The registry lists 8 outcome measures and 8 analyses. The primary outcome is change in walking performance, measured as average strides per day. The listed secondary outcomes concern muscle performance—power and strength—and muscle architecture, including hypertrophy, in high- versus low-frequency SBLTT groups. The registry does not report results for those secondary outcomes in the registry record for interpretation.

When a study examines multiple outcomes or performs multiple comparisons, the chance of at least one apparently unusual result can increase if each test is interpreted separately without adjustment. Statistical methods can address this issue in different ways, but they may also change the balance between false-positive risk and sensitivity. The registry does not report a multiplicity adjustment, a hierarchy among secondary outcomes, or a plan for selecting which results would be treated as confirmatory. Those details are not reported in the registry record.

Missing-data handling is also important for change outcomes. Participants may lack a baseline value, an immediate post-training measurement, or sufficient valid accelerometer wear time. The registry describes a minimum wear-time criterion, but does not state how missing measurements, incomplete monitoring, or other invalid observations were handled in the analysis. It does not report whether imputation was used or how any imputation model was specified.

The analysis population is not defined in the reported information. The registry does not say whether the primary analysis included everyone as randomized, only participants who completed the training, or another population. Nor does it report how deviations from assigned schedules were addressed. These omissions make it difficult to assess how the posted result relates to the effect of assignment versus the effect of completing the intended schedules.

There is no basis in the registry record to infer that missingness or multiple testing did or did not materially alter the result. Clear reporting of analysis populations, missing-data rules, and multiplicity methods would help readers assess robustness and the strength of evidence across outcomes.

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 non-serious adverse event terms

The registry reports no serious adverse events in either group: the serious-event count is 0 among 6 at risk in the High Frequency 5x/Week group and 0 among 6 at risk in the Low Frequency 2x Week group. Deaths are also reported as 0 in both groups. These are the posted counts for the recorded safety categories; they should not be taken as proof that serious events cannot occur.

For other adverse events, the registry reports 2 among 6 at risk in the high-frequency group and 0 among 6 at risk in the low-frequency group. The registry does not provide event descriptions in the registry record, and the listed serious-event terms are empty. The data therefore do not support conclusions about the nature, timing, severity, or relationship of the other events to training.

Adverse-event counts in small groups are descriptive. A count of zero means no event was recorded in that group under the registry's reporting, not that risk is absent. The record does not report the adverse-event collection period, definitions, causality assessments, or whether all events were actively solicited. Safety comparisons are consequently limited and should be kept separate from the primary walking outcome.

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

The result addresses a focused statistical question: how the change in average daily stride counts differed between participants assigned to two schedules for the same treadmill-training program. The posted means are 1872.0 strides per day for the high-frequency group and 1551.0 strides per day for the low-frequency group, with a reported Wilcoxon (Mann-Whitney) p-value of .39. Those figures describe the posted outcome and analysis; they do not establish that one schedule is superior.

A p-value of .39 is not proof that the schedules have identical effects. It also does not establish that a clinically meaningful difference is absent. The registry does not report an interval estimate or a prespecified equivalence margin, so the range of plausible group contrasts cannot be evaluated from an interval. The small groups and the observed spread in changes also limit the precision that can be expected from the posted summaries.

The comparison does not evaluate training versus no training: both groups were assigned SBLTT. It is a comparison of frequency and duration of delivery, with 20 sessions specified for each schedule. A difference between the groups, if estimated precisely, would pertain to that scheduling contrast and the outcome as measured in this study.

The primary measure captures average daily stride activity using StepWatch accelerometry. It is not a complete measure of mobility, walking quality, independence, or participation. The result is tied to the baseline-to-immediate-post assessment; longer-term persistence is not reported. Although secondary outcomes are listed for muscle performance and architecture, no results for them are included in the information available for this analysis.

These findings should be read as a small randomized comparison with a limited primary endpoint and incomplete reporting of precision and analysis details. They describe what the registry reports, but they do not determine the clinical importance of the result, the response of an individual child, or the likely outcome in populations beyond the study participants.

15. Why This Trial Matters Statistically

This trial illustrates a distinction that is often important in clinical research: comparing two ways to deliver an intervention is not the same as comparing the intervention with no intervention. Both randomized groups were assigned short-burst interval treadmill training, with the same total of 20 sessions specified in the arm descriptions. The statistical target is therefore a schedule contrast—5x/week for 4 weeks versus 2x/week for 10 weeks—not an estimate of the effect of receiving training itself.

The outcome also demonstrates why the estimand and test method need to be read together. The registry presents mean changes and standard deviations in a continuous stride-count measure, while reporting a Wilcoxon (Mann-Whitney) test. A rank-based test does not directly test the difference between arithmetic means in the usual sense. Readers need to distinguish the descriptive summaries from the inferential comparison and avoid treating the p-value as an estimate of the mean contrast.

Precision is especially relevant in a small randomized study. Each group has 6 participants in the primary outcome record, and the standard deviations indicate variation in measured changes. Without a confidence interval, the uncertainty around the group contrast is difficult to communicate. A non-small p-value alone cannot establish equivalence or rule out differences that could matter.

The record also highlights reporting questions that shape interpretation beyond the headline test: which participants were analyzed, how missing accelerometer data were handled, whether multiple outcomes were adjusted, and which parties were masked. These details are not reported in the registry record. Clear reporting of those choices helps readers distinguish the effect of assignment from the effect of adherence and understand how robust the findings may be.

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 enrollment is small, and the primary outcome record reports 6 participants in each group. This limits the information available for estimating between-group differences and makes the result sensitive to variation among participants. The registry reports standard deviations but no confidence interval for the contrast, so its precision is not directly shown.

The comparison has no no-training group. It can inform the contrast between the two stated schedules, but not the effect of SBLTT relative to usual care or no treadmill training. The registry describes the intended session schedules, but adherence and treatment deviations are not reported in the information available for this interpretation.

Population details are limited. The condition is listed as cerebral palsy spastic diplegia, and the brief description identifies ambulatory children, but age range, severity, eligibility criteria, baseline clinical characteristics, and recruitment setting are not reported in the registry record. Generalizability beyond the enrolled population is therefore difficult to assess.

The primary endpoint is an accelerometer-based average over 5 days, with a stated minimum of 8 hrs/day of wear. The registry does not describe the amount of valid wear achieved, handling of incomplete data, device processing, or the analysis population. It also does not report a confidence interval, the Wilcoxon test statistic, or a detailed statistical analysis plan.

Masking is classified as single, but the masked party and procedures are not specified. Randomization-sequence generation and allocation concealment are likewise not reported. The registry lists secondary measures and analyses but does not report a multiplicity strategy or the secondary results in the registry record for interpretation. Finally, the primary timeframe is immediately after training; longer-term outcome data are not reported.

17. Related Tutorials

18. Related Statistical Calculators

19. Sources

• ClinicalTrials.gov: Cerebral Palsy: Short-burst Interval Training — NCT03373890.
• Registry record last updated 2019-07-10; status last verified 2019-06.
• Results first posted 2019-07-10.
• 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

NCT03373890 is a completed randomized parallel-group study of short-burst interval locomotor treadmill training in ambulatory children with cerebral palsy; the registry lists cerebral palsy spastic diplegia as the condition. The study compared 20 sessions scheduled at 5x/week for 4 weeks with 20 sessions scheduled at 2x/week for 10 weeks. Enrollment is reported as 12, with 6 participants in each group in the primary outcome record.

The primary outcome was change in average strides per day from baseline to immediately after SBLTT, measured by StepWatch accelerometry. The posted group means were 1872.0 and 1551.0 strides per day, with standard deviations of 880.0 and 952.0. The reported Wilcoxon (Mann-Whitney) result was p .39. No confidence interval or between-group effect estimate is reported in the registry record.

The registry lists secondary outcomes related to muscle performance and architecture. Details about missing-data handling, analysis populations, multiplicity, and the identity of the masked party are not reported in the information available for interpretation.