Comparative Effectiveness of Rehabilitation Services for Survivors of Acute Ischemic Stroke: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Comparative Effectiveness of Rehabilitation Services for Survivors of Acute Ischemic Stroke (NCT02284165), a registered phase not stated study. The primary outcome is Home-time.

2. Clinical Question

The study’s stated purpose was broader than a single comparison of rehabilitation destinations. It was designed as a retrospective assessment of patterns, predictors, and outcomes associated with rehabilitation-service use after ischemic stroke hospitalization. Its data sources were Get With the Guidelines-Stroke, Medicare claims, and the AVAIL longitudinal data sets. The record identifies ischemic stroke as the condition and Duke University as the sponsor.

The clinical question is therefore observational: it concerns service use and subsequent outcomes in recorded patient data, rather than the effect of a treatment assigned by the study. Although the eligible population includes patients discharged to two types of care, the registry does not provide the group definitions or results needed to interpret them as a direct comparison. The central reported endpoint is home-time, while the listed secondary outcomes concern function, quality of life, institutionalization, rehospitalization, and death. Numerical results for those outcomes are not reported in the registry record.

3. Trial Design

This is a retrospective analysis, not a randomized clinical trial. The registry lists the study activity as a retrospective analysis and does not state an allocation method. The record describes linking Get With the Guidelines-Stroke data with Medicare claims and the AVAIL longitudinal data sets to examine rehabilitation-service patterns, predictors, and outcomes after hospitalization for ischemic stroke. These features make the study an analysis of existing data rather than a prospective assignment of rehabilitation services by investigators.

The registry reports total enrollment of 164246. The primary-outcome population description identifies a narrower eligible group: patients with acute ischemic stroke in the Get With the Guidelines registry, Medicare fee-for-service insurance, discharge to inpatient rehabilitation or skilled nursing care, and no survival limitations identified in hospital. It states that 69,212 patients met those criteria. The record does not explain the relationship between total enrollment and this endpoint population in further detail.

The study status is completed. The recorded start is 2013-05, and primary completion is 2016-05. The record defines primary completion as the end of primary analyses, consistent with the end of the PCORI Cycle I grant period. It defines study completion as acceptance of the final research report and lay abstract by PCORI. These are registry descriptions of completion, not dates of patient follow-up for each outcome.

The primary endpoint is mean home-time in days at 12 months post-discharge. Seven outcome measures are reported as registered, including the primary outcome and secondary outcomes. The registry lists zero analyses, and it does not provide analysis-specific estimates, uncertainty intervals, or hypothesis-test results. It also does not report a randomization procedure, treatment allocation, or a protocol-based sample-size justification. The study should therefore be read as an observational, linked-data analysis with outcome results that are not numerically 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: The following exclusions were applied prior to group assignment: death during index admission (n=10,466); loss of FFS eligibility not due to death during follow-up (n=1683); discharged to Federal hospital (n=102); transferred to another acute care hospital or discharged directly to nursing home for long-term care (n=4279).

Overall Study

Analysis population for the primary outcome, as defined in the registry: Acute ischemic stroke patients in the Get With the Guidelines registry with Medicare fee-for-service health insurance, analyzed as one cohort rather than by arm to differentiate the different types of rehabilitation services within the full patient population.

5. Baseline Characteristics

The table reproduces the baseline characteristics posted in the registry, which describe the participants analysed at baseline and help show whether the randomized groups were comparable.

Population: Acute ischemic stroke patients in the Get With the Guidelines registry with Medicare fee-for-service health insurance. Differentiation into study arms by discharge location meets the first primary aim of the study, to describe rehab service use through 90 days.

6. Primary Endpoint

The registry lists the primary outcome as Home-time, measured over 12 months post-discharge.

Home-time is defined as the number of days alive and living outside inpatient care. The registry specifies a mean, measured in days, over the 12 months post-discharge. Its stated population is the eligible group of acute ischemic stroke patients with Medicare fee-for-service coverage who were discharged to inpatient rehabilitation or skilled nursing care and had no in-hospital survival limitations identified.

As general statistical context, a mean gives the arithmetic average of a continuous measure across the people included in an analysis. A measure such as home-time can combine information about survival and time spent outside inpatient care into one summary. This can be useful when the research question concerns how much of the follow-up period participants were alive and living outside such care. It does not, by itself, show the distribution of individual outcomes: the same mean can arise from different patterns among participants.

The registry identifies standard deviation as the dispersion measure accompanying the mean. In general, a standard deviation describes how spread out individual values are around their mean; it is not a measure of uncertainty in the estimated mean. The registry record does not report the numerical mean or standard deviation, nor does it report a confidence interval for home-time. As a result, the endpoint definition and intended population can be described, but the observed level and precision of home-time cannot be assessed from the registry record.

Using a common follow-up horizon and a clearly defined measure can make an outcome easier to interpret across analyses. However, the endpoint does not establish that any rehabilitation service caused a particular outcome. The retrospective design and absence of reported group-level estimates limit causal and comparative interpretation.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The registry describes a retrospective analysis linking Get With the Guidelines-Stroke data with Medicare claims and AVAIL longitudinal data sets. It states that the analysis was intended to assess patterns, predictors, and outcomes associated with rehabilitation-service use after ischemic stroke hospitalization. No analysis-specific numerical results are reported in the registry record, and the record lists zero analyses.

What does a retrospective linked-data design mean?

As general statistical context, retrospective research analyzes data that were collected before the analysis was conducted, rather than assigning participants to study groups prospectively. Linking data sources can allow investigators to bring together information recorded in different settings or for different purposes. The accuracy and completeness of a linked analysis depend on the content and quality of those sources, how records are linked, and how variables and follow-up are defined. The registry identifies the sources but does not report the linkage procedures, validation results, or detailed variable definitions beyond the primary endpoint description.

Because rehabilitation service is not described as assigned by the study, service use may be associated with participant and care characteristics. As general statistical context, confounding occurs when characteristics associated with both the exposure or service and the outcome contribute to an apparent relationship. Statistical adjustment can account for measured characteristics included in a model, but does not by itself remove bias from unmeasured or poorly measured factors. The registry record does not report whether adjustment was performed or which covariates were included.

How should the mean and standard deviation be read?

The primary endpoint is a continuous measure expressed as days of home-time, summarized by a mean and standard deviation. In general, the mean describes the average value among the analyzed participants. The standard deviation describes variation among their individual values. These statistics answer different questions: the mean concerns the average outcome, while the standard deviation concerns how dispersed the observed outcomes are. Neither one alone establishes how precise an estimate of the population mean is; precision is commonly described using a standard error or confidence interval. Neither is reported for this outcome in the registry record.

As general statistical context, a confidence interval is a range calculated using a statistical method that reflects uncertainty in an estimate under the model’s assumptions. Its width depends on factors such as the variability in the data, the amount of information available, and the chosen method. A confidence interval would help readers judge how precisely the mean or a contrast was estimated. Since no interval or numerical mean is reported, the precision of home-time cannot be evaluated from the registry record.

What can be said about comparisons?

The eligible population includes patients discharged to inpatient rehabilitation or skilled nursing care. That description does not establish that the registry reports two analysis arms, a direct contrast, or comparable groups. The record itself notes that this is not a randomized comparison and that the arms are not a like-for-like control comparison. It does not state an allocation method. Without defined group estimates and an analysis description, it is not possible to characterize a contrast, its uncertainty, or the role of adjustment.

As general statistical context, random allocation can help balance both measured and unmeasured baseline characteristics on average, whereas observational comparisons rely on the data and analytic methods to address group differences. Even carefully adjusted observational estimates remain dependent on assumptions about measurement, model specification, and unmeasured confounding. Those are general considerations, not claims about the specific analysis beyond the nonrandomized nature recorded in the registry.

What do the sample-size illustrations represent?

The record describes planning illustrations based on the observed effect, with a two-sided significance level of 0.05, equal allocation between two arms, and a normal approximation. It says the effect to detect is a stated fraction of the observed effect. For time-to-event tables, the total number of events is the quantity described; for other tables, the quantity is participants per arm. Numerical illustration values and the observed effect are not reported in the registry record. These descriptions are planning assumptions, not observed power and not evidence that the original study was designed using those assumptions.

As general statistical context, a sample-size calculation estimates the amount of information needed to detect a prespecified effect under specified assumptions. It does not show that an effect exists, and a planning illustration does not establish the power of a completed analysis. The interpretation depends on the assumed effect, variability or event rate, design, allocation, and statistical method. Where the observed effect used for an illustration is not numerically reported, the numerical planning requirement cannot be independently interpreted from the registry record.

What methods are not described?

The registry record does not report the primary analysis model, covariate adjustment, handling of repeated measurements, missing-data method, sensitivity analyses, interim analyses, or a protocol-based sample-size justification. It does not report the estimate, standard deviation, confidence interval, or p-value for home-time. As general statistical context, each of these choices can affect how an estimate is calculated and interpreted; their absence from the record means those aspects of the analysis cannot be evaluated. No particular method should be assumed.

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("Inpatient Rehab Facility (IRF)", "Skilled Nursing Facility (SNF)")
n   <- c(34815, 34397)
m   <- c(271.2, 195.5)
sp  <- c(112.5, 138.5)
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  Inpatient Rehab Facility (IRF)  34815  271.200  112.500  270.018  272.382
CARM  Skilled Nursing Facility (SNF)  34397  195.500  138.500  194.036  196.964

11. Multiplicity, Missing Data and Analysis Populations

The record lists seven outcome measures and identifies home-time as the primary outcome. The named secondary outcomes include functional dependence or death, quality of life, nursing-home institutionalization, rehospitalization or death, and death. Numerical results for these outcomes are not reported in the registry record. The record does not describe a multiplicity strategy, such as adjustment of significance thresholds or a hierarchy for interpreting multiple outcomes.

As general statistical context, evaluating several outcomes can increase the chance that at least one result appears statistically notable by chance when each is tested separately. A prespecified hierarchy or multiplicity adjustment can help define how evidence across multiple tests should be interpreted. Without a reported plan and outcome-level results, it is not possible to determine how multiplicity was handled or whether it affected interpretation.

Missing-data methods are not reported in the registry record. The registry entries do not state how much outcome information was missing, how missing observations were handled, or whether sensitivity analyses were performed. As general statistical context, missingness can affect estimates when the probability of an outcome being observed is related to participant characteristics or outcome values. Different assumptions and handling methods can lead to different interpretations, so the method used matters.

The analysis populations are not described beyond the primary-outcome eligibility definition, which states that 69,212 patients met the criteria. The record does not provide a flow of participants through data linkage, exclusions, follow-up, or outcome availability. It also does not report whether analyses used all eligible patients or subsets for particular outcomes. These details are necessary to assess the extent to which results reflect the defined population, but numerical outcome results and analysis-population details are not reported in the registry record.

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

The registry record documents a completed retrospective analysis of rehabilitation-service use after hospitalization for ischemic stroke. It identifies linked data sources, a primary-outcome population, and home-time as the primary endpoint. Home-time is defined as days alive and living outside inpatient care over 12 months post-discharge. These details establish the study’s subject and the intended meaning of its primary outcome.

They do not establish an outcome level, an association, or a difference between care destinations. The numerical results for home-time and the listed secondary outcomes are not reported in the registry record. Therefore, the magnitude and precision of any estimate, the direction of any association, and the strength of statistical evidence cannot be described. No claim that rehabilitation services caused an outcome can be supported from the record’s design description alone.

As general statistical context, an observational association is not automatically a causal effect. Patients receiving different services may differ in ways related to their subsequent outcomes. Statistical adjustment may help account for measured differences, but causal interpretation also depends on design, measurement, and assumptions that cannot be evaluated without details of the analysis. The registry record does not report the models, covariates, or sensitivity analyses used.

The status marked completed means the record identifies the study as completed under its registry status field. The record describes primary completion as the end of primary analyses and study completion as acceptance of the final research report and lay abstract by PCORI. Completion status does not provide the missing estimates or show whether an analysis was statistically conclusive.

The study can be understood as addressing a relevant descriptive and observational question about rehabilitation services and outcomes after ischemic stroke, but its reported registry information is insufficient to draw conclusions about comparative outcomes. Readers should distinguish the defined research question and population from results that are not reported. The record supports describing the analysis design and outcome definitions; it does not support a conclusion about the effectiveness of one service relative to another.

13. Why This Trial Matters Statistically

Statistically, this study illustrates the distinction between a clinical question about service use and a randomized test of a treatment. The registry describes a retrospective analysis of linked data intended to examine patterns, predictors, and outcomes. That design can address questions about what happened in recorded care, but it does not assign rehabilitation services or, by itself, establish that a service caused an outcome.

The primary endpoint, home-time, combines survival and location of care into a measure expressed in days. As general statistical context, a composite-style measure can be useful when several aspects of a patient’s course matter to the research question. Its average can also conceal differing individual trajectories, so interpreting it alongside its variability and other outcomes is important. The registry specifies mean and standard deviation but does not report their values.

The stated eligible population is also important for statistical interpretation. It consists of patients with Medicare fee-for-service coverage who were discharged to inpatient rehabilitation or skilled nursing care and met the stated in-hospital survival criterion. A result from such a defined population would not automatically describe all people hospitalized with ischemic stroke. The record does not provide numerical outcome results or detailed analysis populations that would allow broader evaluation.

Finally, the record lists several outcomes but does not report how multiplicity, missing data, or confounding were handled. These are central considerations when analyzing linked observational data. The case therefore underscores that a clear endpoint and a large recorded enrollment do not replace transparent reporting of estimates, uncertainty, analytic choices, and population flow. Here, those numerical and methodological details are not reported in the registry record.

14. Limitations

• The posted summary statistics are the only data these results rest on; patient-level data, full protocols and statistical analysis plans are not available.
• Recalculated results are simple, unadjusted calculations from posted summaries. They do not replace the sponsor's prespecified analysis.

A key limitation for interpreting the record is that numerical outcome findings are not reported. The primary endpoint is defined, but its mean, standard deviation, confidence interval, and any test result are absent. The record also does not provide numerical results for the secondary outcomes. This prevents assessment of outcome magnitude, precision, direction, or statistical evidence.

The study is described as retrospective and nonrandomized, and the registry notes that the arms are not a like-for-like control comparison. This limits direct causal interpretation. As general context for observational research, service selection may be associated with patient characteristics and care circumstances. Whether and how those factors were addressed is not reported in the registry record.

The registry names the linked data sources but does not report linkage methods, data validation, variable construction beyond the primary endpoint definition, or the extent of missing information. It also does not describe participant flow from the total enrollment of 164246 to the primary-outcome population of 69,212, beyond stating the latter population’s eligibility criteria. The record does not provide analysis-population details for the secondary outcomes.

Methods that could affect interpretation are not reported, including the primary statistical model, covariate adjustment, handling of missing data, sensitivity analyses, and any multiplicity approach. A protocol-based sample-size justification is not reported. The sample-size assumptions described for planning illustrations do not supply numerical estimates or demonstrate the design used for the original study. These limitations concern what can be established from the registry record; they do not show that any particular method was or was not used in the full research work.

15. Related Tutorials

16. Related Statistical Calculators

17. Sources

• ClinicalTrials.gov: Comparative Effectiveness of Rehabilitation Services for Survivors of Acute Ischemic Stroke — NCT02284165.
• Registry record last updated 2018-03-01; status last verified 2018-02.
• Results first posted 2017-01-09.
• Data retrieved from ClinicalTrials.gov on 2026-11-09.
• Group sizes, means, event counts and sponsor-reported p-values are reproduced as posted. Effect estimates, confidence intervals and p-values not attributed to the sponsor are independently calculated from the posted summary statistics.

18. Record Summary

This completed record concerns a retrospective analysis of rehabilitation services following hospitalization for ischemic stroke, sponsored by Duke University. Its stated data sources are Get With the Guidelines-Stroke, Medicare claims, and the AVAIL longitudinal data sets. The registry reports total enrollment of 164246 and gives a primary-outcome population definition that 69,212 patients met.

The primary outcome is mean home-time in days over 12 months post-discharge, defined as days alive and living outside inpatient care. The record lists seven outcome measures and describes secondary outcomes involving function, quality of life, institutionalization, rehospitalization, and death. Numerical outcome findings and analysis-specific results are not reported in the registry record.

The study is marked completed, with recorded start 2013-05 and primary completion 2016-05. It is not a randomized comparison, and the registry notes that the arms are not a like-for-like control comparison. The outcome definitions and design can be described, but the comparative results and their precision cannot be assessed from the registry record.