Improving Patient-Centered Communication in Breast Cancer Through Patient and Provider Interventions: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Improving Patient-Centered Communication in Breast Cancer Through Patient and Provider Interventions (NCT04549571), a randomized na study. The primary outcome is Patient Knowledge About Risks and Benefits of Local Regional Treatment.

2. Clinical Question

3. Trial Design

This is a randomized, double-masked, factorial trial with registry enrollment of 614. The study evaluates two intervention components: a patient-facing decision tool and a clinician dashboard. The patient-level comparison is between iCanDecide-ESE and iCanDecide-S. The practice-level comparison is dashboard implementation versus usual care. The factorial structure means the patient intervention is evaluated in the context of both dashboard and usual-care practices, and the dashboard intervention is evaluated with both patient-facing tools.

The record describes staggered implementation of the dashboard across clinics. Clinicians in the dashboard groups received training beginning 4 weeks before their practice began dashboard use; the listed clinic groups began use at different weeks and continued according to the stated schedule. Clinics assigned to usual care continued breast cancer surgical care per usual care. Because implementation is organized by clinic, the record's analysis note identifies the design as cluster-randomized and cautions that a simple individual-level re-analysis would be misleading.

Patients used their assigned website and underwent surgery within 5 weeks of registration. The registry also describes optional audio-recorded phone interviews for patients at 9-12 months post registration and for clinicians. These interviews are not the primary endpoint. The study includes breast cancer conditions across the listed stage categories. The record lists the study as active, not recruiting. Start and primary completion dates are 2021-02-16 and 2024-05-01, respectively.

The phase is listed as NA. The record does not report a detailed allocation procedure, masking implementation, clinic or patient allocation counts, or a statistical analysis plan beyond the primary outcome model description. Those details 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.

Only the first 6 of 7 groups are shown.

T1

Reasons for not completing this period:

T2

Reasons for not completing this period:

T3

Reasons for not completing this period:

T4

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.

Population: No baseline data was collected for clinicians.

6. Primary Endpoint

The registry lists the primary outcome as Patient Knowledge About Risks and Benefits of Local Regional Treatment, measured over 5 weeks.

The primary endpoint is patient knowledge about risks and benefits of local regional treatment. It is defined as the mean percentage of correct answers on a 5-item scale, with a stated range of 0-100%. Higher percentages indicate greater knowledge of treatment risks and benefits. The outcome is assessed at 5 weeks.

A knowledge measure is relevant to a communication and decision-support study because the intervention aims to help patients engage with information about treatment choices. The endpoint focuses on what patients know about treatment risks and benefits, rather than on which treatment they select or on a clinical outcome. The registry describes the instrument as previously developed and pilot tested, but does not identify its items, scoring details beyond the percentage-correct definition, or measurement properties in this trial. Those details are not reported in the registry record.

The endpoint is continuous and is summarized as a mean with standard deviation. The stated primary analysis uses a linear mixed-effects model because the intervention is fully crossed across patient and practice intervention conditions and the outcome is continuous. The record does not report the observed endpoint means, standard deviations, effect estimates, or an uncertainty interval. As a result, the registry description establishes what the outcome is intended to measure, but does not show the observed level of knowledge or the estimated intervention effects.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The primary outcome is continuous: it is a mean percentage of correct answers on a knowledge scale. The registry describes a linear mixed-effects model as the primary analysis. The distinction between the endpoint and the model matters: the endpoint defines the measurement, while the model specifies how outcome variation and the trial's design are handled when estimating intervention effects.

Why use a factorial model?

In general, a factorial design studies more than one intervention component in the same trial. Here, the two components are the patient decision tool and the clinician dashboard. The stated plan estimates the effect of iCanDecide-E and the effect of the CDB in one model because the interventions are fully crossed. That means each intervention's comparison is considered across the levels of the other intervention, rather than treating one intervention as if it occurred in isolation.

In general, a factorial model can also assess whether the effect of one component depends on the presence of the other. The registry description does not say whether such an interaction was included or how it was handled. It also does not specify whether the reported intervention effects, if results were posted, would be marginal averages across the other factor or conditional comparisons. Those model details are not reported in the registry record.

Why does clustering matter?

The dashboard is implemented at the clinic level, and the record's analysis note identifies the trial as cluster-randomized. Participants treated in the same clinic may have more similar outcomes than participants in different clinics because they share clinicians and practice processes. In general, this similarity is called within-cluster correlation. If an analysis treats all patient observations as independent when outcomes are correlated within clinics, its standard errors can be too small and its confidence intervals too narrow. A mixed-effects model can account for such grouping by representing clinic-level variation, though the particular random-effects structure used in this trial is not reported in the registry record.

The scheduled, staggered introduction of the dashboard also makes time relevant to interpretation. In general, implementation timing can be associated with changes in practice or patient mix over time. The record notes that the primary model covers the intervention conditions across every time period, but does not explain how period effects or other calendar-time changes were modeled. Those details are not reported in the registry record.

What does a mean and standard deviation convey?

The mean describes the average percentage-correct score in a group or condition. The standard deviation describes variation in scores around that mean. In general, the standard deviation is not the same as the uncertainty around a mean or treatment-effect estimate: precision also depends on the design, the amount of information, and correlation among observations. The registry does not report the primary outcome means or standard deviations, so their values and the extent of score variability cannot be interpreted.

How should uncertainty be read?

In general, a confidence interval gives a range of effect values compatible with the data and the assumptions of the analysis at a stated confidence level. Its width conveys precision: wider intervals indicate less precision than narrower intervals. A p-value addresses compatibility with a specified null hypothesis under the model assumptions; it does not measure the size or importance of an effect. The registry does not post a confidence interval or p-value for the primary outcome, so neither the precision nor the strength of evidence can be evaluated from the record.

What is over-dispersion?

In general, over-dispersion refers to observed variability greater than a statistical model expects. It is especially relevant to count outcomes when a model assumes a particular mean-variance relationship. The primary endpoint is continuous and the stated analysis is a linear mixed-effects model, so an over-dispersion adjustment is not specified as part of the primary analysis. The record does not describe diagnostics or model checks.

What does the planning illustration mean?

The registry describes a sample-size planning illustration based on the observed effect, with a two-sided significance level of 0.05, equal allocation between two arms, and a normal approximation. It says the illustration varies the effect to detect as a stated fraction of the observed effect. This is a planning exercise, not observed power, and not the trial's own design justification. In general, post hoc calculations that use an observed effect do not establish the probability that the trial detected a true effect. Specific planning values are not included in the registry information available for interpretation, and the protocol's sample-size justification is not reported in the registry record.

10. Confidence Intervals for Each Arm

No two-arm comparison is shown for this trial, but each arm's mean can be given with a t-based 95% confidence interval computed from the posted mean, spread and number of participants.

R code and output

arm <- c("iCanDecide -S, No CDB", "iCanDecide -S, CDB", "iCanDecide -ESE, No CDB", "iCanDecide -ESE, CDB")
n   <- c(99, 101, 99, 104)
m   <- c(62.6, 59.2, 29.1, 60.4)
sp  <- c(33.1, 33.6, 29.1, 31.4)
sd  <- sp      # posted standard deviations
for (i in seq_along(n)) {
  se <- sd[i] / sqrt(n[i]); h <- qt(0.975, n[i] - 1) * se
  cat(sprintf("CARM|%s|%d|%.3f|%.3f|%.3f|%.3f\n", arm[i], n[i], m[i], sd[i], m[i] - h, m[i] + h))
}

CARM  iCanDecide -S, No CDB  99  62.600  33.100  55.998  69.202
CARM  iCanDecide -S, CDB  101  59.200  33.600  52.567  65.833
CARM  iCanDecide -ESE, No CDB  99  29.100  29.100  23.296  34.904
CARM  iCanDecide -ESE, CDB  104  60.400  31.400  54.293  66.507

11. Multiplicity, Missing Data and Analysis Populations

The trial evaluates two intervention components and lists secondary outcomes of patient self-efficacy in dealing with breast cancer and patient cancer worry. The registry identifies one primary outcome, patient knowledge about risks and benefits of local regional treatment, but does not describe a multiplicity strategy for testing the primary and secondary outcomes or for considering both intervention effects. A multiplicity plan can matter when several hypotheses are assessed because the chance of at least one false-positive finding can increase as the number of tests grows. Whether any adjustment or hierarchical testing approach was planned is not reported in the registry record.

The factorial design raises additional questions about which effects and interactions were tested. The primary description identifies the iCanDecide-E and CDB effects in a single mixed-effects model, but does not state whether an interaction was assessed or how any multiplicity across effects was addressed. These details are not reported in the registry record.

The record does not describe missing primary-outcome data, missing-data assumptions, imputation, sensitivity analyses, or procedures for incomplete questionnaires. It also does not specify the analysis population, such as whether the primary analysis included all randomized participants or used another population definition. These details are not reported in the registry record. Without them, the impact of missing measurements and the exact population represented by a future estimate cannot be assessed.

The registry lists zero analyses and posts no primary or secondary outcome estimates. Consequently, it is not possible to determine how many outcome comparisons were actually conducted, whether multiplicity was addressed in the sponsor's analysis, or whether missing data affected the reported findings.

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

The registry record describes a trial designed to test whether patient-facing decision support and a clinician dashboard influence patient-centered communication outcomes, with knowledge about local regional treatment risks and benefits as the primary endpoint. It also describes a factorial analysis intended to estimate the two intervention effects in a mixed-effects model while respecting the clinic-level design.

What the record does not establish is the result of that test. No primary outcome estimate, group mean, confidence interval, or p-value is posted. There is therefore no basis to say whether knowledge scores differed between intervention conditions, how large any difference was, or how precisely it was estimated. The same limitation applies to the listed secondary outcomes of self-efficacy and cancer worry.

Even if an estimate were available, interpretation would depend on the outcome scale, model specification, clinic clustering, missing data, and the set of comparisons considered. A statistical difference would not by itself establish that a communication change is important to patients; conversely, an imprecise estimate would not establish that there is no difference. The registry record does not state a threshold for a clinically meaningful difference on the knowledge scale.

The adverse-event entries are not a substitute for outcome results or an analysis of intervention effects. They do not establish a relationship between an event and an intervention. The planning illustration based on an observed effect is also not a result and should not be read as observed power. Overall, the record permits interpretation of the trial's aims and planned primary analysis, but does not support a conclusion about effectiveness, equivalence, or the size of any outcome effect.

13. Why This Trial Matters Statistically

Statistically, this trial is notable for combining a patient-level decision-support intervention with a clinic-level clinician dashboard in a factorial, cluster-randomized design. That structure creates a clear opportunity to study the separate intervention components while accounting for the fact that dashboard implementation occurs within clinics. It also means that an analysis based only on independent patient observations could misstate uncertainty, as the registry's analysis note cautions.

The primary endpoint is directly tied to the communication aim: it measures patient knowledge of treatment risks and benefits rather than treatment choice or a clinical endpoint. The use of a continuous score and a mixed-effects model is consistent with the record's description of a mean percentage-correct outcome and clustered implementation. How well the analysis accounts for clinic grouping, timing, and the factorial structure would be important to interpretation, but several model details are not reported in the registry record.

The registry also illustrates why a detailed result record matters. A design description alone cannot show the estimated effect or its precision. Without posted outcome estimates, the statistical question remains unresolved in the public record, even though the endpoint and intended model are described. Reporting confidence intervals, analysis populations, missing-data handling, and the approach to multiple intervention and outcome comparisons would help readers assess the strength and scope of the evidence.

Finally, the planning illustration should be kept separate from evidence about results. It is described as a planning exercise derived from an observed effect, not as observed power or the original sample-size justification. Distinguishing design assumptions from estimated effects is essential when interpreting trials with complex allocation and clustered outcomes.

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 the trial is the absence of posted outcome results. The registry identifies the primary endpoint and gives a general model description, but lists zero analyses and provides no primary or secondary outcome estimates, standard deviations, confidence intervals, or p-values. The direction, magnitude, and precision of the intervention effects therefore cannot be assessed from the registry record.

The primary model description is brief. It does not specify the random-effects structure, fixed effects, covariates, handling of time, interaction terms, or model diagnostics. Although the record emphasizes the factorial structure and cluster-randomized design, it does not provide enough detail to reproduce the intended analysis. The allocation procedure and masking implementation are also not described in detail.

Information about the measurement instrument is limited to its 5-item format, percentage-correct scoring, range, and prior development and pilot testing. The registry does not provide the item content, reliability, validity evidence, or a threshold for a meaningful change in the scale. Missing-data procedures and the analysis population are not reported either.

The record lists secondary outcomes but does not describe a multiplicity plan, and it does not state whether an interaction between patient and clinic interventions was tested. The clinic-level staggered implementation makes treatment, practice, and time effects important to distinguish, but the modeling of time is not reported. Finally, the listed sample-size illustration is not the trial's own design justification, and its detailed planning values are not available in the registry description. These reporting gaps limit assessment of both robustness and generalizability.

15. Related Tutorials

16. Related Statistical Calculators

17. Sources

• ClinicalTrials.gov: Improving Patient-Centered Communication in Breast Cancer Through Patient and Provider Interventions — NCT04549571.
• Registry record last updated 2026-10-27; status last verified 2026-10.
• Results first posted 2026-10-27.
• 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

NCT04549571 is a randomized, double-masked factorial trial sponsored by the Alliance for Clinical Trials in Oncology. It evaluates iCanDecide-ESE versus iCanDecide-S and clinician dashboard implementation versus usual care in breast cancer communication. Registry enrollment is 614, and the study status is active, not recruiting.

The primary endpoint is patient knowledge about risks and benefits of local regional treatment, measured as mean percentage correct on a 5-item scale at 5 weeks. The registry describes a single linear mixed-effects model for the two intervention effects and notes that simple individual-level re-analysis would be misleading because the design is cluster-randomized.

The record lists patient self-efficacy and cancer worry as secondary outcomes, but posts no outcome analyses or estimates. Details about missing data, analysis populations, multiplicity, and several elements of the statistical model are not reported in the registry record.