SISTERS: Spasticity In Stroke Study - Randomized Study: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for SISTERS: Spasticity In Stroke Study - Randomized Study (NCT01032239), a randomized phase 4 study. The primary outcome is Change in Average Ashworth Scale (AS) in Affected Lower Extremities From Baseline to Month 6.

2. Clinical Question

3. Trial Design

SISTERS was a phase 4, randomized, controlled, open-label, parallel-group, multicenter study. Its stated purpose was to compare intrathecal baclofen therapy with Best Medical Treatment for severe spasticity in post-stroke patients over six months of active treatment. The registry record lists an enrollment of 61 and a completed status. The recorded start was 2009-11 and the primary completion was 2016-09.

The intervention arm received intrathecal baclofen therapy with an implantable pump. The comparator arm used one or a combination of oral antispastic medications. Because masking was listed as none, the study was open-label: participants and study personnel were not masked to treatment assignment according to the registry record. The registry record does not report further details about allocation concealment, treatment adherence, crossover, or the clinical centers.

The primary outcome summary reports 25 participants in the intrathecal baclofen arm and 26 in the Best Medical Treatment arm. These are the outcome-specific sample sizes; the registry record does not explain why they differ from the enrollment total. The primary analysis population was described as intent to treat, with last observation carried forward imputation. The record lists 12 outcome measures and 14 analyses, and posts 11 secondary outcomes. It does not report the original sample-size justification, a detailed analysis plan, or how the 14 analyses map to the outcome measures.

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: A total of 61 patients were enrolled. One patient did not meet the inclusion criteria (screening failure). 60 patients were then randomized into the study.

Overall Study

Reasons for not completing this period:

Analysis population for the primary outcome, as defined in the registry: Intent to treat analysis. Last Observation Carried Forward (LOCF) imputation method.

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 Average Ashworth Scale (AS) in Affected Lower Extremities From Baseline to Month 6, measured over Baseline and month 6.

The primary endpoint was change in average Ashworth Scale score in the affected lower extremities from baseline to month 6. The registry defines change as the month 6 score minus the baseline score. Thus, a negative change represents a lower score at month 6 than at baseline; the scale’s lower values indicate less resistance to passive movement. The measure is based on assessment of six lower-extremity muscle groups in the affected limb, with their Ashworth scores averaged.

The Ashworth Scale is a manual examination of resistance to passive movement about a joint at varying movement velocities. In the registry description, its score choices range from 1 to 5: a score of 1 indicates no resistance and a score of 5 indicates rigidity. Averaging across the specified muscles gives a summary measure of lower-extremity resistance rather than a separate result for each muscle.

This endpoint was used to assess the study’s primary question about severe spasticity. A continuous change score retains the size and direction of change on the scale, rather than reducing outcomes to categories such as improved or not improved. It is still a scale-based clinical measure: its interpretation depends on the measurement properties and on the study’s definition of the affected extremity and assessed muscles. The registry record does not report a threshold for a change considered clinically important.

7. Results Posted in the Registry

• Best Medical Treatment (BMT): mean -0.43 versus -0.99 on control; difference 0.56 (95% CI 0.16 to 0.96), standardized effect d = 0.76; z-test p = 0.007.

8. Secondary Efficacy Endpoints

The registry also posts results for secondary outcomes. For each outcome below, the endpoint type, the test applied to the posted summary statistics, the result and its interpretation are given in turn. Secondary endpoints are usually not adjusted for multiple testing, so they support rather than confirm the primary finding.

Change in Average Ashworth Scale (AS) in Affected Upper Extremities From Baseline to Month 6

• Endpoint type: Continuous: a mean value per participant, compared as a difference in means.
• Test: Unadjusted two-sample z-test for the difference in means, using the posted standard deviations.
• Result:
  • Best Medical Treatment (BMT): mean -0.17 versus -0.66 on control; difference 0.49 (95% CI 0.14 to 0.84), standardized effect d = 0.76; z-test p = 0.007.
• Interpretation:
  • The 95% confidence interval excludes no difference, so the mean is statistically higher for Best Medical Treatment (BMT) than for the control. This is an unadjusted, secondary comparison, so it is suggestive rather than confirmatory.

Change in Functional Independence Measure (FIM) Score From Baseline to Month 6

• Endpoint type: Continuous: a mean value per participant, compared as a difference in means.
• Test: Unadjusted two-sample z-test for the difference in means, using the posted standard deviations.
• Result:
  • Best Medical Treatment (BMT): mean -2.58 versus 2.68 on control; difference -5.26 (95% CI -11.11 to 0.59), standardized effect d = -0.49; z-test p = 0.078.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Best Medical Treatment (BMT) and the control. The interval gives the range of effects that remain compatible with the data.

Change in Average 10 Meter Time Walking Test (10MTWT) From Baseline to Month 6

• Endpoint type: Continuous: a mean value per participant, compared as a difference in means.
• Test: Unadjusted two-sample z-test for the difference in means, using the posted standard deviations.
• Result:
  • Best Medical Treatment (BMT): mean -2.48 versus 4.86 on control; difference -7.34 (95% CI -46.21 to 31.53), standardized effect d = -0.12; z-test p = 0.711.
• Interpretation:
  • The 95% confidence interval includes no difference, so these data do not show a statistically significant difference between Best Medical Treatment (BMT) and the control. The interval gives the range of effects that remain compatible with the data.

The secondary outcomes cover several domains: affected upper-extremity spasticity, functional independence, walking time, transfers, pain, goal attainment, and health-related quality of life. The registry lists additional measures of these domains, including EQ-5D-3L and SF-12. Across the secondary outcomes for which numerical comparisons are posted, the pattern is mixed rather than uniformly pointing in one direction. The upper-extremity Ashworth result is directionally consistent with the lower-extremity primary outcome, while the posted functional independence and walking comparisons are less conclusive. The registry record does not post numerical results for every listed secondary outcome in the available outcome summaries.

These outcomes relate to the primary result by addressing other aspects of function and experience, but they are not substitutes for the primary endpoint. Change on a spasticity scale does not, by itself, determine change in walking, independence, pain, goal attainment, or quality of life. Each outcome also has its own units, measurement properties, and number of observed participants. For example, the walking outcome has fewer observations per arm than the primary measure; the registry record does not explain the reasons for the outcome-specific sample sizes.

Unadjusted secondary results carry less weight because several outcomes and analyses are reported, and the registry record does not report whether or how multiplicity was controlled. General: when many hypotheses are tested, the chance of obtaining at least one apparently notable result by chance can increase. Without a prespecified hierarchy or adjustment information, individual secondary p-values should be treated as exploratory rather than as definitive confirmation. The differing secondary findings should therefore be considered alongside the primary result and the lack of reported multiplicity procedures, not as a single consistent conclusion across all measured domains.

Secondary outcomes registered for this trial, for reference (only those above were recalculated):

• Change in Average Ashworth Scale (AS) in Affected Upper Extremities From Baseline to Month 6
• Change in Functional Independence Measure (FIM) Score From Baseline to Month 6
• Change in Average 10 Meter Time Walking Test (10MTWT) From Baseline to Month 6
• Number of Participants Who Were Able to Transfer From the Wheelchair to Bed Without Human Assistance
• Change in Numeric Pain Rating Scale (NPRS) From Baseline to Month 6
• Number of Participants Who Achieved Their Primary Therapeutic Goal Assessed With the Goal Attainment Scale (GAS)
• Change in Euro QoL Group-5 Dimensional, 3 Level Version (EQ-5D-3L) From Baseline to Month 6
• Change in SF-12 (12-item Short Form) From Baseline to Month 6

9. Statistical Interpretation

10. Statistical Methods Explained

What does the mean change summarize? General: a mean change is the average of participants’ endpoint changes, where each person’s change is calculated using the study’s stated subtraction order. In this trial, change was defined as the month 6 Ashworth score minus the baseline score. Negative group means therefore indicate that scores were lower at month 6 on average. A mean is sensitive to the values of individual observations and can conceal variation: participants can have different changes even when the group average is the same. The posted standard deviations describe the spread of individual changes around each arm’s mean, rather than the precision of the mean itself.

What does the recalculated mean difference represent? The unadjusted recalculation compares the two posted group means. For the primary outcome it is expressed as Best Medical Treatment minus intrathecal baclofen, so a positive value corresponds to a more negative, or larger downward, mean change in the intrathecal baclofen arm. The difference is a contrast between arm-level averages; it does not adjust for other participant characteristics. It is recalculated from posted summary statistics, not from individual participant records, and therefore cannot reproduce analyses requiring the underlying participant-level observations.

Why is the registry analysis different? The registry reports a Wilcoxon (Mann-Whitney) analysis with a Hodges-Lehmann estimate. General: the Wilcoxon rank-sum approach compares the ordering of observations between groups rather than directly comparing arithmetic means. The Hodges-Lehmann estimate is a rank-based location estimate. It is therefore not the same estimand as the difference between the two arithmetic means, and its sign depends on the order in which groups are contrasted. A rank-based result can differ from a mean-based recalculation, particularly when the distributions are not well represented by a simple mean-and-standard-deviation description. The registry record does not report distributional diagnostics or further details about this test’s implementation.

What do a p-value and a confidence interval convey? General: a p-value describes how compatible the observed data are with a specified null hypothesis under the assumptions of the test. It is not the probability that the null hypothesis is true, nor a measure of clinical importance. A confidence interval gives a range of effect estimates compatible with the data and method at its stated confidence level. For the sponsor-reported primary analysis, the interval excludes zero, consistent with its reported p-value. The recalculated interval also excludes zero. These two inferential results arise from different estimators and methods and should be kept distinct.

What does intent to treat with LOCF mean? The primary outcome population is described as intent to treat, with last observation carried forward imputation. General: an intent-to-treat analysis aims to retain participants in the groups to which they were assigned, regardless of later treatment experience, although the exact implementation depends on the protocol. Last observation carried forward fills a missing follow-up value with that participant’s most recent observed value. This method makes an assumption that the last observed value is an appropriate substitute for the missing month 6 value; the assumption may not reflect the participant’s actual later course. The registry record does not report the amount or pattern of missing data, reasons for missingness, or sensitivity analyses using other missing-data methods.

Why does dispersion matter? General: the standard deviation describes variation among participant-level outcomes. Larger variation makes an estimate of a group mean less precise when sample sizes are held fixed, and can make it harder to distinguish a group contrast from sampling variability. The registry posts standard deviations for the arm-level changes, but not individual observations. This means the reported mean and spread do not reveal the full shape of either outcome distribution, potential outliers, or whether a small number of observations strongly influence the mean.

What can the posted analysis establish? The registry reports a primary superiority analysis, with a rank-based test and estimate, and the summary statistics allow an unadjusted mean-based recalculation. Neither result alone resolves questions about clinical importance, missing-data assumptions, or generalizability. The registry record does not report a detailed statistical analysis plan, the original sample-size justification, interim analyses, multiplicity procedures, or model adjustments. Those elements should not be inferred from the analysis results or the separate planning illustrations.

11. Confidence Intervals and Precision

For the primary outcome, the sponsor-reported rank-based analysis gives a 95.1% confidence interval for the Hodges-Lehmann estimate from -1.000 to -0.167. The entire interval is below zero in the reported group ordering. The corresponding test has p = 0.014. The interval describes uncertainty in the rank-based estimate under that analysis; it is not an interval for the unadjusted mean difference.

The unadjusted recalculation compares mean changes and gives an estimated mean difference of 0.56 units, with a confidence interval that excludes zero. Its confidence interval is based on the posted arm means and standard deviations. It answers a mean-difference question, so it should not be substituted for the confidence interval accompanying the registry’s Hodges-Lehmann result. The registry record does not post confidence limits for the two arm means themselves.

General: a confidence interval is a range of values that are compatible with the observed data and the method’s assumptions at the stated confidence level. A narrower interval indicates greater precision for that particular estimand and method; it does not guarantee that the estimate is free from bias or that the measure is clinically important. An interval containing zero would be compatible with both directions of a difference as well as no difference, while an interval excluding zero indicates that zero is not among the estimates in that interval.

Precision must also be considered alongside the outcome-specific sample sizes: 25 and 26 for the primary outcome. The registry record reports enrollment of 61, but does not explain why the primary outcome sample sizes are smaller. The open-label design, LOCF approach, and unavailable information on missingness and analysis details add uncertainty that is not fully represented by a confidence interval. The interval quantifies sampling uncertainty under the analysis method, not every source of uncertainty in the study.

12. Planning a Similar Trial: Sample Size

This is a planning illustration, not a re-evaluation of the original trial. It asks: if a new study wanted to detect a difference like the one observed, or a smaller one, how large would it need to be? It is deliberately not “observed power” of the finished trial, which is not a meaningful quantity.

The planning figures are participants per arm, not total participants or total events. They are illustrations based on the observed effect, using a two-sided significance level of 0.05, equal allocation between the two arms, and a normal approximation. Each row considers a stated fraction of the observed effect: the full observed effect, 75% of the observed effect, or 50% of the observed effect. The planning assumptions describe these illustrations, not the trial’s original design. The registry record does not report the original sample-size justification or its design assumptions.

The 80% power column gives the illustrated number of participants per arm for a design with 80% power to detect the selected effect under those assumptions. The 90% power column gives the corresponding number per arm for 90% power. General: power is the chance that a specified statistical procedure will reject its null hypothesis if the particular effect used for planning is the true effect, under the model’s assumptions. Higher planned power generally requires more participants for a fixed effect and analysis.

Smaller effects need more participants to detect under otherwise similar assumptions. In the primary outcome illustration, planning for 75% of the observed effect uses 49 participants per arm for 80% power and 65 for 90% power; planning for 50% uses 109 and 145, respectively. These are design illustrations, not the number of participants observed in the trial. They are not observed power: they do not calculate the probability that the completed trial detected its effect after the results were known. The curve should be read as a visual expression of how required sample size changes with the effect being targeted, not as evidence about the trial’s original planning process.

13. How the Sample Size Is Derived

Power is the probability that a study of a given size rejects the null hypothesis when a specific true effect exists. It depends on four things: the effect to detect, the variability of the outcome, the significance level (here a two-sided α of 0.05, so z_1−α/2 = 1.960) and the desired power (z_1−β = 0.842 for 80% and 1.282 for 90%).

Formula

n is the number of participants in each arm, Δ is the difference in means to detect and σ is the common standard deviation.

Worked calculation for this trial

1. Observed difference Δ = 0.560; pooled standard deviation σ = 0.735.
2. n = 2 × (1.96 + 0.842)² × 0.735² / 0.560² = 27.0, rounded up to 28 participants per arm for 80% power.

The smaller-effect rows in the table above repeat the same calculation with the effect multiplied by 0.75 and 0.5 (for ratios, applied to the logarithm of the ratio).

R code

alpha <- 0.05
z_a   <- qnorm(1 - alpha / 2)          # two-sided
delta <- 0.56                      # observed difference in means
sigma <- 0.734846922834953                     # pooled standard deviation
n_per_arm <- function(power, frac = 1) {
  ceiling(round(2 * (z_a + qnorm(power))^2 * sigma^2 / (delta * frac)^2, 6))
}
for (f in c(1, 0.75, 0.5))
  cat(sprintf("%3.0f%% of observed effect: n per arm = %s (80%% power), %s (90%% power)\n", 100 * f,
              format(n_per_arm(0.80, f), big.mark = ","), format(n_per_arm(0.90, f), big.mark = ",")))

# Check against R's built-in function (t-based, so very slightly larger)
cat("power.t.test:", round(power.t.test(delta = delta, sd = sigma, power = 0.80)$n, 1), "\n")

# Simulation: exact two-sample t-test using simulated means and variances
set.seed(1032239)
n    <- n_per_arm(0.80)
reps <- 20000
m_diff <- rnorm(reps, delta, sigma * sqrt(2 / n))             # difference in sample means
s2     <- sigma^2 * rchisq(reps, 2 * (n - 1)) / (2 * (n - 1))   # pooled sample variance
tstat  <- m_diff / sqrt(s2 * 2 / n)
cat(sprintf("Simulated power at n = %s per arm: %.3f (%d replicates)\n", format(n, big.mark = ","), mean(abs(tstat) > qt(0.975, 2 * (n - 1))), reps))

Output from running the code

100% of observed effect: n per arm = 28 (80% power), 37 (90% power)
 75% of observed effect: n per arm = 49 (80% power), 65 (90% power)
 50% of observed effect: n per arm = 109 (80% power), 145 (90% power)
power.t.test: 28 
Simulated power at n = 28 per arm: 0.799 (20000 replicates)

The simulation draws many sets of group means and variances at the planned sample size and counts how often the two-sample t-test rejects the null. It should land close to the target power; at very small required sample sizes it can fall below it, because the normal approximation is rough there.

14. Analyses Reported by the Sponsor

The registry’s own primary analysis used the Wilcoxon (Mann-Whitney) method and reported a Hodges-Lehmann estimate of -0.667, a 95.1% confidence interval from -1.000 to -0.167, and p = 0.014. This is the sponsor-reported analysis as posted in the registry record. Its effect estimate is rank-based and is reported in the registry’s group ordering.

The unadjusted recalculation uses the posted arm means and standard deviations to estimate the difference between arithmetic mean changes. With the contrast expressed as Best Medical Treatment minus intrathecal baclofen, the recalculated estimate is 0.56 units and its unadjusted p-value is 0.007. The sign differs because the contrast order differs, and the estimands differ because a difference in means is not a Hodges-Lehmann estimate. The values therefore should not be compared as if they were two versions of precisely the same statistic.

The registry record does not flag a discrepancy between the reported analysis and posted summaries. The recalculation is useful as a transparent check based on summary statistics, but it does not replace the sponsor’s own analysis or reproduce an individual-level analysis. The reported rank-based result should be interpreted according to its stated method, while the mean-based result describes the unadjusted contrast between posted averages.

15. Multiplicity, Missing Data and Analysis Populations

The registry record lists 12 outcome measures and 14 analyses, with 11 secondary outcomes posted. It reports a primary outcome analysis and several secondary analyses, but does not report a multiplicity adjustment, a hierarchy among secondary outcomes, or a detailed plan for interpreting the family of tests. Consequently, the posted p-values should not automatically be read as if only one hypothesis had been tested. General: testing multiple outcomes can increase the chance of at least one small p-value arising by chance, and a prespecified strategy may address that issue through adjusted tests, a hierarchical sequence, or clearly defined exploratory analyses. The record does not indicate which approach, if any, was planned.

The primary endpoint population is described as intent to treat, with last observation carried forward imputation. The record does not report how many primary endpoint values were missing, when missingness occurred, or the reasons for it. It also does not report whether participants with missing values were handled in other ways for any secondary outcome. LOCF replaces a missing later measurement with an earlier observed value, which assumes that this carried-forward value is an adequate representation of the missing result. If outcomes continue to change after the last observation, that assumption may not hold.

The registry record does not report sensitivity analyses comparing LOCF with alternative missing-data approaches, nor does it provide enough detail to assess whether missingness differed between treatment groups. The outcome-specific sample sizes vary, and their reasons are not reported. The record also does not report the original sample-size justification, interim analyses, covariate adjustments, or a full statistical analysis plan. These omissions limit the ability to assess how analysis choices were prespecified and how robust the reported results are to alternative assumptions. They should be treated as unreported information, not filled in by assumptions about common trial practice.

16. Safety Results

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

Deaths

Other (non-serious) adverse events

Participants with at least one non-serious event at or above the registry's 0% reporting threshold.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: implant site infection (8.0%), urinary retention (8.0%), device dislocation (8.0%), urinary tract infection (5.7%), cystitis (4.0%).

The registry reports serious and other adverse event counts in groups labeled ITB-I and BMT+ITB-NI. For serious adverse events, 12 of 25 participants were affected in the ITB-I group, and 10 of 35 were affected in the BMT+ITB-NI group. For other adverse events, the affected counts were 23 of 25 and 21 of 35, respectively. The record lists 1 death in the ITB-I group and 0 in the BMT+ITB-NI group.

The registry’s most frequent serious adverse event terms include implant site infection, urinary retention, and device dislocation, each listed as 8.0%; urinary tract infection at 5.7%; and cystitis and bacterial pneumonia, each at 4.0%. These are registry-reported terms and percentages. The record does not specify the denominator used for each percentage in this summary or provide enough detail to determine event timing, severity beyond the serious classification, or whether events were judged treatment-related.

These counts are descriptive and should not be interpreted as proof that a treatment caused or prevented an event. The safety group labels and denominators differ from the primary efficacy arm summaries, and the registry record does not explain the relationship in detail. Serious event counts, other event counts, and deaths are distinct summaries; they should not be combined or compared without considering their definitions and follow-up.

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

The primary result describes a between-group difference in change on the average Ashworth Scale in affected lower extremities over the recorded baseline-to-month-6 interval. The posted arm means were negative, and the unadjusted mean contrast favored a larger decrease in the intrathecal baclofen group under the stated contrast order. The registry’s reported rank-based analysis also reported a statistically significant superiority result. These are findings about the measured outcome and the analyses used.

They do not establish that every participant improved, that the change was important to participants, or that the same pattern occurred for every secondary outcome. Group averages can conceal individual variation. The registry record does not report a minimum clinically important difference for the Ashworth Scale, so statistical evidence of a difference cannot be translated from this record alone into a judgment about clinical importance. Nor does a result on a spasticity scale establish a corresponding effect on walking, independence, pain, goal attainment, or quality of life.

The comparison was randomized but open-label. Randomization supports a treatment-group comparison, but lack of masking can influence care, reporting, or assessments, particularly for a manual measure. The registry record does not report further safeguards or assessor masking. The primary analysis also used LOCF, and the record does not report missing-data patterns or alternative analyses. These limitations affect how confidently the result can be generalized beyond the analysis as posted.

The secondary results are mixed, and multiplicity handling is not reported. The available results therefore do not establish a broad effect across all measured domains. They also do not provide a complete assessment of harms: adverse event counts are descriptive, and causal attribution is not reported in the summary. The defensible interpretation is limited to the specific posted comparisons, their statistical uncertainty, and the design and reporting details recorded for this trial.

18. Why This Trial Matters Statistically

This trial illustrates why the estimand and analysis method matter when interpreting a result. The registry’s primary analysis is rank-based and reports a Hodges-Lehmann estimate, while the recalculation from posted summary statistics compares arithmetic means. Both can address a treatment-group contrast, but they do not estimate precisely the same quantity. A sign difference can also arise from the order in which groups are subtracted. Reporting the method and contrast direction is essential for making sense of the result.

The record also illustrates the distinction between statistical evidence and precision or importance. The primary comparison has a reported confidence interval excluding zero, but the registry record does not state a clinically important difference threshold. Statistical evidence alone does not determine whether a scale change matters to participants, and it does not establish effects on other outcomes. A confidence interval helps describe uncertainty in an estimate, while design features and measurement limitations remain relevant to interpretation.

Finally, the registry’s primary population is described as intent to treat with LOCF imputation, while outcome-specific sample sizes vary and secondary results span several domains. The record does not provide a multiplicity plan, missing-data sensitivity analyses, or original design assumptions. These absences show why registry summaries should be read with attention to what is stated and what remains unreported. Planning illustrations based on the observed effect are useful for understanding sample-size dependence on a target effect, but they are not evidence of the trial’s original design or its achieved power.

19. 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.

Several limitations specific to this study affect interpretation. The trial was open-label, with masking listed as none. The primary endpoint relied on a manual assessment of resistance to passive movement, so the absence of masking may be relevant to assessment. The registry record does not report whether outcome assessors were masked or what procedures were used to reduce assessment bias.

The primary outcome used last observation carried forward imputation, but missing-data counts, patterns, reasons, and sensitivity analyses are not reported. This leaves uncertainty about how much the primary estimate depends on the imputation assumption. Outcome-specific participant counts also vary, including fewer observations for walking time than for the primary endpoint; the registry record does not explain these differences. The posted efficacy sample sizes do not account for the full enrollment total in a way explained by the record.

The primary outcome is a scale-based measure, and the registry record does not state a threshold for clinical importance. The record also does not report full individual-level distributions, making it impossible to assess the shape of the data or the influence of unusually large changes from posted summary statistics alone. The primary mean-based recalculation and reported rank-based result estimate different summaries and should not be treated as interchangeable.

Finally, the study reports multiple outcome measures and analyses, but multiplicity procedures, the original sample-size justification, interim analyses, and a detailed statistical analysis plan are not reported in the registry record. Secondary findings therefore require caution, and the safety counts cannot establish treatment causality from the information recorded.

20. Related Tutorials

21. Related Statistical Calculators

22. Sources

• ClinicalTrials.gov: SISTERS: Spasticity In Stroke Study - Randomized Study — NCT01032239.
• Registry record last updated 2018-02-15; status last verified 2018-01.
• Results first posted 2018-02-15.
• 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.

23. Record Summary

SISTERS was a completed, randomized, open-label, parallel-group, multicenter phase 4 study comparing intrathecal baclofen therapy with Best Medical Treatment for severe post-stroke spasticity. The primary endpoint was change in average Ashworth Scale score in affected lower extremities from baseline to month 6, analyzed in an intent-to-treat population with last observation carried forward imputation.

The registry reports mean changes of -0.99 for intrathecal baclofen and -0.43 for Best Medical Treatment. Its primary rank-based analysis reports a Hodges-Lehmann estimate of -0.667, a 95.1% confidence interval from -1.000 to -0.167, and p = 0.014. An unadjusted recalculation from posted summary statistics estimates a mean difference of 0.56 units, with a confidence interval excluding zero and p = 0.007. These are results from different statistical methods and estimands.

Secondary outcome findings are mixed. The registry record does not report multiplicity procedures, missing-data sensitivity analyses, the original sample-size justification, or a detailed statistical analysis plan. Safety counts are descriptive and do not establish causality.