Vanderbilt ICU Recovery Program Pilot Trial: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Vanderbilt ICU Recovery Program Pilot Trial (NCT03124342), a randomized na study. The primary outcome is Number of Components of the ICU Recovery Program Received.

2. Clinical Question

3. Trial Design

The registry record describes a randomized, single-masked, parallel-group trial sponsored by Vanderbilt University Medical Center. The study is listed as completed, with a phase of NA. The reported enrollment was 232. The recorded study start was 2017-05-01, and primary completion was 2017-10-31. The registry record does not identify who was masked, so the meaning of single masking cannot be specified further.

Participants were assigned to the Vanderbilt ICU Recovery Program (VIP) or usual care. The posted outcome results include 111 participants in the VIP group and 121 in the usual-care group for the primary outcome and the listed binary secondary outcomes. These outcome denominators describe the reported results; the registry record does not explain whether they represent every randomized participant or an analysis subset.

The program was multidisciplinary. Its components included inpatient contacts and resources, medication reconciliation, and services or assessments in the ICU Recovery Clinic. The primary outcome counted how many components participants received during the stated period. The control description indicates that usual care was determined by the clinical team and gives examples of medication reconciliation, but the record’s description ends before detailing usual care fully.

The brief summary frames the study as a pilot focused first on feasibility and process measures, including contact with the recovery team and clinic attendance. It also states secondary aims concerning 30-day same-hospital readmission and other clinical outcomes. This design therefore combines a measure of program delivery with exploratory clinical outcomes. A protocol, detailed allocation procedure, sample-size justification, and full statistical 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.

Overall Study

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 Number of Components of the ICU Recovery Program Received, measured over From the time of study enrollment to 30 days after hospital discharge.

The primary outcome was the number of components of the ICU Recovery Program received, measured from study enrollment to 30 days after hospital discharge. The registry defines ten components, spanning nurse practitioner visits, educational material and a contact line, pharmacist medication reconciliation, clinic assessments and consultations, case management, and referrals. The outcome is a count of delivered program elements, not a direct measure of health status or a comparison of clinical events.

The registry reports the outcome as a median with an interquartile range. The VIP group had a median of 2 components, with an interquartile range of 1 to 3. Usual care had a median of 1 component, with an interquartile range of 0 to 1. These summaries describe the distribution of program components received in each group. The record does not report a formal statistical comparison, confidence interval, or p-value for this primary outcome.

A component count is useful in a pilot because it describes whether the intended services reached participants and can identify where implementation occurred. It does not, by itself, establish whether participants experienced better health or fewer hospital visits. Nor does the group-level median reveal how often each individual component was delivered. Component-specific results are not reported in the registry record.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The primary outcome is a count of program components, summarized with a median and interquartile range. As a general statistical concept, the median identifies the middle of an ordered distribution, while the interquartile range describes the central spread of observations. These summaries can be informative when counts are skewed or have many low values. They do not describe every feature of the distribution, and the median difference is not itself a formal treatment-effect estimate unless an analysis is specified.

The registry reports a median of 2 components with an interquartile range of 1 to 3 for VIP, and a median of 1 with an interquartile range of 0 to 1 for usual care. The outcome definition counts ten possible program components. The record does not provide component-by-component uptake, participant-level distributions, or a statistical test of the primary outcome. It is therefore possible to describe the posted group summaries, but not to reconstruct a formal comparison from the registry entry.

The secondary outcomes are counts of participants with binary events. For these outcomes, the registry posts a risk difference comparing the VIP group with usual care, along with a confidence interval and unadjusted p-value. As a general explanation, a risk difference expresses the difference between the event proportions in two groups. A negative value, using the comparison ordering in the registry, points to a lower event proportion in VIP; a value near zero indicates little estimated difference. A confidence interval describes uncertainty around the estimate under the assumptions of the analysis. When the interval includes zero, the data are compatible with no risk difference as well as with values on either side of zero.

The readmission result has a confidence interval that includes zero. The death-or-readmission result has an interval that does not include zero. The emergency department result has an interval that includes zero. These distinctions concern the posted unadjusted comparisons; they do not establish that the analyses were prespecified or that their assumptions were checked. The registry record does not report an adjustment model, covariates, exact testing procedure, or analysis population definition.

The unadjusted p-values are 0.162 for same-hospital readmission, 0.037 for death or readmission, and 0.951 for emergency department visits. In general, a p-value measures how unusual data at least as incompatible with a specified null hypothesis would be under that hypothesis and the chosen statistical model. It is not the probability that the null hypothesis is true, nor a measure of clinical importance. Because these p-values are unadjusted and multiple secondary outcomes were reported, they should not be interpreted as if only one outcome had been tested.

The registry also gives relative risk and odds-ratio summaries for the secondary outcomes. In general, relative risk compares event proportions, while an odds ratio compares event odds. Both are alternative ways to describe binary outcomes, but they can be unstable when events are sparse. The emergency department outcome has one event in each group, and its relative-risk and odds-ratio intervals are very wide. This illustrates a general issue: sparse outcomes can yield estimates that are highly uncertain, even when a point estimate is available. The registry does not identify these relative measures as results from a sponsor-reported analysis.

The posted planning illustrations use the observed effect as a basis for hypothetical future sample sizes. They assume a two-sided significance level of 0.05, equal allocation, and a normal approximation; they show participant counts per arm for 80% and 90% power using 100%, 75%, or 50% of the observed effect. For readmission, these counts range from 461 to 1,986 participants per arm for 80% power, and from 617 to 2,658 for 90% power. For death or readmission, the corresponding ranges are 207 to 894 and 276 to 1,197. For emergency department visits, the ranges are 242,457 to 949,109 and 324,581 to 1,270,586. These are planning illustrations based on observed results, not the trial’s design justification and not observed power. The registry record does not report the original sample-size rationale.

As a general statistical point, planning from an observed effect can be uncertain, especially in a pilot with limited events. A future design would ordinarily need an explicit target effect and assumptions about the outcome rate, allocation, analysis, and desired precision. Those design choices cannot be recovered from the posted planning illustrations. No sample-size assumptions for the primary component-count outcome are reported in the registry record.

10. Multiplicity, Missing Data and Analysis Populations

The registry record lists three secondary outcomes with posted group comparisons and does not report a multiplicity strategy. It does not state whether the outcomes were ordered in a hierarchy, whether one was designated as the main secondary endpoint, or whether p-values or confidence intervals were adjusted for multiple comparisons. This matters because interpreting each unadjusted result as an independent confirmatory finding can overstate the evidence when several outcomes are examined. The composite result is therefore best viewed as nominal and exploratory unless a prespecified analysis plan says otherwise; such a plan is not reported in the registry record.

Missing-data handling is not reported in the registry record. The posted outcomes include group denominators, but the record does not say whether all randomized participants had complete follow-up, whether any outcomes were unavailable, or what assumptions and procedures were used for missing observations. It also does not report imputation, sensitivity analyses, or reasons for missingness. The reported denominators alone cannot establish that there were no missing data or that missingness had no effect on the estimates.

The analysis populations are also not described. The registry record does not say whether the results followed an intention-to-treat approach, included only participants with observed outcome data, or used another population definition. It does not give details of protocol deviations, crossover, or participants excluded from analyses. Randomization is recorded, but the analysis strategy used to preserve or assess the benefits of randomization is not reported.

The record lists 0 reported analyses. That does not replace a full account of statistical methods; it means the posted registry information does not describe additional reported analyses. The registry record also does not report a detailed statistical analysis plan, any interim analyses, or adjustments for baseline factors. Without those details, the secondary p-values and estimates can be described as posted, unadjusted comparisons, but their relationship to prespecified hypotheses and analysis choices remains uncertain.

11. 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.

The registry’s safety information records 0 serious adverse events and 0 other adverse events in the VIP group, with 111 participants at risk in each category. In usual care, it records 0 serious adverse events and 0 other adverse events, with 121 participants at risk in each category. These entries describe what is recorded in the safety fields; they do not establish that no symptoms or events of any kind occurred, since the definitions, collection procedures, and reporting thresholds are not detailed.

The safety record also lists 7 deaths in VIP and 10 in usual care. The registry does not state the timeframe or provide an attribution for these death counts in the safety entry. Separately, the secondary clinical outcome counts participants with death or readmission within 30 days after hospital discharge. Those are different entries and should not be assumed to share an identical definition or observation period.

Zero recorded adverse-event counts do not demonstrate equivalence in safety between groups. The available entry is limited, and the registry does not provide event descriptions or a comparative safety analysis. These figures should be read as registry reporting, not as evidence that the program caused or prevented an event.

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

The posted results describe a pilot trial in which participants assigned to VIP received a median of 2 program components, compared with a median of 1 in usual care. This is evidence about program delivery as summarized by the registry. It does not establish whether all intended services were delivered consistently, which individual components were most often received, or how participants experienced them.

The clinical outcome comparisons are less conclusive as a whole. The unadjusted readmission and emergency department comparisons have confidence intervals that include zero. The death-or-readmission comparison has a nominal unadjusted p-value of 0.037 and a confidence interval that excludes zero, but several secondary outcomes are reported and no multiplicity plan is described. The results therefore do not provide a definitive answer about the program’s effect on post-discharge clinical outcomes.

Statistical uncertainty is not the same as evidence that groups are identical. A confidence interval that includes zero means the data are compatible with no difference under the analysis, while also allowing differences in either direction. Conversely, an interval that excludes zero does not by itself establish clinical importance, robustness, or causality. The registry does not report enough detail to assess all modeling and follow-up assumptions.

The trial’s size and pilot purpose also place the findings in context. The record frames feasibility and process measures as the primary focus, with clinical outcomes as secondary aims. The registry does not report a sample-size justification or a complete statistical analysis plan. The planning illustrations based on observed effects are not a substitute for an original design rationale and do not indicate the power achieved by the completed study.

Accordingly, the results can inform discussion of implementation and help identify outcomes for further evaluation. They do not establish that VIP improves or worsens health outcomes, nor do they establish that usual care and VIP are equivalent. Any broader interpretation would require fuller methods, outcome reporting, and evidence from an appropriately planned evaluation.

13. Why This Trial Matters Statistically

This trial is statistically useful because its primary outcome reflects implementation rather than a conventional clinical endpoint. Counting received program components makes delivery visible and helps distinguish the question of whether a complex service was implemented from the question of whether clinical outcomes changed. That distinction is important when evaluating multidisciplinary programs, where the intended intervention consists of multiple activities rather than one isolated treatment.

The registry also illustrates how a pilot can produce clinical estimates that are interesting but uncertain. The binary secondary outcomes have different event frequencies and therefore different levels of precision. In particular, sparse outcomes can generate very wide relative-measure intervals. A point estimate alone can conceal this uncertainty, so reading the interval alongside the estimate and event counts is essential.

The contrast between the primary implementation measure and the secondary clinical outcomes is another important statistical lesson. Delivery of more program components and changes in clinical events are not interchangeable outcomes. The posted group summaries do not show whether receiving a particular component is associated with an individual participant’s later outcome, and they do not establish a mechanism connecting implementation to readmission or death.

Finally, the record exposes the limits of registry results when analysis details are absent. Without a reported sample-size rationale, analysis population, missing-data method, or multiplicity plan, readers cannot fully evaluate how the clinical comparisons were designed and interpreted. The study can still contribute descriptive evidence about feasibility and inform future research questions, while the uncertainty around its secondary outcomes cautions against treating one nominal result as a definitive conclusion.

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.

Several limitations specific to this trial affect interpretation. The primary endpoint records the number of program components received, but the registry does not post component-level results. A median count cannot show whether participants received the same combination of services or whether some components were rarely delivered. The endpoint also does not directly measure recovery, function, or quality of life, despite those outcomes appearing in the study rationale.

The usual-care description is incomplete in the registry record. It says care was directed by clinical teams and begins to describe medication reconciliation, but the text is truncated. The extent to which usual care overlapped with VIP cannot be established from the record. The masking information is also limited: the trial is labeled single-masked, but the masked party is not identified.

The posted clinical comparisons are unadjusted, and no reported analysis is listed. The record does not explain the analysis population, baseline balance, missing-data handling, or multiplicity approach. Those omissions limit the ability to judge whether estimates reflect the randomized groups as intended and how much confidence to place in secondary comparisons. The record also does not provide results for the listed outpatient clinic visits outcome.

Event counts are limited for some outcomes, particularly emergency department visits. Estimates from sparse outcomes are necessarily imprecise, and the registry’s relative-measure intervals for that outcome are very wide. The safety entry records deaths but does not state their timeframe or attribution. These details constrain interpretation without implying that any unreported event or analysis occurred.

Finally, the registry does not report a sample-size justification, a full protocol, or a detailed statistical analysis plan. The planning illustrations use observed effects and are not the original design rationale. These limitations make the trial most informative as a feasibility-focused pilot with exploratory secondary outcomes, rather than as a definitive comparative-effectiveness evaluation.

15. Related Tutorials

16. Related Statistical Calculators

17. Sources

• ClinicalTrials.gov: Vanderbilt ICU Recovery Program Pilot Trial — NCT03124342.
• Registry record last updated 2019-10-08; status last verified 2019-10.
• Results first posted 2019-10-08.
• 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

NCT03124342 is the Vanderbilt ICU Recovery Program Pilot Trial, sponsored by Vanderbilt University Medical Center and listed as completed. It was randomized, single-masked, and parallel-group, with 232 participants enrolled. The recorded start was 2017-05-01, and primary completion was 2017-10-31.

The primary outcome counted ICU Recovery Program components received from enrollment to 30 days after hospital discharge. The posted median was 2 (interquartile range 1 to 3) for VIP and 1 (interquartile range 0 to 1) for usual care. Secondary outcomes include same-hospital readmission, death or readmission, emergency department visits, and outpatient clinic visits. The registry posts unadjusted comparisons for the first three, but not a result for outpatient clinic visits.

The registry record does not report a full statistical analysis plan, sample-size justification, missing-data methods, analysis populations, or multiplicity approach. The primary result is descriptive, and the secondary clinical comparisons should be interpreted cautiously.