Nudging Provider Adoption of Clinical Decision Support: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Nudging Provider Adoption of Clinical Decision Support (NCT05203185), a non-randomized na study. The primary outcome is Percentage of Patient Visits Where the Provider Adopted the Pulmonary Embolism Risk Kalculator (PERK) Tool.

2. Clinical Question

3. Trial Design

NCT05203185, titled “Nudging Provider Adoption of Clinical Decision Support,” is listed as completed. The registry record gives a start date of 2021-07-01 and a primary completion date of 2022-04-30. Enrollment is listed as 1612. The study is recorded as a parallel, nonrandomized comparison with no masking and phase listed as NA.

The intervention was delivered at an Intervention Site, where full-time employed Emergency Department providers used PERK. At the No Intervention Site, providers used PE CALC without nudges. These descriptions frame the comparison as use of different tools at different sites, rather than a randomized allocation of providers or visits. The registry record does not explain how sites or providers were selected or assigned, or whether any other differences between sites were addressed.

The stated central hypothesis was that adding a theory-informed nudge to clinical decision support would address behavioral barriers to use and improve provider adoption. The registry describes development of the tool using a behavioral theory-informed process, but details of that process and the specific nudges are not reported in the registry record. Nor are protocol details describing the way exposure was measured or the primary comparison was estimated.

The primary outcome was the percentage of patient visits where the provider adopted PERK, assessed using electronic health record data up to Month 6. A secondary outcome was the percentage of PE-designated CTPA tests that were positive for PE. Although both outcomes are identified, numerical outcome results are not reported in the registry record. Thus, the design and intended measures can be described, but the posted record does not establish the size or precision of an observed difference.

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.

Population: Sex, ethnicity and race data were not collected from Provider participants. Sex data were not collected from Patient participants.

6. Primary Endpoint

The registry lists the primary outcome as Percentage of Patient Visits Where the Provider Adopted the Pulmonary Embolism Risk Kalculator (PERK) Tool, measured over Up to Month 6.

The primary endpoint is the percentage of patient visits in which the provider adopted PERK, with measurement specified as up to Month 6. The registry defines adoption behaviorally: during a visit, the provider must accept a recommendation from the tool and order what it recommends. The examples are a recommended D-dimer that is ordered or a recommended CT pulmonary angiogram (CTPA) that is ordered. The outcome was assessed using electronic health record data from the patient visit.

This measure focuses on uptake of the decision-support recommendation rather than on whether the diagnostic result was correct or whether patient health outcomes changed. That makes it relevant to the stated question about whether adding nudges can address barriers to tool use. In general, a visit-level percentage can describe how often a defined behavior occurred among visits meeting the analysis criteria. To interpret such a percentage, readers also need to know how eligible visits and missing records were handled; those details are not reported in the registry record.

The registry record does not report the observed percentage, the denominator used in its calculation, or an uncertainty interval. The secondary outcome concerns the proportion of PE-designated CTPA tests that were positive for PE, but its result is also not reported in the registry record. Therefore, the endpoint definition is available, while the numerical findings needed to assess the observed adoption level or its precision are absent.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The study's primary outcome is a percentage built from visit-level behavior: a provider accepted a recommendation from PERK and ordered the recommended test. This makes adoption a defined event for each visit. In general, a percentage is interpretable only when the eligible set of visits, the event definition, and the handling of missing or incomplete records are clear. The registry record supplies the event definition and names electronic health record data as the source, but it does not report the analysis denominator or missing-data procedures.

What does a visit-level percentage describe?

A visit-level percentage describes the frequency of the specified action across visits included in an analysis. It is not automatically a measure of provider-level behavior, since one provider may contribute multiple visits. When observations from the same provider or site are related, treating every visit as statistically independent can make uncertainty appear smaller than it is. This is a general statistical concern; whether and how the study accounted for repeated visits or clustering is not reported in the registry record.

The outcome definition combines acceptance of a recommendation with an order. A visit where the provider did not order the recommended test would not meet that definition, even if other clinical actions occurred. The registry record does not describe how it classified visits when the recommendation was absent, the record was incomplete, or the provider selected an alternative action. Those operational details matter because different classification rules can change the estimated adoption percentage.

What can a nonrandomized comparison establish?

Random allocation, in general, helps make groups comparable by distributing measured and unmeasured factors across study conditions. This study is recorded as nonrandomized, with PERK at the Intervention Site and PE CALC at the No Intervention Site. A difference between their outcomes could reflect the tools, but could also reflect other differences associated with site or implementation. This is a general limitation of nonrandomized comparisons, not evidence that a particular bias occurred in this study.

To assess comparability, readers would need information about the sites, providers, eligible visits, and any adjustment strategy. The registry record does not report baseline characteristics, an adjusted model, or the covariates that might have been considered. It also does not report whether providers could use other decision-support resources or how implementation fidelity was assessed. Without these details and numerical results, neither the direction nor magnitude of a contrast can be evaluated.

How should uncertainty be understood?

In general, a confidence interval communicates uncertainty around an estimated effect. It reflects the variability expected under a statistical model and the data used, while relying on the model's assumptions. A point estimate alone would not show how precisely an effect had been estimated; an interval would help distinguish a result that is compatible with a range of values from one estimated more tightly. The registry record does not report a point estimate, confidence interval, or hypothesis-test result for the primary outcome.

The study lists a secondary percentage for PE-designated CTPA tests that were positive for PE. This is a test-yield measure, not the same endpoint as provider adoption. In general, test yield can be affected by which patients are tested as well as by test performance. Since the registry record does not report the result or analytic details for this secondary outcome, it cannot be used to infer how adoption affected imaging decisions or diagnostic yield.

What do planning illustrations mean?

The accompanying sample-size material is described as a planning illustration calculated from the observed effect, not as the trial's original design. Its stated assumptions are a two-sided significance level of 0.05, equal allocation between arms, and a normal approximation, with the effect to detect expressed as a stated fraction of the observed effect. For outcomes other than time-to-event outcomes, the illustration is expressed as participants per arm for 80% power or another stated planning level.

Such sample-size values are planning illustrations, not observed power and not evidence that the study had a particular probability of detecting an effect. They depend on the effect and assumptions used in the calculation. The registry record does not report the trial's original sample-size justification, or the statistical plan used for the primary analysis. The planning assumptions do not fill that gap, and the illustration cannot substitute for an outcome estimate and its uncertainty.

10. Multiplicity, Missing Data and Analysis Populations

The registry record identifies a primary outcome and a secondary outcome: adoption of PERK across patient visits, and the percentage of PE-designated CTPA tests that were positive for PE. It does not report numerical results for either outcome. Although 2 outcome measures are listed, the record does not describe a multiplicity strategy, such as whether analyses were treated as confirmatory or whether any adjustment was planned for interpreting multiple tests. The analysis details are not reported in the registry record.

Missing-data handling is also not reported. The primary outcome relies on electronic health record data, but the record does not state how incomplete records, unobserved visits, or visits without a classifiable recommendation were treated. It does not report the number of visits excluded from analysis or whether any sensitivity analyses assessed the effect of missing information. These details affect which observations contribute to an estimate and how confidently it can be interpreted.

The registry lists no analysis records. It therefore does not describe analysis populations, inclusion and exclusion rules, or whether all enrolled participants, providers, or visits were included in the reported analyses. Enrollment is not necessarily the same as the endpoint denominator, and the record does not give that denominator. The analysis plan, population definitions, missing-data procedures, and multiplicity approach are not reported in the registry record, so they cannot be inferred from the study's outcome definitions.

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 lists serious adverse event counts for Intervention Site - Patients and No Intervention Site - Patients. It reports 0 serious events among 793 patients in the intervention-site group and 0 among 735 patients in the no-intervention-site group. The other event counts and deaths are also listed as 0 for each group.

These figures describe what the registry record reports for the listed groups; they do not establish that no adverse events occurred outside the reporting scope or that every possible harm was assessed. The record does not provide details about event collection, definitions, surveillance procedures, or the timeframe covered by these counts. The listed numbers should therefore be read as registry-reported counts, not as proof of an absence of risk.

The safety figures also do not provide a comparison adjusted for differences between sites or details about which patients were exposed to which tool. As with the adoption outcome, the nonrandomized site comparison limits causal interpretation. These counts should not be used alone to draw conclusions about safety or to guide care.

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

This record describes a study designed to examine whether adding behavioral nudges to clinical decision support could encourage provider adoption. It specifies a primary measure based on the percentage of patient visits where a provider accepted and ordered the PERK recommendation, and it identifies PE CALC without nudges as the comparison tool at a separate site. These details clarify the intended question and the operational definition of adoption.

The posted record does not report the numerical result for the primary or secondary outcome. It therefore does not show how often PERK was adopted, whether adoption differed between sites, or how precise any estimated difference was. The available information cannot support a conclusion that nudges increased adoption, had no effect, or changed the positivity of PE-designated CTPA tests.

Even with outcome results, an association in this nonrandomized comparison would require careful interpretation. The intervention and comparison conditions were implemented at different sites, so differences in context could contribute to any observed contrast. The record does not report adjustment methods or details that would allow readers to judge how those influences were handled.

Adoption is also a process measure, not a direct measure of patient health. An order matching a recommendation does not establish that the test was appropriate in every case, that diagnoses improved, or that unnecessary imaging was reduced. The secondary outcome could speak to the positivity of designated tests, but its result is not reported and it would not, by itself, establish a patient-level benefit.

In short, the registry record conveys the intervention concept, comparator, and outcome definitions, but not the empirical findings needed to assess the effect. The distinction between a study's hypothesis and its reported results is essential: the rationale motivates the evaluation, while conclusions about observed outcomes require the missing estimates, uncertainty, and analysis details.

13. Why This Trial Matters Statistically

Clinical decision support can be available in an electronic record without being routinely used. This study's stated focus on nudges addresses a statistical and implementation question: whether a change in how a tool is presented is associated with provider adoption, rather than whether the algorithm alone can generate a recommendation. The primary endpoint makes that implementation target explicit by defining adoption as acceptance and ordering of a recommendation during a visit.

That focus matters because uptake is a step between deploying a tool and realizing any downstream effect. In general, an intervention that does not change behavior may have little opportunity to alter later decisions, while increased use alone does not prove that decisions or patient outcomes improve. Separating these stages helps readers avoid treating an implementation measure as a clinical outcome.

The design also highlights why the comparison structure matters. The nonrandomized, site-based contrast can be informative about implementation in the settings studied, but it is vulnerable to differences between sites that are not caused by the tools. Statistical estimates and transparent reporting of adjustment and uncertainty would be central to evaluating such a comparison. Those results and methods are not reported in the registry record.

Finally, the record illustrates the difference between a well-defined research question and a fully interpretable result. It gives the adoption definition and identifies a secondary diagnostic-yield measure, but lacks numerical findings and analysis details. Readers can understand what investigators intended to measure, while remaining unable to judge effect size, precision, or the robustness of any conclusion from the posted 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 central limitation of the available record is that it does not report numerical results for either listed outcome. There is no adoption estimate, comparison estimate, confidence interval, or significance test to assess. The primary endpoint's denominator is also not reported, so the enrollment figure cannot be treated as the number of visits contributing to that analysis.

The study is nonrandomized, and the intervention and comparison are described at different sites. This structure allows site-level context to be entangled with the tool condition. The registry record does not report site or provider characteristics, an adjustment strategy, or other analysis details that might help evaluate comparability. This means the record alone cannot establish that a difference, if one were reported elsewhere, was caused by the nudges.

The record does not describe the detailed implementation of the nudges, adherence to the tools, or how providers and visits were selected. It also does not report how repeated visits by providers were handled statistically, how incomplete electronic records were classified, or which analysis population was used. These omissions limit assessment of outcome measurement and uncertainty.

The primary measure captures adoption behavior, not appropriateness of care or patient health. The secondary measure concerns positivity among PE-designated CTPA tests and likewise has no posted numerical result. The serious adverse event counts are reported, but event collection procedures and the timeframe for those counts are not described. Overall, the registry record provides study context and definitions, but not enough results or methodological detail to assess effectiveness, precision, or the robustness of conclusions.

15. Related Tutorials

16. Related Statistical Calculators

17. Sources

• ClinicalTrials.gov: Nudging Provider Adoption of Clinical Decision Support — NCT05203185.
• Registry record last updated 2023-06-26; status last verified 2023-06.
• Results first posted 2023-06-26.
• 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

NCT05203185 is a completed, parallel, nonrandomized study with no masking. The intervention site used PERK, a pulmonary embolism decision-support tool with nudges, while the no-intervention site used PE CALC without nudges. Enrollment is listed as 1612.

The primary outcome was the percentage of patient visits where a provider adopted PERK by accepting and ordering its recommendation, assessed using electronic health record data up to Month 6. The secondary outcome was the percentage of PE-designated CTPA tests that were positive for PE. Numerical results for these outcomes and analysis details are not reported in the registry record. The record lists 0 serious adverse events in each of the two patient groups, with the event counts and deaths also listed as 0.