Precision Versus Endurance Training to Improve Walking After Chronic Incomplete SCI: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Precision Versus Endurance Training to Improve Walking After Chronic Incomplete SCI (NCT01765153), a randomized na study. The primary outcome is Change in Spinal Cord Injury - Functional Ambulation Profile (SCI-FAP) at Completion of Study.

2. Clinical Question

3. Trial Design

The registry describes a randomized, single-blind crossover trial with 21 participants enrolled. Participants were randomly allocated to begin with either Precision Training or Endurance Training. These starting sequences were called Training Phase I; after the first training and rest periods, participants trained using the other method in Training Phase II.

Each training phase lasted 2 months, with training 5x/wk, and was followed by a 2-month rest period. After the first rest period, participants returned for the second training phase, then completed another 2-month rest period. The primary endpoint was assessed from Baseline I to the end of Rest Period II, described as approximately 8 months after the study began. The registry says walking speed, distance, skill, confidence, and depression measures were obtained at least 3 times before training and monthly thereafter. Electrophysiological measures were taken before and after each training phase and rest period.

In the posted primary-outcome results, Endurance First has n 8 and Precision First has n 8. Those labels identify randomized order groups, not groups that received only one training method: each sequence included both interventions. The registry also reports 21 enrolled, but it does not explain in the registry record why the posted primary endpoint has results for fewer participants than the enrollment figure.

Single-blind status is recorded, but the registry does not report who was masked or how masking was maintained. The record identifies one analysis and 49 outcome measures, including the primary outcome and multiple secondary outcome time points. Further details of allocation concealment, adherence, protocol deviations, and the analysis plan 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.

Training Phase 1

Reasons for not completing this period:

Rest Phase 1

Reasons for not completing this period:

Training Phase 2

Reasons for not completing this period:

Rest Phase 2

Reasons for not completing this period:

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 Spinal Cord Injury - Functional Ambulation Profile (SCI-FAP) at Completion of Study, measured over Change from Baseline I to the end of the study (i.e., end of Rest Period II, which was ~8 months after the beginning of the study).

The primary endpoint is change in the Spinal Cord Injury Functional Ambulation Profile (SCI-FAP) from Baseline I to the end of the study, at the end of Rest Period II. The registry describes this as approximately 8 months after the beginning of the study. The SCI-FAP is a timed walking test covering 7 tasks. For each task, time to completion is combined with an assistance factor based on walking aids; the item scores are summed into one composite score.

The scale ranges from a minimum score of 7, representing all tasks performed without aids at the mean able-bodied time, to a maximum of 2100, representing inability to complete any task within the allowed time. The reported outcome is final score minus initial score, so a negative change indicates improvement. Both posted order-group means are negative, which means that their average changes are in the direction of improvement on this score.

A composite measure can capture performance across several kinds of walking tasks rather than relying on a single timed task. The registry does not report a separate component-level result for each task in the primary outcome. It also does not report a primary-outcome estimate directly comparing the training methods.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

What does the mean describe?

The primary outcome is continuous, and the registry reports the mean for each order group. A mean gives the arithmetic average of the recorded change scores within that group. Here, each score is final SCI-FAP minus initial SCI-FAP. Since lower SCI-FAP scores represent better task performance and the registry defines negative change as improvement, a negative mean change indicates that the group average moved in the improvement direction. It does not mean that every participant improved; individual results are not reported.

What does the standard deviation describe?

The standard deviation describes how dispersed the observed scores are around their group mean. It is useful for understanding variability among participants. It is not the same as a standard error or a confidence interval. A confidence interval is an interval estimate intended to express uncertainty about an estimated quantity, such as a mean or treatment contrast, under a stated statistical model. The registry record does not report confidence intervals for these results.

Why does the crossover structure matter?

In a crossover design, participants receive both interventions in a planned sequence. Each participant can therefore contribute information under both treatment conditions, and the comparison can account for the fact that one person’s walking performance may differ from another’s. A simple comparison of independent group means is not the natural way to estimate a within-person treatment contrast. The sequence groups reported for the primary outcome are Endurance First and Precision First, and the endpoint is measured after both training methods have been completed. These end-of-study group means therefore combine the two training periods, sequence, and intervening rest periods.

Crossover analyses may need to consider treatment, period, sequence, and possible carryover from an earlier period. Carryover means that an earlier treatment may continue to influence an outcome measured in a later period. The protocol includes 2-month rest periods, but the adequacy of those periods for eliminating carryover is not reported in the registry record. The statistical model used to handle period, sequence, within-participant correlation, or carryover is also not reported in the registry record. These are general design and analysis considerations, not claims about what the investigators did.

What can be inferred from the posted summaries?

The means and standard deviations describe the observed change-score summaries for the two randomized order groups. They do not provide the within-participant difference between Endurance and Precision outcomes. To interpret a treatment contrast, readers would need the relevant treatment-period estimates or paired differences, an estimate of uncertainty, and the analysis method. These are not reported in the registry record. No inferential test for the primary comparison is included in the posted results. Accordingly, no conclusion about statistical evidence for a difference between the training methods can be drawn from these group summaries alone.

The registry lists one analysis and identifies the primary outcome as a mean with standard deviation. That information does not specify whether the model used repeated measurements, adjusted for baseline, or included other covariates. The statistical analysis plan, including the handling of incomplete observations and protocol deviations, is not reported in the registry record. Any such approach would be an assumption rather than a documented feature of this analysis.

How should planning calculations be understood?

General sample-size planning asks how many participants may be needed to detect a specified effect under chosen assumptions. The registry’s planning description states a two-sided significance level of 0.05, equal allocation, and a normal approximation, with the effect set as a stated fraction of the observed effect. Such calculations are planning illustrations based on the observed effect; they are not evidence that the trial was designed with those assumptions, and they are not observed power. The trial’s own sample-size justification is not reported in the registry record. The record also does not describe an adjustment to the planning method for the crossover structure.

A planning result is conditional on its assumptions, including the effect size and variability used. In general, a smaller target effect requires more information to detect than a larger target effect, all else being equal. This is a general statistical principle, not a finding about this study. Any planning illustration should therefore be interpreted as a hypothetical design aid rather than as a result of the trial or a measure of the strength of its findings.

Why can over-dispersion matter?

In general, greater dispersion in outcome scores means that observations vary more widely around their average. If a statistical model assumes less variation than is present, uncertainty can be understated and inferences may be too confident. Conversely, accounting for substantial variability can widen uncertainty intervals and reduce precision. The reported standard deviations show that participant results varied, but the registry does not report a model-based variance assessment, an interval estimate, or diagnostics. No conclusion about over-dispersion or the adequacy of a particular model can be made from the posted summaries alone.

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("Endurance First", "Precision First")
n   <- c(8, 8)
m   <- c(-128.77, -64.07)
sp  <- c(180.73, 77.24)
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  Endurance First  8  -128.770  180.730  -279.864  22.324
CARM  Precision First  8  -64.070  77.240  -128.644  0.504

11. Multiplicity, Missing Data and Analysis Populations

The registry lists 49 outcome measures, with 48 secondary outcomes posted, and identifies one analysis. The secondary outcomes include changes in SCI-FAP and the 6 Minute Walk Test at the ends of both training phases and both rest periods. Multiple outcomes and time points create opportunities for many separate comparisons. In general, when many hypotheses are tested, the chance of at least one apparently notable result arising by chance can increase unless the analysis plan addresses multiplicity. The registry record does not report a multiplicity adjustment, a hierarchy of tests, or which secondary outcomes were designated as confirmatory.

The primary endpoint is described as the change from Baseline I to the end of Rest Period II. Although the registry describes repeated measurement of several outcomes, the analysis method for those repeated observations is not reported in the registry record. It is not possible to tell from the posted primary summaries whether baseline adjustment, repeated-measures modelling, or a within-person crossover analysis was used.

The record reports 21 enrolled, while the primary-outcome results have n 8 in each order group. It does not state why the endpoint counts are smaller than the enrollment figure, how many participants completed each phase, or how missing observations were handled. The analysis population—such as all randomized participants or participants with complete outcome data—is not reported in the registry record. The handling of withdrawals, protocol deviations, and missing data is also not reported in the registry record.

These omissions matter because missingness can affect interpretation if the participants with observed results differ systematically from those without results. That is a general concern, not evidence that such a difference occurred in this trial. The posted summaries do not permit assessment of the reasons for missingness or the robustness of the result to alternative assumptions.

12. Safety Results

The registry reports serious adverse events as affected participants divided by participants at risk.

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 0 among 10 participants at risk in the Endurance First safety group and 0 among 11 participants at risk in the Precision First safety group. It also reports 0 other adverse events among 10 and 11 participants at risk, respectively. These are counts within the safety groups as recorded; they do not describe the severity or types of events because no event terms are listed.

The safety denominators differ from the primary-outcome counts, which are n 8 in each sequence group. The registry does not explain the difference between these counts or specify the safety observation period in the posted results. Death counts are not reported in the registry record. The absence of recorded serious or other adverse events in these groups should be read as a summary of the posted counts, not as proof that no events occurred outside the registry’s reporting scope.

Zero reported events in a group does not establish that the underlying risk is zero. With a small number of participants, uncommon events may not be observed, and the record does not provide uncertainty intervals for the event counts. These data describe reported safety events; they do not support broader conclusions about safety.

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

The posted primary results show that the mean SCI-FAP change was negative in both randomized order groups: -128.8 for Endurance First and -64.07 for Precision First. By the registry’s definition, negative change is in the direction of improvement. The standard deviations, 180.7 and 77.24 respectively, describe variability in those group scores. These summaries describe change from the initial baseline to the end of the study, after participants had been assigned to a sequence involving both forms of training.

They do not establish the comparative effect of Precision Training versus Endurance Training. The endpoint is not a direct contrast between the interventions, and the registry does not report a treatment-effect estimate, confidence interval, or p-value for that comparison. No calculation of the difference between the group means can substitute for the appropriate crossover contrast, particularly when the groups represent treatment order and the outcome was measured after both treatments.

The means also do not show how many individual participants improved, remained unchanged, or worsened. A group average can be negative even when individual responses vary. The posted standard deviations indicate dispersion but do not provide participant-level patterns or the uncertainty of a treatment comparison.

In general, a confidence interval would help show the precision of an estimated effect by giving a range of values compatible with the data and model. No such interval for the comparative treatment effect is reported in the registry record. The available summaries therefore support a limited descriptive reading: average change in both sequence groups was in the direction the registry defines as improvement, while the size and precision of any difference attributable to training method remain unreported.

The study’s stated hypothesis concerned the potential value of visually guided training, and the randomized crossover design was intended to compare two approaches. The posted end-of-study summaries alone do not resolve that hypothesis. The record does not report enough comparative analysis detail to determine whether the observed changes differed by training method, how much uncertainty surrounded any difference, or whether the analysis accounted for the crossover structure.

14. Why This Trial Matters Statistically

This trial illustrates why design labels and outcome summaries must be interpreted together. A randomized crossover study assigns participants to an order of treatments, and each participant is intended to experience both approaches. The resulting treatment comparison is different from comparing two independent groups that each receive only one intervention. Here, the posted primary outcome is a baseline-to-end-of-study change for the randomized sequence groups, so it summarizes the full study experience rather than isolating each treatment period.

The example also shows the importance of identifying the estimand—the precise quantity a statistical analysis is intended to estimate. A mean change within each sequence group is a descriptive quantity. A treatment contrast in a crossover design would instead relate outcomes under Precision Training to outcomes under Endurance Training, accounting for the repeated observations and the order in which treatments were received. The registry does not report that contrast.

Precision is another key issue. Means and standard deviations convey information about central tendency and variability, but they do not replace an uncertainty interval around the treatment effect. With n 8 in each posted primary-outcome group, the amount of available information is limited, and the outcome-specific counts are below the reported enrollment of 21. The reasons for the difference are not reported in the registry record.

Finally, the registry lists many outcome measures and several time points. In general, this makes a prespecified analysis plan and transparent reporting of multiplicity important. The registry record does not describe those procedures. The trial is therefore a useful statistical example of how randomized allocation and repeated measurement do not, by themselves, make a treatment comparison interpretable without a clearly reported contrast, uncertainty estimate, analysis population, and handling of missing data.

15. 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 primary-outcome summaries are reported by treatment order rather than as a direct contrast between Precision and Endurance Training. Because both interventions were part of each assigned sequence, the end-of-study mean change cannot isolate the contribution of either method. A within-participant treatment comparison, including its uncertainty, is not reported in the registry record.

The outcome-specific results include n 8 in each sequence group, while overall enrollment is 21. The record does not explain the difference or report participant flow through the training phases and rest periods. Reasons for missing primary outcome data, completion rates, and the analysis population are not reported in the registry record. Without those details, the potential effect of incomplete follow-up cannot be assessed.

The registry reports a single-blind design but does not identify who was masked. It does not report allocation concealment, adherence, protocol deviations, or the statistical model used. In a crossover design, treatment period, sequence, within-person correlation, and possible carryover are general considerations. The record’s 2-month rest periods are described, but whether they were sufficient to address carryover is not reported in the registry record.

The outcome-measure listing includes multiple measures and time points, but the record does not report a multiplicity strategy or a hierarchy of confirmatory outcomes. Confidence intervals and a p-value for the primary treatment comparison are also absent. These limitations restrict interpretation to the posted descriptive summaries and prevent assessment of the precision and robustness of a comparative treatment effect.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: Precision Versus Endurance Training to Improve Walking After Chronic Incomplete SCI — NCT01765153.
• Registry record last updated 2017-05-02; status last verified 2017-01.
• Results first posted 2017-05-02.
• 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.

19. Record Summary

NCT01765153 is a completed, randomized, single-blind crossover trial sponsored by the University of Alberta. It enrolled 21 participants with chronic motor-incomplete spinal cord injury and compared two physical therapy approaches: Precision Training, emphasizing visually guided over-ground walking, and Endurance Training, emphasizing treadmill walking.

Participants were assigned to start with one method, complete a 2-month training phase and 2-month rest period, then receive the other method in a second training phase followed by another rest period. The primary outcome was change in SCI-FAP from Baseline I to the end of Rest Period II. The posted sequence-group means were -128.8 for Endurance First and -64.07 for Precision First, with standard deviations of 180.7 and 77.24, and n 8 in each group. Negative change is defined as improvement.

The registry does not report a direct treatment comparison, uncertainty interval, p-value, analysis population, or explanation for the difference between enrollment and primary-outcome counts. It reports zero serious and zero other adverse events in the listed safety groups; death counts are not reported in the registry record.