Increasing Screening for Cancer Using EHR-Nudges: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Increasing Screening for Cancer Using EHR-Nudges (NCT06177795), a randomized na study. The primary outcome is Proportion of Patients Who Complete a Screening Mammogram Within 3 Months After the Visit.

2. Clinical Question

I-SCREEN asks whether changes to routine electronic health record workflows and post-visit communication are associated with mammogram completion. The planned approaches differ between the two participating systems: Penn’s intervention combines a default pended order for clinicians with patient text reminders, while the University Hospitals (UH) intervention is described as default pended orders for providers. The registry summary describes two complementary, concurrent, 6-month, cluster-randomized pragmatic trials.

The study also includes a distinct question within Penn’s intervention clinics. Patients identified as high risk for noncompletion were individually randomized to additional bidirectional texting or standard text messaging. This comparison concerns an intensification of the patient messaging intervention, rather than the overall intervention-versus-control comparison.

The registry record lists 21,123 participants enrolled and the study status as completed. It does not report the primary outcome results, so the posted record cannot answer whether completion differed between groups or how precisely any difference was estimated. The clinical question is clear; the result needed to address it is not reported in the registry record.

3. Trial Design

The registry describes I-SCREEN as a randomized, double-masked, parallel study with a total enrollment of 21,123. It lists breast cancer as the condition. The study is titled “Increasing Screening for Cancer Using EHR-Nudges,” and its official title identifies it as a randomized evaluation of EHR-based nudges.

The brief summary describes two concurrent, 6-month, cluster-randomized pragmatic trials conducted with Penn Medicine and Case Western Reserve University-University Hospitals. At Penn, clinics were randomized to control or intervention. The control clinics received standard of care; intervention clinics received clinician and patient nudges. The clinician nudge was a default pended mammogram order, and the patient nudge was a post-visit text reminder. Within intervention clinics, high-risk patients were separately randomized to additional bidirectional texting or the standard intervention messaging.

At UH, primary care providers were randomized to control or intervention. The intervention is described as default pended mammogram orders; the control was standard of care. The registry therefore describes related but not identical intervention packages across the two settings. It does not report a single common intervention delivered identically at both sites.

The study start date is 2023-12-18, the primary completion date is 2025-01-07, and the status is completed. The registry lists masking as double, but it does not report who was masked or how masking was maintained in a study involving visible EHR orders and patient communications. The number of randomized clinics or providers, allocation procedures beyond the listed randomization, and details of the study's statistical 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.

Before assignment: Penn clinics were randomized 2:1 to intervention and control. Penn patients seen at an intervention clinic and identified as high risk were further randomized 1:1 to standard messaging or an intensification nudge. UH primary care providers were randomized 1:1 to intervention or control. No individual randomization took place at UH.

Cluster Level Randomization

Individual Level Randomization (Penn)

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: Baseline characteristics are reported for the Main Study Period only - Cluster Level Randomization (Control and Intervention). Patients in the Penn High Risk arms are accounted for in the Intervention arm as a whole, but were additionally randomized at the Individual Level, as noted in the Participant Flow. At UH, clinicians were randomized but no clinicians were individually enrolled as participants. All analyses were done at the participant level at both Penn and UH.

6. Primary Endpoint

The registry lists the primary outcome as Proportion of Patients Who Complete a Screening Mammogram Within 3 Months After the Visit, measured over 3 months.

The primary endpoint is the proportion of patients who complete a screening mammogram within 3 months after the first eligible primary care visit. The registry describes the outcome as a count of participants and specifies a 3-month timeframe. This is a binary outcome at the participant level: a patient either completed a screening mammogram within the specified period or did not.

In general, screening completion is a direct measure of whether the intended screening step occurred, rather than an intermediate measure such as whether an order was placed or a reminder was sent. That distinction is relevant because the interventions act on clinical workflow and communication, while the primary outcome concerns completion. The registry also lists completion within 6 months after the visit as a secondary outcome. The posted record does not report the number or proportion completing screening at either time point.

In general, a defined follow-up window makes clear which events count toward an endpoint and gives groups a common period for outcome ascertainment. The registry's primary window is 3 months after the first eligible visit. How completion was verified, how patients without observed follow-up were handled, and whether the endpoint definition included any additional eligibility conditions are not reported in the registry record.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

What does cluster randomization mean?

General explanation: In an individually randomized trial, each participant is assigned separately. In a cluster-randomized trial, groups are assigned, and participants in those groups are then observed. The registry describes clinic-level randomization at Penn and provider-level randomization at UH for the main comparisons. This means that participant outcomes are grouped according to the clinic or provider assignment.

Participants treated within the same clinic or provider group may share workflows, staff, and local practices. Their outcomes can therefore be correlated. In general, this correlation reduces the amount of independent information in a dataset compared with treating every participant as unrelated. A suitable analysis accounts for the grouping when estimating uncertainty. The registry record does not report the analysis model, how clustering was handled, or the number of randomized clusters. It does caution that a simple individual-level re-analysis would be misleading and that only the sponsor's own analysis is shown.

Why does the outcome's binary form matter?

General explanation: A binary endpoint records whether an event occurred for each participant. Here the event is completion of a screening mammogram within the specified follow-up window. Analyses of binary outcomes can compare the chance or proportion of an event between groups, but the appropriate estimate and uncertainty depend on the analysis plan and design. The record supplies no estimate for the primary outcome and does not report the analysis method, so no particular effect measure can be attributed to the sponsor.

The primary outcome is stated as a proportion, while its registry parameter is a count of participants. A count by itself would not show a completion proportion without a relevant denominator. The enrollment total is not necessarily that denominator: the number assigned to each group and the number with usable outcome data are not reported. The record also does not state how the screening outcome was ascertained.

What does a confidence interval tell readers?

General explanation: A confidence interval describes uncertainty around an estimated effect under the assumptions of the analysis. It indicates which effect values are reasonably compatible with the observed data and the statistical procedure. A narrow interval indicates greater precision than a wide interval, while an interval that includes no difference may be compatible with both no effect and some effects in either direction. Interpretation also depends on the scale of the estimate and the study design. No confidence interval for the primary outcome is reported in this registry record.

A p-value, if reported, would address how compatible the data are with a specified null model under the analysis assumptions. It would not measure the size or importance of an effect, and it would not replace an estimate with its uncertainty. No primary-outcome p-value is reported. The registry lists no analyses, so readers should not infer that a statistical test was or was not conducted outside the posted record.

How should the sample-size illustrations be read?

The planning illustrations are described as calculations based on the observed effect, using a two-sided significance level of 0.05, equal allocation between two arms, and a normal approximation. The stated effect to detect is a fraction of the observed effect. These are planning illustrations, not the trial's own design justification, and they are not observed power. The registry does not report a sample-size justification or the assumptions used to plan the trial. The illustrations should therefore not be mistaken for the enrollment target, a description of the randomization, or evidence that the trial had a particular probability of detecting an effect.

In general, a sample-size calculation asks how much information a study might need under specified assumptions. It depends on the outcome, the effect being targeted, the allocation, and the analysis. For a clustered design, the correlation within groups and the number and size of groups can also matter. The listed illustration assumptions do not establish how the sponsor designed or analyzed this cluster-randomized study. No additional planning assumptions should be inferred from those calculations.

What can be inferred from the posted methods information?

The registry gives the broad design and intervention descriptions, but it does not report the sponsor's statistical analysis plan, the primary comparison definition, adjustments for clustering, covariates, interim analyses, missing-data approach, or analysis populations. It also does not report the primary outcome results. These omissions limit methodological appraisal: a reader cannot verify from the registry record how the planned comparisons were estimated or whether the reported analysis followed a prespecified plan.

In general, assessment of a randomized trial benefits from knowing which comparison was primary, how outcomes and missing observations were handled, and how uncertainty reflected the assignment structure. Those are methodological questions, not assumptions that can be answered from the study title or intervention descriptions. Until the analysis details and results are posted, the most defensible interpretation is descriptive: the record outlines randomized nudges and a screening-completion endpoint, while leaving the statistical analysis and its findings unreported.

10. Exact Estimates for Each Arm

No two-arm comparison is shown for this trial, but each arm's result can still be summarized with exact interval estimates computed from the posted counts. The Clopper-Pearson interval is exact and conservative; the Wilson interval is a common closer-to-nominal alternative; the Bayesian credible interval uses a uniform Beta(1, 1) prior. For small groups the three can differ noticeably.

R code and output

arm    <- c("Penn: Control", "Penn: Intervention", "Penn: High Risk - Standard Messaging", "Penn: High Risk - Intensification Messaging", "UH: Control", "UH: Intervention")
n      <- c(4513, 10392, 2270, 2260, 3075, 3143)
events <- c(943, 2810, 511, 549, 1167, 1390)
for (i in seq_along(n)) {
  ci <- binom.test(events[i], n[i])$conf.int          # Clopper-Pearson exact interval
  cr <- qbeta(c(0.025, 0.975), 1 + events[i], 1 + n[i] - events[i])   # Bayesian, uniform Beta(1,1) prior
  wi <- (function(x, n) { p <- x / n; z <- qnorm(0.975); d <- 1 + z^2 / n
    c((p + z^2 / (2 * n) - z * sqrt(p * (1 - p) / n + z^2 / (4 * n^2))) / d, (p + z^2 / (2 * n) + z * sqrt(p * (1 - p) / n + z^2 / (4 * n^2))) / d) })(events[i], n[i])
  cat(sprintf("ARM|%s|%d|%d|%.1f|%.1f|%.1f|%.1f|%.1f|%.1f|%.1f\n", arm[i], n[i], events[i], 100 * events[i] / n[i],
              100 * ci[1], 100 * ci[2], 100 * wi[1], 100 * wi[2], 100 * cr[1], 100 * cr[2]))
}

ARM  Penn: Control  4513  943  20.9  19.7  22.1  19.7  22.1  19.7  22.1
ARM  Penn: Intervention  10392  2810  27.0  26.2  27.9  26.2  27.9  26.2  27.9
ARM  Penn: High Risk - Standard Messaging  2270  511  22.5  20.8  24.3  20.8  24.3  20.8  24.3
ARM  Penn: High Risk - Intensification Messaging  2260  549  24.3  22.5  26.1  22.6  26.1  22.6  26.1
ARM  UH: Control  3075  1167  38.0  36.2  39.7  36.3  39.7  36.3  39.7
ARM  UH: Intervention  3143  1390  44.2  42.5  46.0  42.5  46.0  42.5  46.0

11. Multiplicity, Missing Data and Analysis Populations

The record lists two outcome measures and identifies mammogram completion within 3 months after the first eligible primary care visit as the primary outcome. It lists completion within 6 months after the visit as a secondary outcome. It does not report the outcome results or the analyses, and the registry lists 0 analyses. A plan for handling multiplicity across outcomes or across the study's distinct intervention comparisons is not reported in the registry record.

Multiplicity arises when a study evaluates several outcomes, groups, or comparisons. In general, the more questions assessed, the greater the chance that at least one result may appear unusual by chance if each is considered separately. Whether and how to account for this depends on the prespecified analysis plan and the inferential goals. No adjustment method, hierarchy of comparisons, or distinction between confirmatory and exploratory analyses is reported. The existence of primary and secondary outcomes alone does not reveal the sponsor's multiplicity strategy.

The registry does not report missing outcome data, the number of participants with screening status available, or how incomplete follow-up was handled. It also does not describe the analysis population, such as whether participants were analyzed according to their randomized assignment or under another rule. These are important details because missing outcome information can affect the interpretation of group comparisons, and analysis-population definitions clarify which participants contribute to an estimate.

There is no basis in the posted record to assume that outcome data were complete, that missing observations were excluded or imputed, or that a particular population was used. These matters are not reported in the registry record. The record also does not report interim analyses or stopping rules. None should be inferred from the completed status alone.

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

The registry record establishes that I-SCREEN evaluated EHR-based and text-message nudges in randomized, pragmatic comparisons focused on mammogram screening completion. It specifies a primary endpoint of completion within 3 months after the first eligible primary care visit and a secondary endpoint of completion within 6 months. It also describes separate intervention arrangements at Penn and UH and an additional high-risk messaging comparison at Penn.

However, the posted record does not report the primary endpoint results or an analysis. It therefore cannot show whether screening completion was higher, lower, or similar between any intervention and control groups. Nor does it show the magnitude or precision of any difference. The completed status and enrollment total do not supply that missing evidence. Readers should not interpret absence of posted results as proof of no effect, and should not infer an effect from the intervention rationale or study design.

Even if an estimate were available, in general its interpretation would depend on the comparison being made, the uncertainty around the estimate, the clustering in the assignment, and the completeness of outcome ascertainment. The main comparisons and the high-risk texting comparison address different intervention questions. An overall statement about “the nudges” would obscure those distinctions.

The record's adverse-event entries are separate from the screening-completion outcome and do not resolve the effectiveness question. The information posted is useful for understanding the trial's question and broad design, but it is insufficient to determine what the randomized comparisons found. Any conclusion about screening effects, their precision, or differences between participating settings would require results and analysis details that are not reported in the registry record.

13. Why This Trial Matters Statistically

This trial is statistically notable because it evaluates interventions embedded in clinical workflows rather than a single treatment delivered identically to every participant. The Penn comparison combines clinician defaults and patient reminders, while the UH intervention is described as provider-facing default orders. At Penn, an additional randomized comparison tests intensified messaging among patients identified as high risk for noncompletion. These layers make clear definitions of each comparison essential when interpreting results.

The design also illustrates why the randomization unit matters. The primary outcome is measured for individual patients, but the main assignment is described at the clinic or provider level. In general, the analysis needs to reflect the grouping created by randomization; treating all patient outcomes as independent may give misleading uncertainty. The registry itself flags this concern and says a simple individual-level re-analysis would be misleading.

The primary outcome is a concrete completion measure with a stated follow-up window, and the record separately identifies a longer follow-up outcome. Such clarity about outcome timing helps define the question. At the same time, without posted estimates, denominators, uncertainty, and analysis details, readers cannot judge the size or precision of any observed difference.

Finally, the record demonstrates the difference between a completed study and a fully interpretable results record. Enrollment, intervention descriptions, and an endpoint definition document what the study set out to examine, but they do not reveal the findings. Statistical interpretation depends on the results and the analysis that accounts for the design.

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 principal limitation for interpreting this trial from the registry record is the absence of primary outcome results and analyses. Although the study status is completed and the primary endpoint is specified, no group-level completion results, effect estimates, confidence intervals, or p-values are reported. The registry lists 0 analyses. The record therefore supports design description, not an assessment of the intervention comparisons.

Important features needed to appraise the analysis are also not reported. These include the number of clinics or providers randomized, the number of participants in each comparison, the analysis model, how clustering was handled, the definition of the analysis population, and the treatment of missing outcome data. The registry does not report how screening completion was ascertained. It also does not report the statistical plan, multiplicity approach, sample-size justification, or interim analyses.

The intervention packages differ across Penn and UH, so results from the two settings would need to be interpreted in relation to the particular intervention and randomization unit. The high-risk bidirectional texting comparison is a separate component within Penn intervention clinics, not the same contrast as the main intervention-versus-control comparisons. The posted information does not provide results for any of these comparisons.

The record labels the study double-masked but does not explain who was masked or how masking was implemented. The intervention descriptions involve orders and text messages, but the practical masking procedures are not reported. These limitations concern what can be learned from the registry record; they should not be taken as evidence that procedures were absent from the full study.

15. Related Tutorials

16. Related Statistical Calculators

17. Sources

• ClinicalTrials.gov: Increasing Screening for Cancer Using EHR-Nudges — NCT06177795.
• Registry record last updated 2026-07-06; status last verified 2026-04.
• Results first posted 2026-07-06.
• 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

I-SCREEN is a completed randomized, double-masked, parallel study of EHR-based nudges for breast cancer screening. The registry reports 21,123 participants enrolled, with a start date of 2023-12-18 and primary completion date of 2025-01-07.

The record describes concurrent, 6-month, cluster-randomized pragmatic trials at Penn and UH. Penn's intervention includes default pended orders and post-visit patient text reminders; UH's intervention is described as default pended orders. A separate individual randomization at Penn compares additional bidirectional texting among patients at high risk for noncompletion with standard intervention messaging.

The primary endpoint is mammogram completion within 3 months after the first eligible primary care visit; the secondary endpoint is completion within 6 months. The primary outcome results, analysis details, missing-data handling, and analysis populations are not reported in the registry record.