Improving Transplant Medication Safety Through A Technology and Pharmacist Intervention: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Improving Transplant Medication Safety Through A Technology and Pharmacist Intervention (NCT03860818), a randomized na study. The primary outcome is Any Hospitalization or Any Emergency Room Visits.

2. Clinical Question

3. Trial Design

This is a randomized, parallel, cluster-randomized clinical trial in Veteran organ transplant recipients. The registry lists masking as none and the phase as not applicable. Enrollment is recorded as 2196. The sponsor is the VA Office of Research and Development. The intervention is described as a technology-enabled pharmacist intervention, and the control arm receives usual care.

The intervention’s stated components are electronic medical record technology to identify Veterans with potential medication-safety issues and pharmacist-led support for management and coordination of immunosuppression therapy. The summary describes the study as prospective and multicenter. The registry does not state which entities were randomized as clusters, how many clusters participated, how cluster allocation was carried out, or whether cluster sizes differed.

The recorded study start date is 2019-03-18, and the primary completion date is 2021-06-14. The study status is completed. The primary outcome timeframe is 24-months. The registry lists secondary outcomes covering estimated health care costs, graft survival, patient survival, medication-safety issues, clinically relevant alerts, dashboard-accepted interventions, and time responding to dashboard alerts. It records 8 outcome measures and 5 analyses.

The record does not report a sample-size justification, the method used to generate the random allocation sequence, details of allocation concealment, or a statistical analysis plan. It also does not report the cluster-level enrollment or the extent to which participants or clinicians knew the assigned group. The absence of masking is reported, but the registry does not specify which roles were unmasked.

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

Reasons for not completing this period:

5. Baseline Characteristics

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

6. Primary Endpoint

The registry lists the primary outcome as Any Hospitalization or Any Emergency Room Visits, measured over 24-months.

The primary endpoint is titled “Any Hospitalization or Any Emergency Room Visits.” Its timeframe is 24-months, and the registry describes it as comparing hospitalization and emergency room visits between the intervention and usual-care groups. The endpoint is coded as a count of participants, with the recorded outcome type meaning participants with an event. This points to whether a participant had an event, rather than how many visits an individual made.

That distinction matters for interpretation. A participant-level event outcome treats having any qualifying hospitalization or emergency room visit as the relevant status. It does not, on its own, express the total burden of repeated visits. The registry title joins hospitalization and emergency room visits, but it does not give separate primary-outcome counts for either component or explain how a participant with both types of visit is handled.

Hospital use is a direct, observable clinical outcome and can reflect the consequences of health status and care delivery. In this trial, it is relevant to the stated aim of improving medication safety and coordination for transplant recipients. That rationale does not establish what the endpoint result means clinically; the record’s primary-outcome entry does not report group-specific event counts, an effect size, or a confidence interval. The two posted primary analyses give p-values, but without an estimate and its precision the size and direction of the group difference cannot be determined from those entries.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

What does the chi-squared method assess?

In general, a chi-squared test can assess whether a categorical outcome is distributed differently across groups. For a participant-level event outcome, the relevant information is ordinarily the number of participants with and without the event in each group. The test evaluates compatibility between the observed group pattern and a null model of no association or no group difference, depending on how the analysis is specified. The registry names the chi-squared method but does not provide the underlying primary-outcome counts or further test specifications.

A p-value describes how incompatible the observed data are with a specified null model, assuming the model and analysis assumptions apply. It is not the probability that the null hypothesis is true, and it does not measure the size or importance of an effect. The two reported values, p < .001 and p < .0001, are both below the stated planning illustration’s two-sided significance level of 0.05. That comparison alone does not resolve the missing information about effect direction, size, precision, or multiplicity.

Why does cluster randomization matter?

In general, cluster randomization assigns groups of people or care units rather than assigning each participant independently. People within a cluster may share clinicians, workflows, or other influences, so their outcomes can be more alike than outcomes from people in different clusters. A method that treats every participant as fully independent can then understate uncertainty. Proper analysis generally needs to account for the clustering and the way the clusters were assigned.

The registry identifies this as a cluster-randomized trial and cautions that a simple individual-level re-analysis would be misleading. It does not identify the cluster unit, cluster count, or the adjustment used in the sponsor’s analysis. Those details matter for evaluating the assumptions and precision of a cluster-randomized comparison. The available record therefore supports describing the sponsor’s reported analysis, but not reconstructing a cluster-adjusted result from individual-level summaries.

What can and cannot be learned from the p-values?

The registry reports p-values without a primary-outcome effect estimate or confidence interval. In general, an effect estimate communicates the direction and size of a group contrast, while a confidence interval conveys the range of values compatible with the data under the analysis method. A p-value does not supply either piece of information. Because the group event counts and estimate are not reported, the primary result cannot be translated into an absolute difference or another effect measure from this record.

The entries list two chi-squared analyses for the same groups, both labeled superiority. The registry does not explain whether these represent alternative specifications, different analysis populations, or another distinction. They should not be averaged or otherwise combined. The p-values should be quoted as posted, with the lack of explanation noted rather than resolved by inference.

How should multiplicity and missing data be considered?

In general, testing multiple outcomes or conducting multiple analyses can increase the chance of obtaining at least one result that appears noteworthy under a null model. Statistical plans may address this with a prespecified hierarchy or an adjustment, but neither approach is described in the registry entry. The record lists 8 outcome measures and 5 analyses; it does not report which analyses correspond to which outcomes, whether the primary analysis was prespecified in a statistical plan, or whether any multiplicity procedure was used.

Missing data can also affect an endpoint comparison if outcome status is not observed for all eligible participants. General approaches include defining an analysis population and specifying how unavailable outcomes are handled. The registry record does not report missing primary-outcome data, a missing-data method, or the analysis population. Without those details, the impact of missingness on the reported primary analysis cannot be assessed from this entry.

What do the planning assumptions represent?

The registry describes a planning illustration based on the observed effect, not the trial’s own design. Its stated assumptions are a two-sided significance level of 0.05, equal allocation between arms, and a normal approximation; the effect to detect is expressed as a fraction of the observed effect. For time-to-event tables, the stated output is the total number of events using Schoenfeld; other planning outputs are participants per arm. Such illustrations are planning scenarios, not observed trial outcomes or observed power. The record does not provide numerical planning outputs in the registry record for this analysis.

The design’s cluster randomization is important when considering any planning calculation: analyses and sample-size methods may need to account for within-cluster similarity. The assumptions listed for the planning illustration do not by themselves describe how clustering was incorporated. That detail, along with the trial’s own sample-size justification, is not reported in the registry record.

10. Analyses Reported by the Sponsor

The registry presents the sponsor’s own analyses and explicitly cautions that a simple individual-level re-analysis would be misleading because the trial is cluster-randomized. As a result, there is no appropriate unadjusted re-calculation from individual-level summary statistics to compare with the sponsor’s findings. The posted sponsor analyses should take precedence over an individual-level re-analysis.

The sponsor’s analysis entries use the chi-squared method and report p < .001 and p < .0001 for the control and intervention groups. Both are labeled as superiority analyses. The registry does not report an effect estimate, confidence interval, or primary-outcome group counts alongside these values. Nor does it explain why there are two entries or whether they reflect different analytic choices. The p-values therefore cannot be used to infer the direction or magnitude of the result.

There is no reported re-calculated estimate in the record that can be compared numerically with the sponsor’s analysis. This is not evidence of agreement or disagreement between methods; it reflects the stated decision to show the sponsor’s analysis rather than conduct a potentially misleading individual-level re-analysis. Assessment of the sponsor’s cluster-adjusted method would require analysis details not reported in the registry record.

11. Multiplicity, Missing Data and Analysis Populations

The registry lists 8 outcome measures and 5 analyses, including the primary endpoint and secondary outcomes covering costs, survival, medication-safety issues, alerts, accepted interventions, and response time. The record does not map each analysis to an outcome or explain whether all analyses were planned in advance. It also does not state whether the primary outcome was designated within a prespecified testing hierarchy.

When several outcomes or analyses are examined, multiplicity is a general statistical concern: without an appropriate plan, the chance of at least one apparently strong result can increase. Common approaches include adjusting the interpretation or setting an order for testing, but the registry does not report use of such a method. The two p-values for the primary comparison are not explained, so their relationship and any role in multiplicity cannot be determined. They should be retained as separate reported analyses rather than combined or selected between without further information.

The record does not report missing primary-outcome data, the number of participants included in the primary analysis, the analysis population, or a method for handling missing outcomes. It therefore cannot establish whether all enrolled participants contributed to the primary analysis or how incomplete follow-up affected the comparison. The timing of outcome ascertainment within the 24-month timeframe beyond the stated timeframe is not described.

The registry also does not report interim analyses, stopping rules, or a statistical analysis plan. These are not grounds to assume that such procedures were or were not used; their status is not reported in the registry record. Together, the lack of multiplicity, missing-data, and analysis-population details makes it difficult to evaluate the full inferential context of the posted p-values.

12. 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

Serious events with the highest arm-level rate: inpatient hospitalization (42.4%), graft failure (3.6%).

The registry’s adverse-event figures list the control arm with 896 at risk and 412 affected, and the intervention arm with 1300 at risk and 436 affected. It also lists deaths as 144 for the control arm and 195 for the intervention arm. These are adverse-event record figures, not the primary hospitalization-or-emergency-room endpoint analysis. The registry entries do not clarify the relationship between these figures and the primary endpoint counts.

The most frequently listed serious adverse-event terms are “Inpatient hospitalization” at 42.4% and “Graft Failure” at 3.6%. The registry information given does not specify the denominator or explain whether these percentages represent participants, events, or another reporting basis. They should be read as the listed serious-event terms and percentages, not converted into participant risks or treated as a comparison between arms.

These figures do not establish that an event was caused by the intervention, nor do they provide a full account of severity, timing, or exposure. The adverse-event entry does not report a comparative statistical analysis or confidence intervals. In particular, the listed deaths are not accompanied by a time-to-event analysis or further explanation of cause. The safety figures therefore describe what the registry lists, while leaving key details for comparative interpretation not reported in the registry record.

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

The posted results support a limited statistical statement: the registry records two sponsor chi-squared analyses, both labeled superiority, with p-values of p < .001 and p < .0001. The primary outcome concerns participants with any hospitalization or emergency room visit over 24-months. The record does not provide the primary outcome counts by group, an effect estimate, or a confidence interval. It is therefore not possible to determine from these entries which group had fewer events, how large any difference was, or how precise an estimate would be.

A small p-value is not the same as a large or important treatment effect. It is also not a direct measure of the probability that the intervention is beneficial. The result should not be read as proof of a specific mechanism or as evidence about outcomes not reported in the primary analysis. The primary endpoint records whether participants experienced an event; it does not, as described, quantify each person’s number of visits.

The trial’s cluster-randomized design affects how the analysis should be interpreted. The registry cautions against a simple individual-level re-analysis and presents the sponsor’s analyses. Without further details of the cluster-adjusted methods, the assignment unit, and the analysis population, independent assessment of the analysis assumptions is limited.

The trial summary describes an intervention intended to improve medication safety through electronic medical record identification of potential issues and pharmacist-led coordination of immunosuppression therapy. The registry entry does not establish from the primary outcome alone whether medication-safety issues changed, whether costs changed, or whether graft or patient survival differed. Those are listed among the secondary outcomes, but their results are not reported in the registry record for this analysis.

Accordingly, the result does not justify a broad conclusion about effectiveness or safety. It describes the reported statistical tests and their limitations. Interpreting clinical importance requires outcome estimates, uncertainty, and enough methodological detail to understand how the cluster-randomized comparison was conducted.

14. Why This Trial Matters Statistically

This trial is statistically notable because its randomized comparison is cluster-based, while the primary endpoint is recorded at the participant level. That combination makes the unit of assignment and the unit of outcome distinct. In general, outcomes within a cluster may be correlated, so treating all participants as independent can misrepresent uncertainty. The registry’s warning against an individual-level re-analysis is therefore central to reading the result, not a minor technical qualification.

The trial also connects a technology-enabled process for identifying potential medication-safety issues with pharmacist-led coordination. The primary outcome—any hospitalization or emergency room visit over 24-months—is a broader clinical-use outcome, rather than a direct measure of the process of finding and responding to alerts. The listed secondary outcomes span costs, survival, medication-safety issues, and dashboard activity. This range makes it especially important to distinguish the primary question from the other outcomes and to know how multiplicity was handled; that information is not reported in the registry record.

Finally, the record illustrates why p-values alone are an incomplete account of a trial. The posted analyses have small p-values, but there is no reported primary effect estimate, confidence interval, or group-specific event count in the analysis entries. Readers can recognize that the sponsor reports evidence against a null hypothesis while still being unable to assess effect magnitude, direction, or precision. The design and reporting details needed for a fuller evaluation are equally important to the statistical result.

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.

Several trial-specific details needed to assess the primary comparison are not reported in the registry record. The cluster unit, number of clusters, cluster-level allocation process, and cluster adjustment used in the sponsor’s analysis are not described. These are important because the record identifies a cluster-randomized design and cautions that a simple individual-level re-analysis would be misleading.

The primary analysis entry does not provide group-specific event counts, an effect estimate, or a confidence interval. It also does not explain why two chi-squared superiority analyses are listed for the same groups or how their results differ. Without those details, direction, magnitude, precision, and the exact relationship between the analyses remain unclear.

The endpoint description and coding emphasize participants with an event, while the title combines hospitalization and emergency room visits. The record does not report separate component results or describe how the combined outcome was constructed. It does not provide the primary analysis population, missing-data handling, or follow-up completeness. The relationship between the adverse-event figures and the primary endpoint is also not reported.

The registry lists multiple outcomes and analyses but does not describe a statistical analysis plan, a multiplicity approach, interim analyses, or a sample-size justification. The planning illustration is described as based on the observed effect rather than the trial’s own design; numerical outputs for that illustration are not reported in the registry record for this analysis. These omissions do not establish that the procedures were absent, but they limit what can be concluded from the registry entry.

Finally, the summary gives the intervention’s intended components but not enough implementation detail to assess how consistently technology alerts or pharmacist activities were delivered. The record does not report adherence, contamination, or the extent of usual-care variation. Those considerations may affect interpretation of a pragmatic care-delivery comparison, but their status is not reported in the registry record.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: Improving Transplant Medication Safety Through A Technology and Pharmacist Intervention — NCT03860818.
• Registry record last updated 2024-08-01; status last verified 2024-05.
• Results first posted 2024-08-01.
• Data retrieved from ClinicalTrials.gov on 2026-11-09.
• Group sizes, means, event counts and sponsor-reported p-values are reproduced as posted. Effect estimates, confidence intervals and p-values not attributed to the sponsor are independently calculated from the posted summary statistics.

19. Record Summary

NCT03860818, “Improving Transplant Medication Safety Through A Technology and Pharmacist Intervention,” is recorded as completed. It was a randomized, parallel, cluster-randomized trial with no masking, sponsored by the VA Office of Research and Development. The registry records enrollment as 2196, a start date of 2019-03-18, and primary completion on 2021-06-14.

The intervention arm received a technology-enabled pharmacist intervention, compared with usual care. The primary outcome was any hospitalization or emergency room visit over 24-months, recorded as participants with an event. The sponsor’s chi-squared superiority analyses report p < .001 and p < .0001. The registry does not report primary-outcome group counts, an effect estimate, or a confidence interval, and cautions that a simple individual-level re-analysis would be misleading because of cluster randomization.

Secondary outcomes include costs, graft and patient survival, medication-safety issues, alert and intervention measures, and response time. Further analysis and reporting details are not reported in the registry record.