Implementation Outcome Assessments of the Emergency Department Delirium Screening and Detection Program: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Implementation Outcome Assessments of the Emergency Department Delirium Screening and Detection Program (NCT05606328), a randomized na study. The primary outcome is Accurate Delirium Screening by Emergency Department (ED) Nurses (RE-AIM: Efficacy).

2. Clinical Question

3. Trial Design

The registry describes a pilot stepped wedge cluster randomized trial of the ED-DDP across three diverse emergency departments, with a second aim assessing implementation outcomes using mixed methods. The study is marked randomized, has no masking, and lists a sequential model. These registry descriptions are compatible with a site-level rollout in which intervention timing differs by site, but the detailed randomization sequence and allocation procedures are not reported in the registry record.

The program includes a multicomponent, one-day delirium champion workshop, direct observation and training of champions via telehealth, practical training of nurses by champions, and patient chart review. During the implementation period, ED leadership at each site identifies 5-10 champions, including nurse educators and managers and bedside nurses. The implementation period is described as lasting 3 months. The intervention period is described as beginning 3-9 months after program implementation and extending over a 15 month period, depending on when a site is randomized to receive the intervention.

Enrollment is listed as 18. The primary measure separately describes a random sample of N=30 nurses, 10 at each ED site, for real-time screening observation, and says site nurses were not considered enrolled in the study for that measure. These are distinct registry descriptions; the record does not explain their relationship. The study is marked completed, with start date 2023-02-28 and primary completion date 2024-01-12. The registry record does not report the sequence of site transitions, the timing of observations at each site, or the handling of time effects in analysis.

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.

Months 1-3

Months 4-6

Reasons for not completing this period:

Months 7-9

Months 10-12

Analysis population for the primary outcome, as defined in the registry: Site nurses were not considered enrolled in the study but were assessed for this measure.

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: Nurse champions enrolled in the ED-DDP program.

6. Primary Endpoint

The registry lists the primary outcome as Accurate Delirium Screening by Emergency Department (ED) Nurses (RE-AIM: Efficacy), measured over Intervention period (3-9 months post program implementation).

The primary endpoint is titled accurate delirium screening by emergency department nurses. It measures the proportion of nurses who meet the specified accuracy threshold when using the bCAM tool, assessed through tele-observation and a validated Spot Check Form. The endpoint description defines a nurse as meeting the criterion if the nurse scores at least 80% accuracy on bCAM use. The observed nurses are described as a random sample of N=30, 10 at each of the three ED sites.

This endpoint focuses on whether training translates into accurate use of a screening procedure in practice, rather than merely whether staff attended training. That makes it relevant to the program’s efficacy aim: a screening program depends on staff applying the tool accurately. The registry also sets a success criterion: the computed proportion of nurses using the bCAM accurately is 80% or higher.

The record does not report the number of nurses who met the threshold, the resulting proportion, or its precision. It also does not report the accuracy scores themselves, the number of observations per nurse, or how incomplete or unscorable observations were handled. The nurse-level endpoint should not be read as a patient-level delirium detection rate.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

What does a stepped wedge cluster randomized design mean?

In a cluster randomized study, groups rather than individual participants are assigned to intervention conditions. Here, the registry describes three emergency departments as the sites in a pilot stepped wedge cluster randomized trial. In a stepped wedge design, sites cross from a comparison condition to the intervention at different times, with the sequence randomized. This structure can allow each site to contribute observations during different stages of rollout, but it also means that intervention status and calendar time can be linked.

General statistical practice for this design accounts for the possibility that measurements from people in the same site are correlated. Staff in one ED share a work setting and may have similar training or procedures. Treating all observations as unrelated can make uncertainty appear smaller than it is. Statistical models may also need to account for time, since outcomes can change for reasons unrelated to the program. The registry record does not report the analysis model, the site sequence, or how time and clustering were handled.

How is the primary outcome defined?

The outcome is the proportion of observed nurses who reach at least 80% accuracy on bCAM use, based on a validated Spot Check Form. The description states that tele-observations will assess screening accuracy and that a random sample of N=30 nurses, 10 at each ED site, will be observed in real time. The nurse is the unit represented by the proportion: each nurse is classified according to whether the accuracy threshold is met.

A proportion is a summary of how many assessed units satisfy a definition among the units assessed. Its interpretation depends on the definition of success, how observations are collected, and whether the assessed units represent the intended population. The registry gives the accuracy threshold and sampling plan, but does not report the number meeting the threshold, an uncertainty interval, or details such as repeated observations per nurse. The record also does not state how missing or incomplete observations would affect the proportion.

What is the role of the stated success criterion?

The registry says a successful outcome will be considered if the computed proportion of nurses accurately using bCAM is 80% or higher. This is a prespecified-looking criterion in the endpoint description, but the record does not report a measured proportion or whether the criterion was met. A threshold for success is not itself a result. It also does not communicate precision: an estimate close to a threshold can have uncertainty, and a proportion without an interval does not show how stable the estimate may be.

General statistical interpretation distinguishes an observed estimate from the uncertainty around it. A confidence interval is a range produced by a statistical procedure to express uncertainty about an estimate under the model’s assumptions. No confidence interval is reported in the registry record for this endpoint, so the precision of any observed nurse proportion cannot be assessed from the record.

Why does clustering matter?

Clustered data can show over-dispersion relative to a model that assumes independent observations. In general, over-dispersion means that the observed variability is greater than the model expects. Shared site conditions can contribute to this pattern. Accounting for clustering affects uncertainty estimates and potentially the interpretation of comparisons. The trial’s cluster structure is described, but the registry record does not report a variance model, an intracluster correlation estimate, or an adjustment method.

Simple recalculation from aggregate summaries may be inappropriate when the randomized units are sites and the design is stepped wedge. The record itself indicates that an individual-level re-analysis would be misleading and that only the sponsor’s own analysis is shown. No sponsor analysis is listed in the posted results, so there is no analysis estimate to compare with a recalculation.

How should implementation outcomes be understood?

The secondary outcomes cover the proportion of delirium champions completing training, the proportion of ED nurses completing training, the proportion of nurses screening eligible patients, the proportion of program components adhering to protocol, and the time required to train champions and nurses. These address implementation dimensions such as reach, adoption, fidelity, and time. The brief summary also describes a mixed methods assessment using training logs, tele-observation, interviews with champions and nurses, and electronic medical record screening.

These outcomes are conceptually different from the primary accuracy endpoint. Training completion measures participation; adoption concerns whether screening is used; fidelity concerns whether program components follow protocol; and training time describes resource use. None substitutes for another, and none alone establishes a patient-level effect. The registry lists these outcome concepts but does not report their findings, measurement details beyond the descriptions, or analysis methods.

10. Multiplicity, Missing Data and Analysis Populations

The registry lists a primary outcome and five secondary implementation outcomes. It does not report how multiplicity across these outcomes was addressed, whether any outcome was designated as confirmatory for a statistical test, or whether adjustments to significance thresholds were planned. The primary endpoint description includes a success criterion, but a criterion does not by itself specify a multiplicity procedure or a hypothesis-testing plan.

Missing-data methods are not reported in the registry record. In particular, the record does not explain how unavailable nurse observations, incomplete Spot Check Forms, missing training logs, or unavailable screening records would be handled. It also does not report whether missingness would be described by site or period, or whether any sensitivity analysis was planned.

Analysis populations are not reported in detail. The registry lists enrollment as 18, while the primary outcome says nurses were assessed but were not considered enrolled and describes a random sample of N=30, 10 per ED site. The record does not explain how the enrollment figure relates to staff observations, patient chart review, or implementation activities. It does not define an analysis set, document exclusions, or say how participants or observations were assigned to analysis populations.

The record lists zero analyses and does not report inferential results, a statistical analysis plan, interim analyses, or a sample-size justification. The sponsor’s analysis therefore cannot be assessed for multiplicity, missing-data handling, or analysis-population decisions using the registry record.

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 listing gives site-specific counts for serious adverse events, other adverse events, and deaths. For Hospital Site A, Hospital Site B, and Hospital Site C, the number at risk is 6 for serious events and 6 for other events; the reported counts affected are 0, and deaths are 0 at each site. No serious-event terms are listed.

These are descriptive counts in the registry safety record. They do not, by themselves, establish that no safety events occurred in every person involved in the project or outside the registry’s reporting scope. The record does not give further definitions, observation windows, event ascertainment procedures, or context for the at-risk groups. The counts should therefore be read as the posted safety figures, not as proof that the program had no risks.

The intervention is a behavioral training and screening program, but its nature does not change the limits of the reported safety data. The registry record does not provide additional safety analyses or explanations of the site-specific group sizes.

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

The record establishes that the study was designed to evaluate a delirium screening and detection program and that its primary measure concerns observed nurse accuracy using bCAM. It also describes a site-level rollout and a set of implementation outcomes. These details clarify the questions the project was intended to address, but they are not findings about whether the program changed practice.

The primary endpoint is operationally defined: nurses are assessed with a validated Spot Check Form, and accuracy is judged against a threshold of at least 80%. The stated success criterion is a computed proportion of nurses accurately using bCAM of 80% or higher. However, the observed proportion and whether it met that criterion are not reported in the registry record. No confidence interval, comparison estimate, or sponsor analysis is reported. The result’s magnitude and precision are therefore unknown from this record.

The measure is about nurse screening performance, not patient outcomes. It does not alone show how many patients were screened, how many cases of delirium were detected, whether management changed, or whether clinical outcomes improved. The brief summary describes these broader aims, but the registered primary outcome is narrower.

Likewise, the listed implementation outcomes describe dimensions the investigators intended to assess, not observed levels of reach, adoption, fidelity, or training time. The registry record does not report those results. The trial’s completed status and the presence of outcome definitions should not be mistaken for the posting of findings.

Finally, this is a pilot cluster-randomized evaluation across three EDs. Its design has features that require analysis attentive to site clustering and rollout timing. Because the analysis methods and results are not reported, no conclusion about statistical significance, effect size, or generalizability can be supported from the posted record.

13. Why This Trial Matters Statistically

This trial is statistically notable because the intervention is implemented at the emergency-department level, while the primary outcome is measured on nurses. That difference between the unit of implementation and the unit of measurement is important. Nurses working at the same site may have shared training and local procedures, so their observations may not be statistically independent. In general, study analyses need to reflect the cluster structure when estimating uncertainty.

The stepped wedge structure adds timing to the comparison. Sites receive the program at different points, and implementation status can therefore be entangled with calendar time. A suitable analysis needs to distinguish changes associated with the intervention from changes that might occur over time for other reasons. The registry record identifies the design as a pilot stepped wedge cluster randomized trial, but does not report the model used to address these features.

The endpoint also illustrates the distinction between a threshold-based performance measure and a complete account of performance. Classifying nurses by whether they reach at least 80% accuracy makes the intended outcome concrete, while the proportion meeting that threshold would still need an estimate of uncertainty to show precision. The record gives the criterion and planned observation approach, but not the observed proportion or interval.

Finally, the registry lists implementation outcomes alongside the accuracy endpoint. This matters because staff training, adoption, fidelity, and time are different dimensions of program delivery. Separating these questions can help clarify whether a program was delivered and used as intended, independently of whether the primary accuracy criterion was achieved. In this record, those outcomes are defined but their results and analyses are not reported.

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.

The main limitation for statistical interpretation is that the registry record does not report primary or secondary outcome results. Without an observed nurse proportion, an estimate of uncertainty, or a sponsor analysis, the stated accuracy criterion cannot be evaluated. The completed status does not fill that gap.

The design has three ED sites, and the registry describes a pilot cluster-randomized stepped wedge trial. With site-level allocation and implementation, the number of sites is central to how much independent information the design contains. The registry record does not report the analysis strategy, the site-specific outcome estimates, or how clustering and rollout timing were handled.

The record also gives enrollment as 18 while describing a random nurse sample of N=30, 10 at each ED site, and explicitly says nurses were not considered enrolled for the primary measure. It does not reconcile these descriptions. The number of nurses observed, whether planned observations were completed, and the relationship between staff activities and enrollment are not reported in the registry record.

Further details needed to assess measurement and potential bias are not reported, including the number of observations per nurse, missing-observation handling, blinding of outcome assessment, and the statistical analysis plan. The trial is listed as having no masking, but the record does not specify how that applied to outcome assessment. Details of the site randomization sequence and the precise timing of measurements are also absent.

Finally, the program is evaluated across three diverse EDs, but the record does not report site characteristics or results by site. The extent to which findings would apply to other settings cannot be assessed, particularly because no outcome findings are posted.

15. Related Tutorials

16. Related Statistical Calculators

17. Sources

• ClinicalTrials.gov: Implementation Outcome Assessments of the Emergency Department Delirium Screening and Detection Program — NCT05606328.
• Registry record last updated 2026-01-30; status last verified 2026-01.
• Results first posted 2026-01-30.
• 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

NCT05606328, titled “Implementation Outcome Assessments of the Emergency Department Delirium Screening and Detection Program,” is listed as completed. The registry describes a randomized, unmasked, sequential pilot stepped wedge cluster randomized trial across three emergency departments. Enrollment is listed as 18, with a start date of 2023-02-28 and primary completion date of 2024-01-12.

The behavioral intervention is the Emergency Department Delirium Screening and Detection Program. The primary outcome concerns the proportion of observed nurses meeting the bCAM screening-accuracy threshold using a validated Spot Check Form. Secondary outcomes address training reach, screening adoption, fidelity, and training time. The registry record defines these outcomes but does not report outcome results or sponsor analyses.