This page provides an independent statistical analysis and educational interpretation of publicly reported results. ClinicalTrials.gov provides the official trial registry record. Numerical trial results on this page are restricted to the results posted in the ClinicalTrials.gov record; calculations added by Clinical Biostats are labelled as such.
1. Trial at a Glance
This page covers the registry results for The Impact of Yoga-based Physical Therapy for Individuals With Traumatic Brain Injury (NCT03701594), a randomized na study. The primary outcome is Heart Rate Variability.
| Feature | Registry value |
|---|---|
| ClinicalTrials.gov identifier | NCT03701594 |
| Phase | NA |
| Condition | Traumatic Brain Injury |
| Allocation | Randomized |
| Design model | Crossover |
| Masking | Single |
| Enrollment | 13 |
| Arms in the primary outcome | 3 |
| Lead sponsor | Shirley Ryan AbilityLab |
| Results posted | Yes |
2. Clinical Question
Population
People with traumatic brain injury receiving inpatient care were the intended population. The registry describes rolling enrollment as people were admitted at Shirley Ryan AbilityLab. It records enrollment of 13, while the brief summary describes a plan to enroll up to 30 subjects. The posted heart rate variability summaries give an n of 11 for each condition; the registry does not explain the difference between the enrollment figure and the condition-level counts.
Intervention
The intervention was a one-hour yoga-based physical therapy session in a group setting. The description specifies a quiet, enclosed space, dimmed lights, and calming instrumental music. The study's central question was whether heart rate variability changed during this session. The registry also describes anxiety, fatigue, agitation, and sleep as outcomes of interest, although the posted secondary outcomes identify anxiety and fatigue.
Comparator
Each participant was to take part in three conditions in random order across three days: yoga-based physical therapy, one-on-one conventional physical therapy, and group seated rest in a relaxing environment. The conventional session was led by a different physical therapist from the yoga session. Seated rest was designed to take place in the same quiet environment as the yoga condition. This crossover structure allows each participant to experience each condition, while the posted analysis compares all three groups or conditions.
Primary question
The primary question was whether heart rate variability, measured in milliseconds with chest-worn MC-10 sensors, differed during the yoga-based session compared with the two other conditions. The registry's hypothesis anticipated improvement after yoga in heart rate variability and several secondary measures. The posted primary analysis reports a mixed models p-value of 0.598 for the three conditions, without a model-based effect estimate or confidence interval. It therefore does not quantify the size or precision of a specific pairwise difference.
3. Trial Design
This was a randomized, single-masked, crossover pilot study, listed as phase NA. The condition was traumatic brain injury, and the sponsor was Shirley Ryan AbilityLab. The record lists a start date of 2018-10-22, primary completion on 2019-04-01, and status COMPLETED. The recorded enrollment is 13. Separately, the study description says it would enroll up to 30 inpatient subjects on a rolling basis over a 12 month period. The record does not explain how the enrollment figure relates to the number contributing data to the posted primary summaries.
Participants were to complete three conditions in a random order across three days. These were one hour of yoga-based physical therapy in a group, one hour of one-on-one conventional physical therapy, and one hour of seated rest in a relaxing group setting. The rest condition was intended to reproduce the yoga session's environment. Conventional therapy was led by a different therapist from the yoga session; the description says there were no restrictions on what could occur during conventional therapy. The record identifies masking as single but does not report who was masked or how masking was implemented.
The crossover approach means that participants were intended to experience all conditions, rather than being assigned to only one condition. In general, this can allow comparisons that account for stable differences between individuals. It also makes the handling of within-person data and the order of conditions relevant. The registry reports use of mixed models for the primary analysis, but the detailed model specification, order effects, period effects, and any washout procedure are not reported in the registry record. The endpoint description says sensors were worn throughout the three one-hour conditions.
The study is described as a pilot intended to explore the immediate response to a yoga-based physical therapy session and inform possible larger-scale research on regular participation as an adjunct to traditional physical therapy. This record concerns the described session-based comparison; it does not report a result for a longer-term program.
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: Two participants dropped out of the study after randomization but before they completed any of the interventions.
Overall Study
| Stage | Yoga-based Physical Therapy Group, Then Conventional Physical Therapy, Then Seated Rest | Yoga-based Physical Therapy, Seated Rest, Conventional Physical Therapy | Seated Rest, , Conventional Physical Therapy, Yoga-based Physical Therapy | Seated Rest, Yoga-based Physical Therapy, Conventional Physical Therapy | Conventional Physical Therapy, Seated Rest, Yoga-based Physical Therapy | Conventional Physical Therapy, Yoga-based Physical Therapy, Seated Rest |
|---|---|---|---|---|---|---|
| STARTED | 2 | 2 | 2 | 2 | 1 | 2 |
| COMPLETED | 2 | 2 | 2 | 2 | 1 | 2 |
| NOT COMPLETED | 0 | 0 | 0 | 0 | 0 | 0 |
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.
| Characteristic | Unit | All Participants |
|---|---|---|
| Participants analysed | 11 | |
| Age, Categorical: <=18 years | Participants | 0 |
| Age, Categorical: Between 18 and 65 years | Participants | 11 |
| Age, Categorical: >=65 years | Participants | 0 |
| Age, Continuous | years | 45.31 (SD 14.23) |
| Sex: Female, Male: Female | Participants | 5 |
| Sex: Female, Male: Male | Participants | 6 |
| Race and Ethnicity Not Collected | Participants | — |
| Region of Enrollment: United States | participants | 11 |
| Time since injury (days) | days | 15 (12 to 20) |
6. Primary Endpoint
The registry lists the primary outcome as Heart Rate Variability, measured over The sensors will be worn throughout the three 1 hour long conditions..
| Element | Registry entry |
|---|---|
| Result type | Median |
| Measure of spread | Inter-Quartile Range |
| Unit | ms |
| Endpoint classification | not reduced to a single comparison |
The primary endpoint is heart rate variability, reported in milliseconds. The registry says MC-10 sensors placed on the chest measured it, and that the sensors were worn throughout the three one-hour conditions. The primary summary parameter is the median, with inter-quartile range identified as the dispersion measure. The posted condition summaries give a median of 21.29 for yoga-based physical therapy, 27.36 for seated rest, and 18.49 for conventional physical therapy. The accompanying bounds are 16.57 to 40.3, 12.31 to 42.13, and 12.82 to 32.75, respectively.
Heart rate variability describes variation in the time intervals between heartbeats. In general, it is used as a physiological measure that may reflect changes in autonomic regulation. A session-based comparison can ask whether the measure differs during or around different activities, but the registry's endpoint description does not specify a particular physiological mechanism or define a clinically meaningful difference.
The posted summaries are condition-level medians rather than a model-based estimate of the change between conditions. The interval bounds should be read according to the registry's stated summary format; they do not substitute for a confidence interval for a treatment contrast. The record does not report a clinical threshold for interpreting the size of a difference in milliseconds.
7. Why No Re-analysis Is Shown
8. Statistical Interpretation
The registry reports condition-level median heart rate variability values of 21.29 for yoga-based physical therapy, 27.36 for seated rest, and 18.49 for conventional physical therapy. Its inferential result is a mixed models analysis across the three conditions, with p = 0.598. The record does not provide a model-based effect estimate, confidence interval, or pairwise test results. As a result, the p-value cannot be used to determine the magnitude or precision of each condition-to-condition contrast.
In general, a p-value assesses how compatible the observed data are with a specified null hypothesis under the assumptions of the analysis. A value of 0.598 is not a measure of the probability that the null hypothesis is true, nor does it establish that the conditions are equivalent. The registry record does not report the significance threshold or a prespecified equivalence margin. Without an effect estimate and its uncertainty, the result does not show how large a difference might be consistent with the observed data.
The medians are descriptive summaries. They do not by themselves show the within-participant change across conditions, which is especially important in a crossover study. A difference between condition medians is not necessarily the same as the estimated within-person contrast from a mixed model. The record does not report participant-level values, the distribution of paired differences, or the model terms used to account for repeated observations. The posted bounds describe the condition summaries, not uncertainty in a model-estimated comparison.
There are n = 11 observations reported for each of the three conditions. Because the design is crossover, these counts should not be added or interpreted as independent groups of people. The record does not state whether every participant contributed a complete set of measurements or explain how missing condition data were handled. Enrollment is listed as 13, but the record does not provide a participant flow accounting for the difference between enrollment and condition-level n.
The study was identified as a pilot and the brief summary describes the intention to inform a possible larger-scale study. A pilot result can describe the observed response and the feasibility context reported by the registry, but this record does not report a formal sample-size justification or a clinically meaningful effect threshold. No conclusion about the effects of a regular yoga-based program follows from this immediate, session-based comparison. The posted result supports a cautious description: the three-condition mixed-model test has p = 0.598, while the size and precision of condition contrasts are not reported in the registry record.
9. Statistical Methods Explained
What does a crossover design compare?
In a crossover study, participants receive multiple conditions in a sequence rather than remaining in only one assigned condition. Here, participants were to experience yoga-based physical therapy, seated rest, and conventional physical therapy in random order across three days. In general, this design can make within-person comparisons possible: each person's response under one condition can be considered alongside that person's response under another. This may reduce the influence of stable differences between participants on a contrast.
Repeated observations from one person are not independent observations. A suitable analysis therefore needs to account for the relationship among measurements from the same participant. The registry identifies mixed models analysis for the primary endpoint, a general class of methods that can model correlated observations. The precise model, its included terms, and the assumptions used for this study are not reported in the registry record. The record does not say how the analysis handled condition order, time or period, or incomplete sequences. Those details matter because a crossover comparison can be affected by sequence and by what happens between conditions.
What does the mixed-model p-value convey?
The reported mixed models analysis compares the three conditions and gives p = 0.598. In general, a p-value describes how unusual data at least as incompatible with a null hypothesis would be if that hypothesis and the analysis assumptions were correct. It is not an effect size and does not measure the probability that a result occurred by chance. The registry does not report a model estimate, confidence interval, or detailed null hypothesis for the analysis. The p-value therefore cannot tell the reader the estimated direction, magnitude, or precision of a particular pairwise comparison.
A confidence interval, in general, gives a range of values compatible with data and a specified statistical procedure. For a comparison, it can show both the estimated contrast and its uncertainty. The posted condition summaries have bounds, but those are associated with the reported median and inter-quartile range format. They are not a confidence interval for the mixed-model contrast. No confidence interval for that contrast is reported in the registry record.
Why distinguish summaries from comparisons?
A median describes the middle of an observed distribution and is less influenced by extreme observations than a mean. The registry identifies median as the primary summary and inter-quartile range as its dispersion measure. The posted medians are 21.29 for yoga-based physical therapy, 27.36 for seated rest, and 18.49 for conventional physical therapy. These values summarize each condition separately. In a crossover study, the inferential comparison may instead depend on paired measurements and on how the model represents the repeated observations. Comparing the medians alone does not recover that analysis.
Dispersion helps describe how spread out observations are. In general, the inter-quartile range describes the span from the lower to the upper quartile, thereby representing the middle portion of a distribution. It does not directly describe uncertainty in an estimated condition difference. The record does not report the individual measurements or the distribution of within-person changes, so the summaries cannot establish how consistently participants changed across conditions.
What can missing details change?
For repeated-measures data, incomplete observations can affect which information contributes to an estimate. In general, mixed models can use available repeated measurements under specified assumptions, but the method's handling of missingness depends on the model and the data pattern. The registry does not report missing-data counts, reasons for missingness, or a missing-data approach. It also does not report an analysis population definition or a participant flow describing inclusion in the primary analysis.
Random order helps address systematic allocation of conditions to sequence, but it does not by itself show that order effects were modeled or absent. Likewise, an intended shared setting for yoga and seated rest helps explain the planned comparison but does not establish that all environmental factors were identical. The record does not report a washout procedure or an assessment of carryover between sessions.
How should the sample-size illustrations be read?
In general, a sample-size calculation is a planning exercise: it asks how many participants per arm might be required under chosen assumptions for detecting a specified effect with a specified power. The sample-size assumptions associated with the posted planning illustrations describe a two-sided significance level of 0.05, equal allocation between two arms, and a normal approximation, with the effect expressed as a fraction of the observed effect. Such illustrations are not observed power, do not establish the study's original design rationale, and should not be treated as results from this crossover study. The registry record does not report the trial's own sample-size justification, and the planning assumptions do not describe the reported three-condition mixed-model result in full.
10. Analyses Reported by the Sponsor
| Method | Estimate | 95% CI | P-value |
|---|---|---|---|
| Mixed Models Analysis | not reported | 0.598 |
The registry's posted inferential result is a mixed models analysis across yoga-based physical therapy, seated rest, and conventional physical therapy, with p = 0.598. It does not report a model-based estimate or confidence interval. The condition summaries report medians and associated bounds, but those summaries are not a recalculated contrast and do not provide the same information as a model estimate.
No separate recalculated estimate or p-value for the primary three-condition comparison is reported in the registry record, so there is no numerical re-calculation to set alongside the mixed-model result. In general, a simple recalculation from summary statistics may not reproduce a crossover mixed model because it may not account for the pairing of measurements within participants or other model terms. The data needed to independently reproduce that analysis are not reported in the registry record.
The record does not flag a discrepancy between a sponsor analysis and a recalculation. The sponsor-reported mixed-model result is therefore the posted inferential result to interpret. Its p-value alone does not give an estimated difference, a confidence interval, or evidence of equivalence between conditions.
11. Multiplicity, Missing Data and Analysis Populations
The registry lists a primary endpoint of heart rate variability and two posted secondary outcomes: change in the Spielberger State-Trait Anxiety Inventory (STAI) - Short Form and change in the Global Fatigue Index-Modified. The study description also names agitation and sleep quality as secondary objectives, and says sleep would be assessed with an activity monitor. The record lists three outcome measures and one analysis, but does not explain the relationship between those counts and the named outcomes or provide posted results for every objective mentioned in the study description.
Multiplicity arises in general when several outcomes, time points, or comparisons are tested. The more tests considered, the greater the chance that at least one result may appear notable by chance if each is interpreted separately. A prespecified testing strategy or adjustment can address this concern, but the registry record does not report a multiplicity plan, a hierarchy among secondary outcomes, or an adjustment method. The primary p-value should therefore be read as the reported test for the primary endpoint, not as a complete accounting of all possible outcome and contrast tests.
The record does not provide an analysis population definition, a participant flow, or missing-data counts. It also does not report whether the primary analysis included all enrolled participants, only participants with measurements for every condition, or another analysis set. Although the condition-level summaries each have n = 11 and enrollment is listed as 13, the registry does not explain this accounting. It does not report reasons for missing measurements, how partial crossover data were handled, or whether any sensitivity analyses were conducted.
These omissions limit assessment of selection into the analysis and the robustness of results to missingness. They do not establish that missing data or multiplicity changed the reported result; their handling is simply not reported in the registry record.
12. Safety Results
The registry reports serious adverse events as affected participants divided by participants at risk.
| Arm | Participants with serious adverse events | Participants at risk | Percentage |
|---|---|---|---|
| Yoga-based Physical Therapy Group | 0 | 11 | 0.0% |
| Seated Rest | 0 | 11 | 0.0% |
| Conventional Physical Therapy | 0 | 11 | 0.0% |
Deaths
| Arm | Deaths (participants affected of at risk) |
|---|---|
| Yoga-based Physical Therapy Group | 0 of 11 |
| Seated Rest | 0 of 11 |
| Conventional Physical Therapy | 0 of 11 |
Other (non-serious) adverse events
Participants with at least one non-serious event at or above the registry's 0% reporting threshold.
| Arm | Participants affected of at risk |
|---|---|
| Yoga-based Physical Therapy Group | 0 of 11 |
| Seated Rest | 0 of 11 |
| Conventional Physical Therapy | 0 of 11 |
The posted adverse-event figures report zero serious adverse events and zero deaths for yoga-based physical therapy, seated rest, and conventional physical therapy. The record lists 11 at risk for each condition for serious adverse events and for other adverse events. It reports zero affected for each of those categories in each condition. No top serious adverse-event terms are listed.
These figures describe what was recorded in the registry for the reported groups and conditions. In general, a count of zero does not establish that an event is impossible or that no event could occur in a broader population. It indicates that no events were reported in the posted figures for the listed categories and denominators. With a pilot study and limited reported counts, the safety information is descriptive and does not provide a precise estimate of event risk.
The crossover design means the same participants were intended to experience multiple conditions. The condition-specific at-risk figures should not be added to infer a count of unique people. The registry does not report the adverse-event collection period, detailed event definitions, severity grading beyond the serious/other categories, or how events were attributed to a condition. The record also does not explain whether the at-risk counts reflect complete exposure to all conditions. Accordingly, these figures should be read as the posted safety record, not as proof that a condition carries no risk.
13. What the Result Does, and Does Not, Mean
The reported data describe heart rate variability during three planned conditions in a randomized crossover pilot study involving people with traumatic brain injury. The registry reports median values of 21.29 for yoga-based physical therapy, 27.36 for seated rest, and 18.49 for conventional physical therapy. The mixed models analysis across the conditions has p = 0.598. The record does not provide an estimated contrast or confidence interval, so it does not quantify the size or precision of a difference between any two conditions.
In general, a non-small p-value does not prove that conditions are identical. Establishing equivalence or ruling out a meaningful difference requires a defined margin and an analysis designed to address that question. The registry record does not report an equivalence margin or such an analysis. Conversely, the numerical ordering of condition medians alone does not establish a reliable difference, because those medians are descriptive and do not capture the model's within-person comparison and uncertainty.
The result concerns an immediate, session-based comparison as described in the registry. It does not test the effects of regular participation in a yoga-based program over time, nor does it report evidence about long-term outcomes. The study summary says the pilot could inform a larger-scale investigation of regular participation as an adjunct to traditional physical therapy; that is a research aim, not a result established by this record.
The registry also describes anxiety, fatigue, agitation, and sleep quality as objectives, while it lists anxiety and fatigue as secondary outcomes. Their results are not included in the registry record recorded for this analysis. No conclusion about those outcomes can be drawn from the primary heart rate variability result. The appropriate reading is limited: the posted primary analysis did not provide a model estimate or confidence interval, and the record does not establish either a clinically important benefit or equivalence among the conditions.
14. Why This Trial Matters Statistically
This trial is statistically informative as an example of how design and reporting shape interpretation. A crossover design exposes participants to multiple conditions and can support within-person comparisons. That potential depends on a suitable analysis of repeated observations and clear reporting of the sequence, model, and handling of incomplete data. The registry identifies a mixed models analysis, which is consistent with analyzing related observations, but does not provide the model specification or a contrast estimate.
The posted condition summaries are medians, while the inferential result is a three-condition mixed-model p-value. Those answer different questions. The medians describe the observed condition distributions; the model test addresses a statistical comparison under the model. Without an estimate and uncertainty interval, the reader cannot tell how large the modeled difference was or how precisely it was estimated. This is why a p-value should not be read as a complete account of the evidence.
The record also illustrates why reporting participant flow and analysis populations matters in a small pilot. Enrollment is listed as 13, while each condition summary reports n = 11, with no explanation of the accounting. This does not identify a particular source of missingness, but it leaves the reader unable to assess who contributed data to each analysis or how incomplete observations were handled.
Finally, the study's stated purpose was exploratory and session-based. Pilot findings can help frame questions for later studies, but the record does not report a design justification, a clinically meaningful threshold, or a formal plan for multiple outcomes. The statistical contribution is therefore best understood as a limited, descriptive and inferential snapshot, rather than a definitive comparison or evidence about a sustained program.
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 interpretive limitation is that the registry reports a mixed-model p-value of 0.598 but no model-based estimate or confidence interval. This prevents assessment of the magnitude and precision of the primary contrast. The record also does not report pairwise results, a prespecified significance threshold, or an equivalence framework. The condition-level medians and bounds cannot replace those missing inferential details.
The record identifies a randomized crossover design but does not specify the mixed model, its covariance structure, or whether it included terms for condition sequence or period. It does not report a washout procedure or assessment of carryover. These are relevant design and analysis details for repeated measurements collected over separate conditions. Their absence from the registry record does not establish that they were not addressed in the study; it means they cannot be evaluated from this record.
Enrollment is recorded as 13, and the planned enrollment description says up to 30, while each condition summary reports n = 11. The record does not provide a flow of participants, the analysis population, reasons for missing measurements, or the handling of incomplete condition data. The resulting uncertainty about the analysis set limits assessment of possible selection effects.
The record describes a pilot and an immediate session-based question, but does not report a sample-size justification or clinically meaningful difference. Secondary aims include anxiety, fatigue, agitation, and sleep, while the posted secondary outcome list names anxiety and fatigue. The record does not provide results for these outcomes or describe multiplicity handling. Finally, the single-masking designation is not accompanied by information about who was masked, and the record does not describe how masking was maintained.
16. Related Tutorials
17. Related Statistical Calculators
18. Sources
- ClinicalTrials.gov: The Impact of Yoga-based Physical Therapy for Individuals With Traumatic Brain Injury — NCT03701594.
- Registry record last updated 2026-06-04; status last verified 2026-06.
- Results first posted 2026-06-04.
- 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.
Continue through the Clinical Biostats statistical pathway
Use this trial's endpoint as a starting point for deeper study of effect measures, confidence intervals, and sample-size planning.
19. Record Summary
NCT03701594 is a completed randomized, single-masked crossover pilot study sponsored by Shirley Ryan AbilityLab. It examined yoga-based physical therapy, seated rest, and conventional physical therapy among people with traumatic brain injury. The record lists enrollment of 13 and reports n = 11 for each condition's primary endpoint summary.
The primary endpoint was heart rate variability in milliseconds, measured with chest-worn MC-10 sensors during the three one-hour conditions. The posted medians are 21.29 for yoga-based physical therapy, 27.36 for seated rest, and 18.49 for conventional physical therapy. The reported mixed models analysis across conditions has p = 0.598. A model-based estimate and confidence interval are not reported in the registry record.
The registry lists anxiety and fatigue as secondary outcomes and reports zero serious adverse events and zero deaths in the posted condition figures. Details on the analysis population, missing data, model specification, and multiplicity are not reported in the registry record.